Subtitle Edit Changelog ----------------------------------------------------------------------------------------------------- v5.3.0-beta8 (21st of September 2026) * Add editor-style layout 14 with a track timeline (video row and subtitle rows above the waveform) * Add sound on Linux to the FFmpeg video player (PulseAudio / PipeWire) * FFmpeg video player: real frame index, seek-free frame stepping (also variable frame rate) and faster scrubbing * Add "Move lines: shorten previous/next line instead of overlapping it" option - thx gambar3 * Add width setting for the "initial text of selected line" waveform toolbar box - thx kadrimarzouki * WhisperX: new standalone build with live output (no more "hang" at "Performing transcription..."), and an update prompt - thx gambar3 * Batch convert: lock Add/Remove/Clear while converting - thx yozongu * OCR "Add character" windows: show subtitle images on the image preview background - thx Nino-kun * Fix many ffmpeg video generation bugs (burn-in, cut, re-encode, remux, write chapters, blank video, add/remove embedded subtitles) * Fix "Add/remove embedded subtitles" dropping extra audio tracks and font attachments * Fix burn-in of Blu-ray sup showing subtitles too early when the first subtitle starts late * Fix window position lost when closing SE with a minimized window - thx GrampaWildWilly * Fix format properties button missing after restart - thx GrampaWildWilly * Fix waveform subtitle blocks vanishing when scrolling in TTS review / Improve time codes - thx gambar3 * Fix underscores missing from file, audio track and plugin names in menus * Fix error log flooded with "disposed mpv player" entries after closing Visual sync / Point sync * Update Chinese (Simplified) translation - thx emsb8888-prog * Update Japanese translation - thx hisui3393 * Update Korean translation - thx 12si27 * Update Polish translation - thx potplayer-fanpack ----------------------------------------------------------------------------------------------------- v5.3.0-beta7 (20th of September 2026) * Add "Improve time codes (forced alignment)" in Tools: re-time a roughly synced subtitle with a Crisp ASR forced aligner, with waveform preview * Add "Isolate speech" option to Crisp ASR speech-to-text (removes music/noise before transcribing) * Add "Show speech only" waveform option * Add "Remove original speech" option to text to speech * Add Supertonic-3 (Crisp ASR) text to speech engine * Add Voxtral F16 model to Crisp ASR speech-to-text * Add "Move text after cursor, go to next, play and pause at end" shortcut - thx kadrimarzouki * Add optional "initial text of selected line" box to the waveform toolbar - thx kadrimarzouki * Add two "line break" items to the waveform toolbar (choose where the toolbar wraps) * Remember the second subtitle file after restart - thx kadrimarzouki * Point sync: open at the line selected in the main window - thx TristisOris * Settings backup: interval 0 = backup at every start, unchanged settings are skipped - thx GrampaWildWilly * Spell check: accept capitalized lowercase-only words at sentence start (e.g. Dutch "Oktober") - thx fraternl * OCR word split and auto-break of 3+ lines now find the best split - thx ivandrofly * Zonos (Crisp ASR) is no longer presented as a voice-cloning engine * Fix Google Translate V2 API adding the detected source language to the translation - thx DaTim69 * Fix waveform play-head jumping to the start when changing settings/docking - thx kadrimarzouki * Fix text to speech "add audio to video" failing with TrueHD audio - thx nielka98 * Fix text to speech "generate video" failing on XviD video without time stamps * Fix white title bar on some dialogs in dark theme - thx Surlime * Fix tooltips not showing when another SE window is active (undocked) - thx GrampaWildWilly * Fix error log being written on every start with routine diagnostics - thx GrampaWildWilly * Fix plain text import "split to four lines" leaving long text unbroken * Fix seconv reading an MP4 without subtitle tracks as text - thx Ro-meo * Fix untranslatable strings in waveform themes and OCR "Add character" windows * Update Crisp ASR to v0.8.34 (Linux CUDA builds are back) * Update Japanese translation - thx hisui3393 * Update Polish translation - thx potplayer-fanpack * Update Turkish translation - thx bilimiyorum ----------------------------------------------------------------------------------------------------- v5.3.0-beta6 (19th of September 2026) * Add actor picker, actor shortcuts now work everywhere and keep a stable order - thx zbik1 * Add optional "go to changed line" on undo/redo - thx kadrimarzouki * Add "Set end and go to next" and "Play from just before text" to the waveform toolbar - thx JonEvanCook * Add two missing SE4 options to Settings > Tools (remember "Use always" list, short display times may move start time) - thx ChocOranger * Spell check: accept Spanish imperative + pronoun forms (dímelo, cuéntame, ...) - thx fraternl * Text to speech: a failed per-line voice clone line falls back to a neighbouring line's reference clip - thx nielka98 * Waveform toolbar wraps to a second row instead of clipping - thx JonEvanCook * Space always types in text boxes, even with single-letter shortcuts allowed - thx kadrimarzouki * Pastel theme: refresh palette * Fix auto-translate "No usable translation" with KoboldCpp (llama.cpp advanced engine) - thx Giorgio800 * Fix edit text box scrollbar covering the text - thx kadrimarzouki * Fix invisible toolbar icons in the Pastel theme * Fix misspelled entries in the Spanish SE word list * Update Japanese translation - thx hisui3393 * Update Polish translation - thx potplayer-fanpack ----------------------------------------------------------------------------------------------------- v5.3.0-beta5 (18th of September 2026) * Add "Generate background music" (ACE-Step, local) in Video > More and in the text to speech window * Add "Delete selection (everywhere)" shortcut - thx ruonghieu * Add burn-in of subtitles into half side-by-side / top-bottom 3D video * Add 3D subtitle export at the depth from a 3D Blu-ray's 3D-Plane (.ofs) file * Add "Search rules..." box to Fix common errors - thx zcraber * Add search icon and clear button to all search/filter boxes * Add FFmpeg parameter editing and M4A input to Remux - thx pdjdev * Add reorder, undelete, context menu, file sizes and video drop to "Add/remove embedded subtitles" * Second subtitle dialog: remember style, optionally skip the dialog - thx kadrimarzouki, muaz978 * Frame mode: snap re-timed lines to frames automatically - thx Ingo * Split at text cursor: cursor at the end now splits at the line break - thx hisui3393 * Return focus to the opening control when a dialog closes - thx Shaima-Almarzooqi * Settings: Ctrl+PageUp/Down also works with focus on a setting - thx Shaima-Almarzooqi * Video OCR: "Test current frame" result readable by screen readers - thx Shaima-Almarzooqi * Fix Perplexity translate always returning no translation * Update audio.cpp runtime to v0.8.0 * Update Italian translation - thx bovirus * Update Korean translation - thx 12si27 ----------------------------------------------------------------------------------------------------- v5.3.0-beta4 (17th of September 2026) * Add "Translate in place" option and a model drop-down to auto-translate - thx fointypinger * Add 22 missing languages (Malayalam, Tamil, Telugu, ...) to the AI auto-translate engines - thx zcraber * Allow frame-based values in Bridge gaps - thx Ina-Ch * Use macOS-standard default shortcuts on macOS (Replace, Find next/previous, Go to line) - thx shotfirer, GitGianluc * Auto-translate selected lines with only one subtitle open now makes the subtitle the original - thx fointypinger * Settings: show one category at a time, Ctrl+PageUp/Down switches category - thx Shaima-Almarzooqi * Screen readers: readable names for list rows and combo box values - thx Shaima-Almarzooqi * Keep Lambda Cap ruby/tate-chu-yoko and vertical columns in Netflix IMSC 1.1 Japanese export * Fix freeze/100% CPU from mpv's clipboard thread on Linux Wayland - thx zbcoding * Fix auto-translate hanging on Gemini replies without text - thx fointypinger * Fix auto-translate target language opening on German for short lines - thx fointypinger * Fix video position when reopening a file with a video offset * Fix DVB teletext export not including the video offset * Fix actor rename not being undoable * Fix burn-in AMF quality falling back to blank for settings stored by older versions * Fix mirrored Latin words and digits in RTL lines in the fast waveform renderer * Fix "click to generate waveform" hint disappearing when switching waveform renderer ----------------------------------------------------------------------------------------------------- v5.3.0-beta3 (16th of September 2026) * Add "Guess start and end time from waveform" shortcut - thx m0ck69 * Add optional "move lines" button groups to the waveform toolbar - thx gambar3 * Add 3D image export (half side-by-side / top-bottom) from SE 4, also in seconv * Add an NVENC tune list to burn-in (hq/ll/ull/lossless) * Add experimental fast waveform renderer (SkiaSharp) - Options > Settings > Waveform * Add "Auto translate" for selected lines with only one subtitle open - thx fointypinger * Ask before auto-opening a video that is stored online only - thx FredOldenburg * Settings: name file type check boxes, download buttons and favorites list buttons - thx Shaima-Almarzooqi * Settings: picking a category moves focus into its section - thx Shaima-Almarzooqi * Fix auto-translate target language defaulting to Abkhaz instead of the last used or UI language - thx RealRUFUS * Fix freeze when opening a subtitle next to an online-only Dropbox video - thx FredOldenburg * Fix waveform jump and the cursor running a frame ahead when playback starts after a seek - thx mariajuferreira1958-creator * Fix Video OCR dropping everything after the 4th line of a frame - thx MengDaGBoss * Fix burn-in failing with the NVENC presets ffmpeg 9 no longer accepts - thx bichitoxxx * Fix the edit box "Hide" time code ignoring the video offset * Fix a remembered video offset being lost from the recent-file entry * Fix subtitle preview margin not allowing 0 * Update Crisp ASR to v0.8.33 * Update Italian translation - thx bovirus * Update Japanese translation - thx hisui3393 ----------------------------------------------------------------------------------------------------- v5.3.0-beta2 (15th of September 2026) * Add API Route auto-translate engine - thx DennyHo0917 * Add "Rename..." to the subtitle grid's Actors context menu - thx ghostminhtoan * Add "all lines" and custom-milliseconds move shortcuts - thx gambar3 * Add "OpenAI Compatible API" engine to batch convert auto-translate * Add Apple Vision OCR engine to seconv (--ocr-engine:applevision) - thx Hansie9999 * Video offset: save the grid's time codes to the file, like SE 4 * Accessible names for content buttons and text-less check boxes - thx Shaima-Almarzooqi * Fix Google Vision OCR splitting words on large fonts and adding a space before French "." - thx Codling * Fix OCR reading uppercase "I" as lowercase "l" in all-caps words (MlSSlSSlPPl) - thx Codling * Fix nOCR punctuation voting a line italic - thx Codling * Fix playhead hidden for a few seconds after seeking to where the video already is - thx Surlime * Fix waveform playhead not drawn at 0 s (e.g. after Stop) * Fix auto-translate moving a closing guillemet or curly quote to the next line - thx gambar3 * Fix text-navigation shortcuts not firing in RTL text when opted in - thx kadrimarzouki * Fix Export custom text format writing the original time codes and text instead of the edited ones * Fix time up/down controls drawn past the edit section with End time or Layer shown * Fix eight FFmpeg video player teardown, seek and clock defects * Fix merge of roll-up captions creating zero-length lines and losing lines * Fix Restore auto-backup wiping all settings when a settings backup is not valid * Fix Remux video leaving ffmpeg running when the window closes, and metadata escaping * Fix batch burn-in skipping a row again after a subtitle was picked for it * Fix Move captions showing a 0 px bar height for the 2.39:1 preset * Fix a plugin missing from the Plugins menu when two plugins share a shortcut action name * Fix "_Plugins" group label in Options > Shortcuts * Fix voice pack install progress not updating on the UI thread * Update CrispEmbed to v0.17.11 * Update whisper.cpp to v1.9.4 * Update Italian translation - thx bovirus * Update Japanese translation - thx hisui3393 ----------------------------------------------------------------------------------------------------- v5.3.0-beta1 (14th of September 2026) * Add FFmpeg video player (FFmpeg.AutoGen) with VideoToolbox (macOS) and D3D11VA/DXVA2 (Windows) hardware decoding * Add "Remux video" (Video > More) for video, multi-track audio and soft subtitles - thx ghostminhtoan * Add Text to speech voice manager: waveform, transcripts, copy voices between engines and voice packs * Add "Rename voice..." to the Text to speech voice combo box context menu * Add Voikko (libvoikko) backend for Finnish spell checking, bundled in Flatpak and macOS builds * Add layouts 12 and 13 with the text box below the video player - thx tuming2025 * Add "Font scale (%)" setting next to UI scale - thx tuming2025 * Add auto-backup of Settings.json with a Settings tab in "Restore auto-backup" * Add shortcuts to move selected lines X ms back/forward - thx gambar3 * Add per-plugin shortcuts in Options > Shortcuts (like SE 4) * Add "Do not ask again" to the plugin apply-to-which-lines prompt - thx fointypinger * Add "Show formatting, keep non-visual tags" subtitle grid mode * Add "Adjust durations...", "Apply duration limits..." and "Change formatting..." to the "Selected lines" context menu * Add settings UI for the hidden rc-stage settings (text-navigation shortcuts, Google Cloud speech to text) - thx kadrimarzouki * Add "use only front center channel" setting for 5.1 audio - thx willian-as * Add calculator button (frames to milliseconds) in "Apply minimum gap" - thx kadrimarzouki * Add configurable spell check unknown-word highlight color - thx mnerec * Add SE 4 EBU STL "Box" toggle to the grid and text box context menus, also drawn in the video preview * Add Transport Stream output settings to batch convert (SE 4 "TS settings...") * Add "Add folder recursive..." to batch convert * Add "Move captions" and "Remove fade in/out" to the image-based edit (BDSup2Sub features) * Add engine settings/info dialog for every video OCR engine * Add merge of roll-up (scrolling) captions to "Merge lines with same text" * AI review + Auto-translate: edit the suggested text in place in the grid * AI review: show the previous and next line under the selected suggestion - thx fraternl * Auto-translate: keep DeepL's sentence split on the row boundary - thx fraternl * Netflix check: allow applying the "Two lines maximum" fix - thx gambar3 * Settings: redesign toolbar settings as an icon toggle grid - thx pdjdev * Accessible names for inputs, lists and grid rows in all tool windows - thx Shaima-Almarzooqi * Wire the help shortcut in 110 more dialogs * Video player: show the GL renderer in the badge and log software-rendering fallback - thx zbcoding * macOS: new macOS-style rounded app icon - thx nitrodox * Improve Turkish (Windows-1254) encoding auto-detection - thx feuer133 * Faster language auto-detection (single tokenizing pass) * Source-generated LibraryImport and async file I/O in view models - thx ivandrofly * Fix edit box growing when a long text is pasted - thx kadrimarzouki * Fix single-line-length labels pushing the main view layout on long texts - thx kadrimarzouki * Fix split line losing ASSA tags and keeping the dialog dash on tagged lines - thx VaLvOnAuTa1981 * Fix AI review prompt editor not scrollable, so long prompts could not be saved - thx Morv55555 * Fix burn-in failing in ffmpeg when generating a video without subtitles - thx zbcoding * Fix spell check flagging valid Dutch words (WeCantSpell COMPOUNDRULE comment bug) - thx FredOldenburg * Fix nOCR double-detecting a glyph via expanded match on neighbour ink - thx Lukewarm1141 * Fix OCR replace lists turning "toi" into "tol" (trailing i->l rule needs a 3-letter stem) - thx fraternl * Fix waveform min-gap clamp landing one frame early with snap to frames * Fix EBU STL teletext line width (37) being applied in the grid for open subtitling files * Fix open-subtitling EBU STL preview using a different row count than the writer * Fix source view selection with extra original rows - thx pdjdev * Fix missing gap between the text and original text boxes * Track pickers: show subtitle count with thousands separator * Docs: bring the user documentation up to date with 5.2, explain start-time sorting in Join subtitles * Update Croatian OCR fix replace list - thx diomed * Update Italian translation - thx bovirus * Update Korean translation - thx 12si27 * Update Turkish translation - thx bilimiyorum ----------------------------------------------------------------------------------------------------- v5.2.0 (10th of September 2026) Subtitle Edit 5.2 collects the work of 32 betas and 7 release candidates. A summary of the changes since v5.1.0 (full details in the pre-release entries below): New features: * Assisted split and Assisted move - ranked one-click suggestions with full previews * Chapter editor with Matroska/MP4/OGM chapter formats (Video > More > Chapters) * Edit original - edit the original reference rows in place, open a non-matching original as a read-only reference, and show the original subtitle on the waveform * Forced narrative lines - a "Forced" column and "Save forced lines as..." * Error list with summary cards (Tools > List errors..., batch convert) - export to clipboard, text, Excel or web page * Statistics dashboard with tiles, meters, checks and a CPS histogram * Subtitle grid: "Columns..." dialog to reorder/hide columns, "Hide tags" formatting mode, and type-to-search in all combo boxes * Text boxes: drag-and-drop text (e.g. original to working text), configurable "Search via" shortcuts, "Google it", macOS "Look up", "Sentence case" and more "Surround with" slots * Waveform: guess start/end time from the waveform, snap to shot changes, editable video position box, per-track audio picker, and copy/paste at the video/waveform position * Beautify time codes using the video's real frame times, and in batch convert * Video offset remembers recent offsets; "Open recent video" and "Go to video position..." in the Video menu * Multiple replace: import/export rule categories, select all/none/invert, move shortcuts and regex match timeouts * Image-based subtitle editor: open DVD sup, XSUB, MP4 VobSub and more, "Video resolution..." scaling, and burn in Blu-ray sup * Update check settings with a stable/beta channel and a startup notification * Default save location, auto-break "do not break after" lists with an editor, and cut the subtitle with "Cut video" SE 4 parity: * Train nOCR, the minimum gap frame rate calculator, and "Remove/replace Unicode characters" (SE 4 plugin port) * WebVTT style manager, voices and browser preview, and configurable WebVTT cue settings * Import an SE 4 Settings.xml, and 22 more SE 4 shortcuts on shortcut import * Shortcuts: go to next empty line, go to first/last line, bookmarks, underline, toggle custom tags, recalculate duration, go to next/prev subtitle (play translate), move first word to previous subtitle, break at first space from cursor * Layout 10 (edit box under the waveform), "Center text in subtitle grid", the four grid double-click actions, and a toolbar frame rate that works like SE 4's * "Set up like Subtitle Edit 4" also arranges the waveform toolbar; "Open second subtitle file..." is back in SE 4's Video menu spot * Remove blank lines when opening a subtitle, full frame image export, regex snippet context menu in Find/replace, and Alt+F/Alt+R in Replace * Tesseract OCR: SE 4's 10 px margin and resize retry passes * Copy-to-clipboard grid menu items and plain text column paste are back Auto-translate and AI: * New "llama.cpp advanced" and "Ollama advanced" engines with batch context, system prompt and schema-forced output - also in batch convert and seconv (--translate-prompt) * MiLMMT-46 translation models; DeepL offers every API language; Google Translate retries via a fallback endpoint when the free service is blocked * Re-break translated rows that break the line profile, keep the chosen languages when the engine changes, and the original text is no longer lost after translating * AI review: in the grid context menu, apply suggestions in passes, Play current, start/stop the llama.cpp server, delay between requests, and new models (Gemma 4 E2B, EuroLLM, Granite 4.1) Speech to text: * New engines and models: WhisperX (standalone, no Python), Google Cloud Speech-to-Text v2, Voxtral (Crisp ASR) and anime-whisper for Japanese * Transcription quality report with non-speech and repeated-line removal * Crisp ASR: "Auto detect" language on all backends, a VAD off switch and a retry without VAD; live progress for WhisperX * The MLX Whisper engine was removed Text to speech: * New engines: IndexTTS 2.5, Higgs Audio v3, Fish Audio S2 Pro and FireRedTTS3 (audio.cpp), dots.tts, Confucius4-TTS and Pocket TTS (Crisp ASR), plus VibeVoice again, CosyVoice3 RL models and multilingual Chatterbox V3 (23 languages) * Voice cloning from the video - "Find voices in video and clone them all", and per-line cloning on Qwen3, audio.cpp, VibeVoice, MOSS-TTS, CosyVoice3 and VoxCPM2 - with a consent prompt before the first clone * Speaker-name detection, "leave sound/music lines silent", a Zonos language picker and custom Piper voice models OCR: * New engines and models: Apple Vision OCR (macOS), Paddle OCR 3.7 (PP-OCRv6), CrispEmbed PP-OCRv6 and DeepSeek-OCR-2, and llama.cpp HunyuanOCR 1.5, LFM2.5-VL 3B and custom vision models * Video OCR (burned-in subtitles): far more subtitles read, better text and start times, CrispEmbed engine, OCR fix engine and spell check coloring * OCR fix engine: new Spanish/Portuguese/Italian and Cyrillic rules, around 115 repaired replace list entries, and apostrophe straightening * Auto-detect the OCR language, open the matching video after OCR, "Save all images with HTML index", and faster nOCR matching Subtitle formats: * New: EBU-TT (Tech 3350), Manzanita DVB teletext (.dvbttx), Csv Excel, CANVASs SSTG1 (.sdb), Wistia json, DVD Junior SPC, Sonic DVD Producer, YouTube srv3 (.ytt), DaVinci Resolve marker EDL and Adobe Premiere markers * New exports: Final Cut Pro Xml Captions, BDN/xml 8-bit and Audacity/Tenacity labels * Read subtitles from fragmented MP4 (DASH/CMAF), ARIB STD-B24 captions from transport streams, and SMPTE-TT bitmap captions * Much better import of unknown formats (generic XML importer), "Import plain text" from the unknown format prompt, and spreadsheets from File > Open * EBU STL/teletext: color picker, alignment dialog, TT column, video preview with box/justification/double height, and frame-based time codes * SCC: colors as CEA-608 mid-row codes, a line length warning, and frame-accurate import timing * ASSA: embed and trim fonts, offer resampling on a resolution mismatch, new advanced effects, and "Change ASSA style properties" in batch convert * Ruby and emphasis in Lambda Cap and Netflix IMSC 1.1 Japanese; font colors in EBU-TT-D and IMSC Rosetta Video and image export: * Image-based export: advanced text effects (gradient, neon glow, 3D extrude), per-line ASSA colors and outlines, inline font face/size, and Blu-ray sup fades and overlapping subtitles * Burn-in: VideoToolbox hardware encoders on macOS, .webm/.ts output, letter spacing and libass style previews * Sync dialogs: subtitle on the video, resizable waveform, and the selected audio track * Smoother waveform/video cursor with mpv, faster scrubbing on long-GOP video, and fixed audio dropouts when pausing or seeking Batch convert and seconv: * Batch convert: Beautify time codes, convert colors to dialog, snap time codes to frames, "Add folder...", translation progress, keep the source file date/time, and every teletext page per PID * seconv: --json output and --help-json, --ocr-prompt, --translate-prompt, --override-position, --output-filename-append, full frame image export, .avi/XSUB reading and batched Paddle OCR Translations and engines: * New UI languages: Central Kurdish and Azerbaijani - and all language files synced with English * Updated engines: Crisp ASR v0.8.32, llama.cpp b10840, whisper.cpp v1.9.3, CrispEmbed v0.17.9, Paddle OCR 3.7, Tesseract 5.5.3, ffmpeg 9.0.1 (Windows), libmpv 20260814 (Windows) and yt-dlp 2026.08.19 A big thank you to everyone who tested the pre-releases, reported issues and contributed code and translations - and a special thanks to Anthropic for sponsoring a Claude subscription used in the development of this release :) ----------------------------------------------------------------------------------------------------- v5.2.0-rc7 (9th of September 2026) * Add CANVASs SSTG1 (.sdb) subtitle import - thx magsmike * Add opt-in setting to let shortcuts fire on Ctrl+Left/Right and Home/End in a text box - thx kadrimarzouki * Merge short lines: add an "Apply" checkbox column to pick which lines to merge - thx ghostminhtoan * Auto-translate: re-break translated rows that break the line profile - thx ruudhaf02-lgtm * Fix SCC import timing - apply control codes at their frame within the line - thx magsmike * Fix grid paste overlapping the selected line in time - thx rosilucia-hub * Fix Blu-ray sup save from the binary editor dropping the first line and duplicating the last - thx Lukewarm1141 * Fix video preview not refreshing when right-to-left mode is toggled - thx kadrimarzouki * Fix repeated shortcuts while holding modifiers, and restore focus after undo/redo - thx ghostminhtoan * Export image: right/center justify Arabic lines precisely - thx kadrimarzouki * OCR: keep a row selected after deleting lines - thx FLAV1N * OCR fix engine: apostrophe straightening honors "use hardcoded rules" * Tesseract OCR: match retry-merge unknown words whole, not as substrings * Qwen3 ASR CPP: opening quotes and dialog dashes start the next cue * Waveform original overlay: keep a long cue overlapped by a shorter one visible * seconv: don't parse input paths in the parameter table as markup - thx TWiStErRob * Add Azerbaijani translation - thx jamalkamaladdin * Update German translation - thx FunkyKnilch * Update Italian translation - thx bovirus * Update Japanese translation - thx hisui3393 * Update Korean translation - thx 12si27/sinfancy * Update Portuguese (Brazil) translation - thx igorruckert * Update Turkish translation - thx bilimiyorum ----------------------------------------------------------------------------------------------------- v5.2.0-rc6 (8th of September 2026) * Add option to show the original subtitle on the waveform - thx pdjdev * Add "remember window size and position" to the Matroska/MP4/transport stream/VobSub track pickers * Add --override-position and --output-filename-append to seconv * Tesseract OCR: add SE 4's 10 px margin and resize retry passes, and merge the retry word-wise - thx andor1999 * OCR fix engine: straighten typographic apostrophes, keep "didn't" as one word, and capitalize a line after a finished sentence * Compare: keep the word-level diff marking in the HTML export * Batch convert: convert every teletext page per PID in transport streams * seconv: use the GUI's PaddleOCR install and add a run timeout - thx tekk42 * FireRedTTS3: require the reference transcript instead of generating noise - thx invio-a11y * Fix "Undo" unexpectedly removing the original subtitle - thx Surlime * Fix subtitle numbering after batch OCR - thx tekk42 * Fix Qwen3 ASR CPP cue segmentation (break on sentence punctuation, no spaces in CJK) - thx tuming2025/TheOriginalNedKelly * Fix memory leaks in video preview dialogs, row handlers, spectrogram and auto-backup * Faster subtitle format detection, "convert colors to dialog", CJK character counts and more * Update llama.cpp to b10840 * Update Bulgarian translation - thx jekovcar * Update Central Kurdish translation - thx dkakaie * Update Chinese (Traditional) translation - thx love80312 * Update German translation - thx FunkyKnilch * Update Hungarian translation - thx Zityi * Update Japanese translation - thx hisui3393 * Update Russian translation - thx jekovcar ----------------------------------------------------------------------------------------------------- v5.2.0-rc5 (7th of September 2026) * Add FireRedTTS3-Base as an audio.cpp text to speech engine (zero-shot voice cloning in 24 languages) * Add "Import plain text" to the unknown subtitle format prompt, and open plain .txt files in the importer paired with a matching audio file - thx NiggleHub * Toggle dialog dashes: only the focused text box, per-column direction, and no stacked dashes - thx kadrimarzouki * Transcription: append the language code to the unsaved file name when "Save as: append language code" is on - thx GrampaWildWilly * Text to speech: per-line voice cloning from the video no longer echoes the reference clip (Fish Audio S2 Pro, CosyVoice3) - thx invio-a11y * Text to speech: Confucius4-TTS switches language without reloading the model * audio.cpp: offer a runtime update when the installed build lacks the engine's model family * Fix "guess start from waveform" moving the whole line instead of trimming the start on long text - thx m0ck69 * Fix "guess start/end from waveform" giving up silently, staying put when the cue is more than 1 s off, or stopping at the first matching boundary - thx muaz978 * Fix the undocked audio visualizer staying behind the main window after "Open original" - thx GrampaWildWilly * Fix a closed "New window" editor keeping its timers and undo polling running * Fix 46 bugs found in random-file hunts (subtitle format round-trips and detection, CSV/TSV quoting, ARIB B24 tables, ASSA style usages, SE 4 shortcut import, batch convert min gap) * Faster Timed Text/Rosetta/Netflix Japanese reading, TTML saving, "merge lines with same text", statistics, pre-translation and name list lookups * Update Italian translation - thx bovirus * Update Korean translation - thx 12si27 ----------------------------------------------------------------------------------------------------- v5.2.0-rc4 (6th of September 2026) * Add Google Cloud Speech-to-Text v2 as an online speech-to-text engine (lossless audio, dynamic batching, word-timed cues) - thx muaz978 * Add "Video resolution..." to the image-based subtitle editor (Tools menu), scaling images and positions like BDSup2Sub * Add per-line voice cloning to VibeVoice, MOSS-TTS, CosyVoice3 and VoxCPM2 (Crisp ASR) - thx invio-a11y * Add Spanish (Mexico), (Argentina) and (Latin America / US) spell check dictionaries * Add keyboard navigation between changes in Beautify time codes * Add live progress, elapsed time and estimate for WhisperX speech-to-text * Show a drop-target bar while dragging text in the subtitle text boxes * Make the OK button the default in dialogs so Enter accepts them (e.g. Multiple replace) - thx rRobis * Fix short durations: also run after split/merge, and always extend the duration - thx andor1999 * Generate blank video: theme-aware, larger time code with a contrast box, and no window resizing while generating * Burn-in: smaller fixed-height preview player without empty bands around the video * Fix "guess start/end from waveform" not working on quiet audio - thx m0ck69 * Fix text drag-and-drop shrinking the selection, and space the drop like SE4 - thx ruonghieu * Fix undocked windows staying above other applications that took the foreground - thx cvrle77 * Fix Higgs Audio v3 voice clone hiss at the end of clips - thx invio-a11y * Remove the Vietnamese Parakeet model from Crisp ASR speech to text (quality too poor) - thx ruonghieu * Update Chinese (Simplified) translation - thx emsb8888-prog * Update Italian translation - thx bovirus ----------------------------------------------------------------------------------------------------- v5.2.0-rc3 (5th of September 2026) * Add text drag-and-drop in the subtitle text boxes (e.g. from original to working text) - thx bichitoxxx * Add "Move first word to previous subtitle" shortcut, and make the word-moving shortcuts follow the focused text box - thx saltarob * Add "Export..." to the transcription quality report - thx andor1999 * Add the Vietnamese Parakeet model to Crisp ASR speech to text - thx ruonghieu * Add LFM2.5-VL 3B to the llama.cpp OCR model list * Accept OpenRouter transcription models that reject verbose_json - thx muaz978 * Batch convert: hide the main window while the tool window is open - thx m0ck69 * Let a user-assigned Alt+Space shortcut beat the Windows system menu - thx kadrimarzouki * macOS: change the default "Open Subtitle Edit folder" shortcut to Ctrl+Option+Shift+Cmd+D - thx GitGianluc * Sync/Cut video: zoom the waveform on Shift + main-row plus/minus too - thx perkesfurmah-collab * Text to speech: per-line voice clone pads short references, and audio.cpp engines can share a models folder - thx invio-a11y * Fix waveform cursor and time display freezing during playback, and play starting away from the cursor - thx Davanix * Fix undo/redo scrolling the current subtitle to the top of the grid - thx saltarob * Fix auto-translate rows shifting when an abbreviation period is added to a merged row - thx claudemartin * Fix "Show Plugins menu" not toggling the native macOS Plugins menu - thx GitGianluc * Fix OCR fix word list rejecting case variants of an existing word - thx Pemicope * Fix Generate button unreachable in burn-in/transparent video dialogs on short screens - thx sukuna0322 * Fix audio.cpp TTS server not restarting after an engine update * Update whisper.cpp to v1.9.3 * Update Crisp ASR to v0.8.32 * Update audio.cpp TTS runtime (Higgs Audio v3 end-of-clip hiss fix) * Update French translation - thx Need74 * Update Italian translation - thx bovirus * Update Japanese translation - thx hisui3393 * Update Portuguese (Brazil) translation - thx igorruckert * Update Turkish translation - thx bilimiyorum ----------------------------------------------------------------------------------------------------- v5.2.0-rc2 (4th of September 2026) * Add waveform vertical zoom shortcuts to the Point sync and Visual sync dialogs - thx perkesfurmah-collab * Add "Auto detect" language to the remaining Crisp ASR backends (Ark, Canary, Cohere, FireRed, Fun-ASR Nano, Granite, Kyutai) - thx subof * Add "Guess end time from waveform", and offsets for both guess commands - thx m0ck69 * Add the subtitle type (normal / hearing impaired) to the DVB teletext export * Burn in Blu-ray sup subtitles from the image-based editor and in batch convert - thx josemorgado107 * Export image: honor inline and per segment - thx Victory61 * Second subtitle: follow the preview font and add a Bold option - thx GrampaWildWilly * Keep "Open second subtitle file..." available while a second subtitle is shown - thx GrampaWildWilly * Text to speech: judge silence relative to the clip's peak so quiet voice clones keep their last word - thx invio-a11y * Fix "guess start time from waveform" only firing while the waveform had focus - thx m0ck69 * Fix the subtitle grid churning and slowing down when merging with a read-only original - thx torvchen * Fix the grid selection and view jumping after merging lines with the same text / time codes * Fix an extra character from DVD OCR when the leftmost glyph has a descender - thx FunkyKnilch * Update Catalan translation - thx jmontane * Update Italian translation - thx bovirus * Update Japanese translation - thx hisui3393 * Update Korean translation - thx 12si27 ----------------------------------------------------------------------------------------------------- v5.2.0-rc1 (3rd of September 2026) * Add "Open recent video" to the Video menu - thx ghostminhtoan * Add "Break at first space from cursor position" text box shortcuts - thx saltarob * Text to speech: per-line voice cloning on the audio.cpp engines (IndexTTS 2.5, Higgs Audio v3, Fish Audio S2 Pro) - thx invio-a11y * Text to speech: a language picker for Zonos (CrispASR) - thx subof * Ask before the first voice clone in Confucius4-TTS and Pocket TTS too * Blu-ray sup export: render overlapping subtitles together instead of dropping one - thx josemorgado107 * Sync dialogs: make the waveform under the video drag-resizable - thx perkesfurmah-collab * Waveform: offer the ffmpeg download instead of failing silently when ffmpeg is missing - thx Marethyu2000 * Source view: use the appearance font and readable light-mode syntax colors - thx GrampaWildWilly * JSON syntax highlighting: color string values that follow a property name * Update llama.cpp to b10760 * Fix the waveform cursor landing a GOP away and hopping a frame forward on click and mouse wheel - thx bichitoxxx * Fix "Set position" previewing rotation differently from libass - thx bichitoxxx * Fix split and merge not keeping an editable original in sync - thx saltarob * Fix closing Settings jumping back to the first subtitle - thx saltarob * Fix recalculate duration landing just over max CPS, and the CPS readouts disagreeing - thx saltarob * Fix a plugin result dropping the original column - thx fraternl * Fix the spell check staying on the old language after a translation - thx fraternl * Fix the video player not opening files on Windows paths longer than MAX_PATH - thx HindiAnimeVerse * Fix the beautify time codes profile editor clipping its numeric fields - thx matmaggi * Fix Compare lining the two grids up before the mirrored selection had scrolled * Fix "set end time and go to next" firing on key down instead of key up - thx ghostminhtoan * Fix the SMPTE preview stretch not reaching the VLC reloader and the secondary subtitle - thx ghostminhtoan * Fix the theme not following OS light/dark changes after the first switch * Fix 16 bugs found in a random-file hunt (subtitle formats, Google Lens, speech to text, TMPGEnc) * Minor performance improvements from immutable brushes and pens for every static color * Update French translation - thx Need74 * Update Italian translation - thx bovirus * Update Japanese translation - thx hisui3393 * Update Portuguese (Brazil) translation - thx igorruckert ----------------------------------------------------------------------------------------------------- v5.2.0-beta32 (2nd of September 2026) * Add a "Columns..." dialog to reorder and hide subtitle grid columns - thx m0ck69 * Export the "List errors" report to clipboard, text file, Excel or web page - thx AuroraMartell * Binary edit: open DVD sup, XSUB (avi), MP4 VobSub, image based WebVTT/TTML and more transport streams, and export DVD sup and D-Cinema * ASSA: offer to resample the subtitle when the video resolution differs from the script resolution - thx hisui3393 * Add Confucius4-TTS (CrispASR), Higgs Audio v3 and Fish Audio S2 Pro (audio.cpp) voice cloning TTS engines * "Point sync via other subtitle": a gap column that highlights likely sync points - thx alexchexes * Move "Open second subtitle file..." to SE 4's spot in the Video menu - thx m0ck69 * macOS: open Finder files in a new window, and start the video picker in the subtitle's own folder - thx muaz978 * seconv: add the full frame image export option - thx smt-joen * Faster waveform scrubbing on long-GOP video - thx GrampaWildWilly * Faster Sami and IMSC 1.1 reading, PAC/EBU STL/D-Cinema saving and Blu-ray image export * Fix audio dropouts when pausing or seeking - the beta 31 mpv option did nothing - thx GrampaWildWilly * Fix Alt+Tab still not bringing undocked windows to the foreground - thx GrampaWildWilly * Fix the playhead jumping when repeat playback wraps to the next line - thx marcpbailey * Fix the "Guess time codes" window losing its buttons at a high UI scale * Update Italian translation - thx bovirus ----------------------------------------------------------------------------------------------------- v5.2.0-beta31 (1st of September 2026) * New subtitle format: Csv Excel - thx matmaggi * Mark lines as forced narrative, with a "Forced" column and "Save forced lines as..." - thx matmaggi * New subtitle format: EBU-TT (Tech 3350), carrying teletext colors, boxing, rows and the GSI metadata * EBU-TT-D and IMSC Rosetta: read and write font colors - thx Scryper * DVB teletext (.dvbttx): write Level 2.5 colors, and make it a toolbar format with alignment picker, video preview and batch convert - thx René * Add Pocket TTS (CrispASR) voice cloning, and offer VibeVoice again * CrispASR: update to v0.8.31, and offer Windows CUDA 13 * Subtitle image previews: a configurable background color, and a Blu-ray sup overlay that scales with the video - thx Lukewarm1141 * Advanced TTS settings: hint icons instead of a wall of text, so the window fits the screen again - thx nielka98 * Fix audio dropouts by keeping the mpv audio device open while paused - thx GrampaWildWilly * Fix Alt+Tab not bringing undocked windows to the foreground - thx GrampaWildWilly * Fix playing selected lines in repeat mode replaying a line at the cue boundary - thx marcpbailey * Fix the subtitle grid jumping to the end of the list after cut + auto-break - thx St0rmXtr00per * Fix the Matroska track picker opening twice on macOS - thx marcpbailey * Update Italian translation - thx bovirus * Update Japanese translation - thx hisui3393 * Update Korean translation - thx 12si27 * Update Turkish translation - thx bilimiyorum ----------------------------------------------------------------------------------------------------- v5.2.0-beta30 (31st of August 2026) * Read Wistia json, DVD Junior SPC, Sonic DVD Producer and YouTube srv3 (.ytt) subtitles * Add "Center text in subtitle grid" (SE 4 parity) - thx matmaggi * Settings import: accept an SE 4 Settings.xml - thx JDario16 * Teletext: write G2 characters like the music note, and read the national option sub-set and the Level 2.5 colours - thx René * Make dialogs keyboard-navigable: initial focus and Tab, across the UI - thx GrampaWildWilly * ASSA style window: scale the preview with the window and give it the whole panel * Auto-translate: make Enter press the default button * Keep ASSA override blocks out of auto-translate too - thx MrGoraj * Sort the OCR fix list alphabetically in Options - Word lists - thx Anth70 * seconv: give Ollama OCR the same generation limits as the GUI - thx gmariani * Fix seconv reporting its version as 5.0.0 - thx gmariani * Fix the mouse wheel setting the video position with the option turned off - thx rookes * Fix Compare's ignore options, and remember its settings - thx FunkyKnilch * Fix the video preview showing EBU STL (and TTML positions) after the format was changed * Fix 20 bugs found in a code sweep * Update French translation - thx Need74 * Update Italian translation - thx bovirus * Update Japanese translation - thx hisui3393 ----------------------------------------------------------------------------------------------------- v5.2.0-beta29 (30th of August 2026) * Add configurable "Search via X" shortcuts, with a "Search via" submenu in the text box context menu - thx kadrimarzouki * Add "toggle translation and original in video/audio preview" shortcut - thx Surlime * Add shortcut to copy the text of the selected lines to the original - thx hisui3393 * Add macOS "Look up" to the subtitle text box context menu - thx routineCode * Video offset: remember recently used offsets, and an "Apply" that keeps the window open - thx kadrimarzouki * EBU STL: a font picker with a preview sample in the save options, color shortcuts snapping to the nearest teletext color, and clearer labels in the teletext alignment dialog * Waveform: give right-to-left text the direction its letters ask for - thx kadrimarzouki * seconv: per-image OCR progress, and percentages on OCR and translate - thx gmariani * Fix the waveform cursor drifting off the frame when stepping forward - thx rookes * Fix "one frame forward/back (with play)" not stepping smoothly - thx rookes * Fix waveform seek drift at SMPTE frame rates, unpinned nudges and a mixed-up blip clock - thx Davanix * Fix the mpv seek target being dropped after a click in the waveform - thx Davanix * Fix Compare no longer scrolling the two grids together - thx Davanix * Fix the subtitle grid not centering on the last prev/next-line paths - thx St0rmXtr00per * Fix the blank waveform and the rest of the rewinding after leaving Settings - thx GrampaWildWilly * Fix .dvbttx files with a preamble over 200 KB reading back as empty * Fix a translation being re-split inside a number - thx fraternl * Fix WebVTT cues placed in different vertical bands being merged - thx MrAnter * Fix six more bugs found in a code sweep * Update Bulgarian translation - thx jekovcar * Update Italian translation - thx bovirus * Update Japanese translation - thx hisui3393 * Update Persian translation - thx rmtjokar * Update Russian translation - thx jekovcar ----------------------------------------------------------------------------------------------------- v5.2.0-beta28 (29th of August 2026) * Read and write Manzanita DVB teletext files (.dvbttx) - thx René * EBU STL video preview: draw the box, the justification, double height rows and an optional custom font * Surround with: more slots, and fix the slots getting lost in settings export/import - thx kadrimarzouki * SCC: write colors as CEA-608 mid-row codes instead of putting the color tags on screen - thx Jessecar96 * Video OCR: a better default prompt for two-line subtitles, and "--ocr-prompt" for seconv - thx tekk42 * Update CrispASR to v0.8.30 * Fix the subtitle grid not centering on the selected row during prev/next navigation - thx St0rmXtr00per * Fix "inverse selection" selecting a single row and jumping to the top of the file * Fix the video rewinding after leaving Settings - thx GrampaWildWilly * Fix the duration field in frame mode not accepting a value typed without the colon - thx René * Blu-ray sup: fix the export of bitmaps over 64 KB * Update Italian translation - thx bovirus * Update Japanese translation - thx hisui3393 ----------------------------------------------------------------------------------------------------- v5.2.0-beta27 (28th of August 2026) * Add "Assisted split" and "Assisted move" windows, with ranked one-click suggestions and full previews - thx GitGianluc * Add "Justify lines" for the mpv video preview in Options - Settings - Video - thx kadrimarzouki * Image based export: let the line justification follow the alignment tag - thx smt-joen * Much better import of unknown subtitle formats - a new generic XML importer plus a dozen JSON fixes * Netflix check: measure gaps, bridges and shot changes at the video's frame rate, and fix two "spell out numbers" defects * Fix "generate video with burned-in subtitles" silently doing nothing when the output file already existed - thx rRobis * Fix the waveform cursor jumping away a moment after a click - thx Davanix * Fix the undocked waveform window taking the foreground when another program is closed - thx GrampaWildWilly * Fix around 290 more bugs found in 13 code sweeps - subtitle formats, batch convert, seconv, dialogs, settings, binary edit and core helpers * Update Chinese (simplified) translation - thx emsb8888-prog * Update Italian translation - thx bovirus * Update Japanese translation - thx hisui3393 * Update Korean translation - thx 12si27 * Update Polish translation - thx potplayer-fanpack * Update Portuguese (Brazil) translation - thx igorruckert ----------------------------------------------------------------------------------------------------- v5.2.0-beta26 (27th of August 2026) * Video OCR: read far more subtitles, with much better text and start times - thx caiweihan * Video OCR: run the OCR fix engine and spell check coloring on the lines, plus an edit dialog and a context menu * Video OCR: order the engines best-measured first, add setting hints, a dictionary download button and window layout polish * Add "Change ASSA style properties" (spacing and alignment) to batch convert - thx adeltalon * Add "Play current" to spell check, so you can hear the line a flagged word is in - thx fraternl * Add "Google it" to the subtitle text box context menu * Add the SE 4 "go to next/prev subtitle (play translate)" shortcuts - thx kadrimarzouki * Add Central Kurdish translation - thx dkakaie * Spreadsheets (csv/xlsx/ods) can now be opened from File - open, and the import is documented - thx GrampaWildWilly * Lambda Cap: read and write ruby, emphasis dots and positioning control codes - thx magsmike * EBU STL: keep the colors when the header claims open subtitling * Split long lines: keep a sentence-ending line break as an event boundary - thx AuroraMartell, Triathlon-rally * "Fix RTL via Unicode" now wraps the sentence inside the tags, and closes the embedding * Netflix check and fix: the timing fixes (min/max duration, gaps, shot changes) are now actually applied * Faster XML format detection, Timed Text saving and DVB/VobSub image decoding * WSB is now import only - Subtitle Edit could not read back the files it wrote * Fix "apply minimum gap", "change formatting" and "sort by" clearing the subtitle when OK was pressed quickly * Fix the first save after a translation overwriting the source file's ASSA styles * Fix cancelling "save as" saving anyway, and "save original as" using the wrong format * Fix a crash when dropping an mkv file on the subtitle grid, and the doubled "parsing Matroska file" wait * Fix Arabic subtitle text showing up as boxes in the subtitle grid - thx adeltalon * Fix auto-translate blanking every untranslated line when pressing OK after a partial translation * Fix "play current and pause" running past the end of the line - thx kadrimarzouki * Fix Multiple replace losing the selection when moving a rule up or down - thx madsoft-ha * Fix the undocked waveform window coming to the front instead of the main window - thx GrampaWildWilly * Fix export/import of settings resetting video settings and rules, and wiping shortcuts * Fix "whole word" find and replace matching inside words * Fix spell check losing user words, names and "use always" entries * Fix speech to text losing word level highlighting, and cancel not stopping a batch * Fix text to speech review edits dropping italics and line breaks * Fix several burn-in and transparent video bugs (cut length, target file size, output properties) * Fix image based export duplicating text after a "<", and ASSA fade-outs exporting fully transparent * Fix five bugs in the csv/xlsx/ods importers (pipe separator, merged cells, sheet order) * Fix shot changes import losing precision, decimal commas and "frame_time" files * Fix CEA-608 roll-up captions scrolling the wrong rows * Fix around 200 more bugs found in six code sweeps - subtitle formats, dialogs, settings, OCR, downloads and core helpers * Update German translation - thx FunkyKnilch * Update Italian translation - thx bovirus * Update Polish translation - thx potplayer-fanpack * Update Turkish translation - thx bilimiyorum ----------------------------------------------------------------------------------------------------- v5.2.0-beta25 (26th of August 2026) * Add advanced text effects to image based export - gradient, neon glow, 3D extrude and more * Add speaker-name detection and a "leave sound/music lines silent" prompt to text to speech - thx subof * Add "Clone from video" to Qwen3 TTS, so each line can be dubbed in the voice heard in the video - thx invio-a11y * Add an editable video position box to the waveform toolbar - thx 95Midnight * Add custom llama.cpp vision models to the OCR model list * Add the "Ollama advanced" translate engine to batch convert * Add "--translate-prompt" to seconv for the local LLM translate engines - thx Makar8000 * Image based export now honors per-line ASSA colors and outline/shadow widths * EBU STL: show frame-based time codes while the format is active - thx Triathlon-rally * "Normalize strings" now folds curly quotes, dashes and ellipsis to plain ASCII - thx fraternl * DeepL: offer all the languages the API supports, and use the host that matches the key * Google Translate: use the right language list for the free and the paid engine, and drop two broken entries * Auto-translate: take the CrispASR MADLAD language list from the model - thx ecotycoon * OCR: list the most accurate llama.cpp models first * Italian OCR: fix "E'" to "È" - thx GitGianluc * Speech to text: say when Crisp ASR was killed instead of reporting no speech - thx marwanlhabti5-coder * Update llama.cpp to b10625 * Fix the original text being erased after translating - thx fraternl * Fix EBU STL export doing nothing when a font color tag has no quotes, and remember the save options * Fix regenerating a cloned line after importing a text to speech session - thx invio-a11y * Fix "find voices in video and clone" not setting actors when a subtitle is open - thx subof * Fix "convert actors" missing a speaker tag split across a line break - thx subof * Fix OCR removing the French space before "?" and "!" * Fix Microsoft Translator throwing an error when the language list cannot be fetched * Update Italian translation - thx bovirus * Update Japanese translation - thx hisui3393 * Update Polish translation - thx potplayer-fanpack ----------------------------------------------------------------------------------------------------- v5.2.0-beta24 (25th of August 2026) * Add dots.tts text to speech engine with voice cloning * Add WhisperX speech to text engine, standalone - no Python needed - thx muaz978 * Add Apple Vision OCR on macOS * Add a format properties button next to the subtitle format picker * Add "List errors..." to the Tools menu * Add the SE 4 bookmark, underline and "guess start" shortcuts * macOS: add the standard Window menu, close menu-bar gaps and fix the file types the app registers * macOS: let Ctrl+click open the shortcuts list's Import/Export menu * Blu-ray sup export can now fade lines in/out via ASSA "{\fad}" tags * Update Paddle OCR to 3.7 (PP-OCRv6) * Google Translate: show a clear message when Google blocks the free service, and retry via a fallback endpoint - thx kadrimarzouki * Speech to text advanced: the parameter buttons now toggle their parameter on and off - thx thealainpaul * Fix EBU STL save options being silently lost or corrupted * Fix subtitle times drifting by one millisecond while editing - thx Davanix * Fix the original text disappearing after applying a tool dialog's result - thx madsoft-ha * Fix "Open with" on an mkv sometimes opening only the subtitle, not the video - thx wixfigur * Fix "convert actors" not writing the actor column for inline "Name: text" lines - thx rotj * Fix Faster Whisper XXL, CTranslate2 and Const-me asking to install again every time - thx senaplz * Fix Qwen 3.5 leaking "thinking" text into llama.cpp translate and AI review * Fix Whisper suggesting int8 again after cuBLAS rejected the compute type * Update Russian translation - thx jekovcar * Update Bulgarian translation - thx jekovcar * Update Chinese Traditional translation - thx love80312 * Update Italian translation - thx bovirus ----------------------------------------------------------------------------------------------------- v5.2.0-beta23 (24th of August 2026) * OCR: add accent, l/I and preposition-split rules for Spanish, Portuguese and Italian * OCR: fix Latin letters and digits misread inside Cyrillic words (Russian, Macedonian) * OCR: repair about 115 broken, dead or duplicated entries in the fix replace lists * Faster Scenarist Closed Captions export (5x), Netflix quality checks and "remove text for hearing impaired" * Faster Blu-ray palette decoding, transport stream scanning and Cavena 890 reading * Fix ASSA attachments being written with the last two bytes zeroed * Fix DVD Studio Pro "with space" files failing to load * Fix csv lines containing a semicolon being dropped * Fix ASSA resampling giving up on tags with a space, like "\pos (10,11)" * Fix words per minute being reported as infinity for zero duration lines * Fix "convert actors" losing the name and mangling lines with leading spaces * Fix crashes and wrong results when parsing malformed MP4, transport stream and RIFF files * Fix WebVTT thumbnail sprite sheets in .jpeg being rejected * Fix video OCR ending a subtitle too early after an unreadable frame * Fix teletext line length errors missing from the error lists in the main window and batch convert * Fix beautify time codes using a cancelled or failed frame time extraction when the video duration is unknown * Fix the speech to text quality report naming the wrong line for removed repeats * Fix Netflix "spell out leading number" doing nothing outside Windows * Fix Unknown 33/34/59 writing 00:00:60 instead of 00:01:00 * Fix Matroska SSA lines with fewer commas than expected throwing an error * Fix plain text import hanging when the max line length setting is zero * Fix OpenRouter speech to text with GPT transcription models - thx muaz978 * Update German translation - thx FunkyKnilch * Update Japanese translation - thx hisui3393 * Update Korean translation - thx 12si27 ----------------------------------------------------------------------------------------------------- v5.2.0-beta22 (23rd of August 2026) * Beautify time codes can now use the video's real frame times instead of assuming a constant frame rate - thx warrentc3 * Add a transcription quality report to speech to text, plus non-speech and repeated-line removal - thx l0uev4-lab * Add an error list with summary cards and per-error rows, in the main window and in batch convert * Add shot-change snap distances to Settings * Snap to the nearest shot change in the waveform when a cue is close - thx m0ck69 * Warn before saving SCC when lines exceed 32 characters or 4 lines * Add a "Use external server" toggle for llama.cpp in batch convert, independent of the main setting - thx emsb8888-prog * Remember the batch convert "Remove formatting" check boxes - thx matake31 * Use the selected audio track in the sync windows - thx dmist * Use display dimensions for rotated video when burning in subtitles - thx thealainpaul * Speed up tag counting in the subtitle grid - thx ivandrofly * Fix split long lines producing overlong lines, gaps between split events and duplicate numbers - thx Triathlon-rally * Fix selecting a text to speech review line by clicking or grabbing its waveform block - thx cvrle77 * Fix Google Translate failing on transient server errors instead of retrying - thx JoaoDuarte-code * Fix the subtitle grid scroll position jumping while the reference projection rebuilds - thx torvchen * Fix the minimum gap not being kept when nudging a cue one frame - thx m0ck69 * Fix the AI review sending ASSA override blocks, and send actor/style as context - thx MrGoraj * Fix continuation style fixes being applied twice to each line * Update Italian translation - thx bovirus * Update Polish translation - thx potplayer-fanpack ----------------------------------------------------------------------------------------------------- v5.2.0-beta21 (22nd of August 2026) * Add an option to keep the source file's date/time on converted files (batch convert + seconv) - thx TucocoTucoco * Add a Start/Stop server button to the AI review window, and stop llama-server when a review is cancelled - thx emsb8888-prog * Snap whole-paragraph waveform drags to shot changes - thx m0ck69 * Update yt-dlp to 2026.08.19 * Fix "snap to shot change" ignoring the cue gap and often doing nothing at all - thx m0ck69 * Fix waveform edge drags running away when the video position is centered - thx GrampaWildWilly * Fix OCR reading "l" as "i" or "I" without "try to guess unknown words" turned on - thx Miggu82 * Fix the subtitle grid jumping after merging lines with a read-only original open - thx torvchen * Fix focus, scroll and selection in the subtitle grid after delete, insert and merge commands * Fix WebVTT showing a line twice when the same cue is repeated with the same time codes * Update Chinese (simplified) translation - thx emsb8888-prog * Update French translation - thx Need74 * Update Italian translation - thx bovirus * Update Japanese translation - thx hisui3393 * Update Korean translation - thx 12si27 * Update Polish translation - thx potplayer-fanpack ----------------------------------------------------------------------------------------------------- v5.2.0-beta20 (21st of August 2026) * Add "Go to sub position and pause" to the video shortcuts - thx BBB2K * Add a teletext color picker with a "No color" option for EBU STL - thx Triathlon-rally * Add the minimum gap frame rate calculator from SE 4 - thx m0ck69 * Add the SE 4 "toggle custom tags" and recalculate-duration shortcuts - thx m0ck69 * Layout 10: put the edit box under the waveform like SE 4/Aegisub - thx Surlime * Let the preview subtitle use the letterbox bars - thx m0ck69 * Keep the chosen languages when the translate engine changes - thx GrampaWildWilly * Apply new waveform colors without restarting SE - thx m0ck69 * Retry a Crisp ASR clip without VAD when it comes back empty - thx Davanix * Explain the cuBLAS compute type failure and offer the fix - thx tomelephant-git * Smooth the waveform/video cursor and take mpv work off the UI thread * Teletext/EBU STL: preview follows row changes, header round-trip, TS charset crash/leak fixes, honest TT column * Hide burn-in codecs the bundled ffmpeg lacks (Flatpak) - thx zbcoding * Keep the settings column out of the video generating progress bar row - thx zbcoding * Docs: add the chapters page, rebuild the settings page, refresh the engine/format/shortcut lists * Update CrispEmbed to v0.17.9 * Update Italian translation - thx bovirus * Perf: vectorize the OCR bitmap primitives and the spectrogram pixel write * Fix Whisper XXL failing to load on glibc 2.41+ (Fedora 42, Arch, Ubuntu 25.10) - thx zbcoding * Fix the grid color guard dropping legible colors - thx St0rmXtr00per * Fix the position slider jumping while dragging during playback - thx Davanix * Fix clipboard managers failing to paste into text boxes on Windows * Fix a working Crisp ASR build not being downloaded on Intel Macs - thx tolikdgan * Fix update-metainfo-version.sh flattening the AppStream release history, and re-pin the Flathub manifest to v5.1.0 - thx CaptechOmar ----------------------------------------------------------------------------------------------------- v5.2.0-beta19 (20th of August 2026) * Add a Teletext alignment dialog, a TT grid column, and better EBU STL Teletext handling - thx Triathlon-rally * Add "Save all images with HTML index" to the OCR window - thx WAusJackBauer * Use the position the subtitle file itself carries when previewing on the video (TTML/PAC/EBU STL) - thx gavinhonl * Remember which "Multiple replace" categories are expanded - thx KaDeeKe * Let the user own the llama.cpp launch flags in batch convert - thx emsb8888-prog * Start the text-to-speech export folder picker in the subtitle's own folder - thx cvrle77 * Delete the temporary files text-to-speech and ffmpeg leave behind - thx subof * Let clipboard managers paste into the focused text box on Windows (WM_PASTE) * Update the pinned llama.cpp build to b10507 * Fix black video and no sound on GPUs without a Vulkan runtime - thx jugorakita-creator * Fix the batch convert window freezing after minimize/restore, and llama-server surviving cancel - thx emsb8888-prog * Fix the UI freezing during OCR, and line editing breaking after a minimize - thx FunkyKnilch * Fix the Delete key doing nothing after a shift+click range selection in a scrolled grid - thx Davanix * Fix ruby and bouten not being read from real Netflix IMSC 1.1 Japanese files - thx w72k24 * Fix D-Cinema XML two-line events collapsing into one line - thx clundible * Fix Gemma 4 models returning empty translations - thx Oplay66 * Fix the Chatterbox crash after switching to another cloned voice - thx subof * Fix black in-progress status text in the batch convert file list * Update Italian translation and installer - thx bovirus * Update Japanese translation - thx hisui3393 ----------------------------------------------------------------------------------------------------- v5.2.0-beta18 (19th of August 2026) * Add a "Hide tags" mode for formatting in the subtitle grid - thx MrGoraj * Add letter spacing to "Generate video with burned-in subtitles" and "Generate transparent subtitle" * Add a shortcut for pasting clipboard text to a new waveform selection - thx Xenos71 * Add custom server parameters, a stall watchdog, and a token cap to llama.cpp translate - thx emsb8888-prog * Add completion-format translate prompts to LM Studio, KoboldCpp, Ollama and the other local engines - thx subof * Add a way to switch VAD off in Crisp ASR (use --chunk-seconds) - thx AmineI * Apply AI review suggestions in passes instead of applying and closing - thx fraternl * Make Shift+Delete cut in every text box, not only the main edit boxes - thx GrampaWildWilly * Import 22 more SE 4 shortcuts, and fix the four "keep gap" frame moves always being skipped * Render the style previews with libass in the burn-in, transparent video, and ASSA styles dialogs * Show the text of tag-heavy lines in the grid's "show formatting" mode - thx MrGoraj * Extend to shot change: stop at the first cut, and never shorten a line - thx m0ck69 * Find OCR replace lists that only have a "_User" file, and use them everywhere - thx Lentzeris * Retry the update check like downloads do, and log why it failed * Keep the waveform responsive while shot changes are auto-extracted * Speed up waveform generation for 16-bit stereo audio * Update the audio.cpp IndexTTS 2.5 engine to the 2026-08-18 build * Fix speech to text doing nothing at all when ffmpeg cannot be started - thx GrampaWildWilly * Fix "key 'DYLD_LIBRARY_PATH' was not present" starting whisper.cpp on macOS - thx GitGianluc * Fix speech to text not finding the Purfview XXL output when using own --output_dir - thx GrampaWildWilly * Fix crash in batch auto-translate when the API key is missing - thx wjcarpenter * Fix "Remove text for HI" check boxes changing state after a partial Apply - thx Davanix * Fix Netflix IMSC 1.1 Japanese files without bouten being read as EBU-TT-D - thx w72k24 * Fix the subtitle list jumping to the top/bottom after replace or "delete empty lines" - thx St0rmXtr00per * Fix the waveform cursor falling behind the audio during long playback - thx hisui3393 * Fix OCR reading words with more than one "l" as uppercase "I" - thx Miggu82 * Update Chinese (traditional) translation - thx love80312 * Update German translation - thx Need74 * Update Korean translation - thx 12si27 * Update Polish translation - thx potplayer-fanpack ----------------------------------------------------------------------------------------------------- v5.2.0-beta17 (18th of August 2026) * Trim embedded ASSA fonts to the characters actually used (attachments, font collector, batch convert) - thx hisui3393 * Add show/hide settings for the format specific toolbar icons - thx GrampaWildWilly * Add a "Select none" button to the AI review window - thx fraternl * Add the CosyVoice3 RL talker models in both quantizations - thx subof * Show the subtitle on the video in "Visual sync" and "Set sync point" - thx gabiacas-ops * Make the toolbar frame rate usable like the SE 4 one - thx smt-joen * Use an index-mapped scrollbar for the subtitle list in the OCR window * Update ffmpeg to 9.0.1 (Windows) / 9.0 (macOS arm), Tesseract to 5.5.3, whisper.cpp to v1.9.2, libmpv to shinchiro 20260814, and Paddle OCR standalone to v1.4.0 * Fix Shift+Delete still not cutting in a text box - thx GrampaWildWilly * Fix per-line voice cloning hanging on Windows - thx invio-a11y * Fix capitalization after "EE.UU." and other doubled Spanish abbreviations - thx fraternl * Fix the Spanish OCR replace rule for "clón" breaking words like "ciclón" - thx fraternl * Fix audio clip extraction progress jumping after Subtitle Edit regains focus - thx Davanix * Fix speech to text failing on videos where the audio is not the first stream - thx emsb8888-prog * Fix speech to text transcribing a different audio track than the engine picks on multi-track files * Fix the main window being stuck on screen when a modal dialog is minimized - thx emsb8888-prog * Fix text to speech losing the whole run when a local engine crashes mid-generation - thx subof * Fix the waveform cursor running ahead of the audio when playback resumes - thx hisui3393 * Fix ASSA styles from the style storage changing values after OCR - thx FunkyKnilch * Update Bulgarian translation - thx jekovcar * Update German translation - thx Need74 * Update Italian translation - thx bovirus, GitGianluc * Update Japanese translation - thx hisui3393 * Update Russian translation - thx jekovcar ----------------------------------------------------------------------------------------------------- v5.2.0-beta16 (17th of August 2026) * Add a chapter editor with Matroska/MP4/OGM chapter formats (Video > More > Chapters) - thx alex64-pb * Add "Find voices in video and clone them all", plus cloning the voice of each line from the video - thx invio-a11y * Add the Voxtral backend to Crisp ASR speech to text - thx AmineI * Add "Final Cut Pro Xml Captions" export * Add DaVinci Resolve marker EDL import/export * Add reading of Adobe Premiere Pro "Markers" panel csv exports * Add text box "Copy (alternative)"/"Paste (alternative)" shortcuts, and make Shift+Delete really cut - thx GrampaWildWilly * Add .avi/XSUB reading to "seconv" * Add reading of IEEE float and WAVE_FORMAT_EXTENSIBLE wav files for the waveform * Open BDN xml and Final Cut Pro image xmeml from the main window (via OCR) * Show translation progress in batch convert - thx emsb8888-prog * Name batch auto-translated files after the target language - thx emsb8888-prog * Name forced container tracks with a ".forced" token in "seconv" * Update Crisp ASR to v0.8.29, and Chatterbox to the versioned V3 models with cross-lingual voice cloning * Improve container support after sweeps of MP4Box, ffmpeg, mkvmerge, TSDuck and Bento4 output * Improve broadcast format round-trips (PAC, Cavena 890, SCC, Cheetah, MacCaption, EBU-TT-D, Ayato, CapMaker Plus, IMSC 1.1) * Improve Apple format support - read Final Cut Pro captions and current fcpxml versions * Improve Adobe format support - read the EDLs and XMP markers Premiere/After Effects write, fix Encore round-trips * Fix negative time codes being zeroed when selecting a line - thx DexKen * Fix reading of Matroska files on a network share still being slower than in beta 13 - thx 1476523 * Fix Ctrl+V in the subtitle grid not selecting/scrolling to the pasted lines - thx Davanix * Fix Ctrl/Cmd+V not overwriting the selected lines in the subtitle grid - thx rosilucia-hub * Fix Sub Station Alpha (.ssa) style colors being written as unreadable 32-bit values - thx schabau * Fix speech to text clipping the audio sent to the engine - thx AmineI * Fix Faster-Whisper XXL/CTranslate2 getting the extracted wav instead of the source file - thx popoche * Fix DVB subtitle timing when a cue starts at the video's first frame * Fix the DaVinci Resolve csv "Play" flag being read from the wrong column * Update Bulgarian translation - thx jekovcar * Update French translation - thx Need74 * Update Polish translation - thx potplayer-fanpack * Update Russian translation - thx jekovcar ----------------------------------------------------------------------------------------------------- v5.2.0-beta15 (16th of August 2026) * Add IndexTTS 2.5 as a text to speech engine, with emotion and speaking rate control * Add MiLMMT-46 translation models to the llama.cpp translate engine * Add HunyuanOCR 1.5 to the llama.cpp OCR model list * Add the "llama.cpp advanced" engine to batch convert, with custom prompt and parameters - thx emsb8888-prog * Add "Convert colors to dialog", "Snap time codes to frames" and "Add folder..." to batch convert * Add a "Play current" button to the AI review window * Add a reset-to-default button for the auto-translate prompt * Add an engine download/update button to the llama.cpp OCR settings - thx fraternl * Ask before the first voice clone in "Text to speech" * Open SMPTE-TT files with bitmap captions * Show "List shot changes" whenever shot changes exist, and keep edits made in the list * Show the errors that could not be fixed in "Fix common errors" - thx radektuma * Make "seconv" fail on unknown options, and add --json output everywhere plus --help-json * Improve subtitle format detection and reading after a sweep of the VideoLAN/CCExtractor sample files * Improve OCR line splitting with an adaptive minimum line height, plus an un-italic retry in nOCR * Fix reading of Matroska files on a network share being much slower in beta 14 - thx 1476523 * Fix whisper.cpp not starting on Linux - thx keremyavuz * Fix crash when closing "Text to speech" after reviewing audio - thx cvrle77 * Fix new lines in ASSA getting the first style in the header instead of the neighbouring one - thx Khaztaroth * Fix changing format to Advanced Sub Station Alpha ignoring the default styles storage category - thx windias * Fix column paste going to the wrong row when the rows were selected top-to-bottom - thx rosilucia-hub * Fix "Remove text for hearing impaired" removing dialog dashes in 3-line subtitles - thx ubeccp6a * Fix "Remove text for hearing impaired" with a speaker name on a line of its own - thx ubeccp6a * Fix speech to text with Qwen3 ASR failing with a JSON error on comma-decimal locales - thx zielinou * Fix switching audio track generating a waveform even with auto-generate off - thx Davanix * Fix italic OCR losing word spacing and reading "l" as "i" with binary image compare - thx Miggu82 * Fix OCR replace lists dropping all but the first "if spelled correctly" section - thx Miggu82 * Fix six small regressions found in the previous day's changes * Update Korean translation - thx 12si27 * Update Polish translation - thx potplayer-fanpack * Update Portuguese (Brazil) translation - thx igorruckert * Minor performance improvements in seconv and auto-break ----------------------------------------------------------------------------------------------------- v5.2.0-beta14 (15th of August 2026) * Add back .webm output for "burn-in subtitle", add .ts, and offer only containers the codec can use * Let "Visual sync" find or open a video when only a subtitle is open * Show install status dots for llama.cpp in the OCR window engine and model lists * Fix OCR with llama.cpp returning nothing but blank lines - thx fraternl * Fix "AI review" pairing corrections with the wrong lines - thx coreyjv * Fix undo stepping back through a waveform drag instead of to before it - thx thelabcat * Fix overlapping lines from speech to text with Crisp ASR and Parakeet - thx perkesfurmah-collab * Fix the subtitle grid jumping away from the line being edited when a row changes height - thx radektuma * Fix column shortcuts not working while the text box has focus - thx SiaL8er * Fix speech to text failing to extract audio with "-map 0:1 matches no streams" - thx slothsh * Fix "Fix short display time" listing fixes that change nothing - thx Pemicope * Fix Google Translate V1 inserting blank lines in multi-line texts - thx radektuma * Fix "Keep partial transcription?" discarding the transcribed lines anyway * Fix cancelling speech to text not stopping the rest of a batch * Fix engine and ffmpeg processes left running after closing the speech to text window * Improve reading of Matroska files on a network share further, now walking clusters in parallel - thx 1476523 * Update German translation - thx Need74 * Update Japanese translation - thx hisui3393 ----------------------------------------------------------------------------------------------------- v5.2.0-beta13 (14th of August 2026) * Add an "Edit original" mode, and make the original reference rows editable in place - thx bichitoxxx * Add back the SE4 option to remove blank lines when opening a subtitle - thx Morv55555 * Add Ctrl+Shift+End/Home and friends for extending the selection in the subtitle grid * Make WebVTT cue settings configurable again - thx apartmandairesi * Make the default waveform paragraph backgrounds more visible * Give the subtitle grid an index-mapped scroll bar so the thumb no longer jitters - thx St0rmXtr00per * Update the speech-to-text advanced parameter help to match the engine versions being downloaded * Fix waveform dragging not following the pointer through mid-drag scrolls - thx bichitoxxx * Fix "Remove text for hearing impaired" reporting pasted multi-line text as changed - thx Davanix * Fix slow reading of Matroska files on a network share - thx 1476523 * Update Japanese translation - thx hisui3393 * Update Polish translation - thx potplayer-fanpack * Minor performance improvements in waveform scrubbing, video loading, OCR and text handling - thx ivandrofly ----------------------------------------------------------------------------------------------------- v5.2.0-beta12 (13th of August 2026) * Add CrispEmbed as an engine for "OCR burned-in subtitle", with its own settings dialog - thx tal-maker * Add Chatterbox Base in F16 and Q4_K alongside Q8_0 - thx subof * Add select all/none/invert to the "Multiple replace" preview list - thx Davanix * Show what each engine is in the "OCR burned-in subtitle" engine list * Remember the numbers and formats seven more tool dialogs ask for * Use the usual tick-all shortcuts in the "Multiple replace" rule category picker * Make the AI review window a little wider * Keep Google Lens OCR going when a single image fails instead of stopping the run - thx fartle * Fix crash when closing "Text to speech" after reviewing audio - thx cvrle77 * Fix undocked video player being in front of the main window at startup - thx GrampaWildWilly * Fix "Multiple replace" freezing on a slow regular expression rule * Fix batch convert failing every file when one replace rule does not compile * Fix the same original line being shown on two rows when opening a non-matching original * Fix transparent subtitles ignoring the output folder picked in its own settings dialog * Fix Perplexity API key, model and URL being forgotten on restart * Fix "Apply minimum gap" mixing up frames and milliseconds * Fix "Export plain text" saving its settings when cancelled * Fix "OCR burned-in subtitle" repeating a line, and saving engine settings while browsing the lists * Fix "OCR burned-in subtitle" running to the end without reporting a broken CrispEmbed/llama.cpp server * Fix batch convert dialogs showing the main window's file name in the title bar * Fix hardcoded English in the transparent video, burn-in and batch convert output folder pickers * Fix Chatterbox retrying an impossible reference voice repair once per line * Fix timers and preview images left running after closing several dialogs * Update Japanese translation - thx hisui3393 * Update Korean translation - thx 12si27 * Update Chinese (Traditional) translation - thx love80312 * Update French translation - thx Need74 * Minor performance improvements in string handling ----------------------------------------------------------------------------------------------------- v5.2.0-beta11 (12th of August 2026) * Add import/export of selected rule categories to "Multiple replace" - thx KaDeeKe * Add move up/down/to top/to bottom shortcuts to "Multiple replace" - thx KaDeeKe * Add select all/none/invert to the "Remove text for hearing impaired" preview list - thx fraternl * Add rule selection to seconv --remove-formatting - thx Hlsgs * Show the whole rule while editing it in "Multiple replace" - thx KaDeeKe * Show the subtitle file name in dialog title bars * Give user-entered regular expressions a match timeout, and mark broken rules in "Multiple replace" - thx KaDeeKe * Fix "Multiple replace" preview freezing after a failed rule - thx KaDeeKe * Fix split/rebalance long lines forgetting its settings - thx cvrle77 * Fix the scroll bar trough paging past the cursor in the subtitle grid - thx ivandrofly * Fix context menu hiding behind the undocked audio visualizer - thx GrampaWildWilly * Fix unchanged text being invisible in the diff previews - thx fraternl * Fix the find window trimming edge spaces from the restored search pattern - thx nms42 * Fix Chatterbox failing on reference voices that are not 24 kHz mono - thx subof * Fix Subtitle Edit staying in front of other applications during transcription and at startup * Update Japanese translation - thx hisui3393 * Update Russian translation - thx jekovcar * Update Bulgarian translation - thx jekovcar * Update Polish translation - thx potplayer-fanpack * Minor performance improvements in name lists, grid rendering and the waveform ----------------------------------------------------------------------------------------------------- v5.2.0-beta10 (11th of August 2026) * Add "BDN/xml 8-bit" export with palette-indexed PNGs * Add "Multiple replace" to the toolbar * Add 23-language selection to Chatterbox TTS (and re-download of the English-only models) * Bring back the "Full frame image" option when exporting images * Open a non-matching original as a read-only reference instead of dropping lines - thx bichitoxxx * Let the replace window pick which columns to search - thx jugorakita-creator * Honor the Subtitle Edit proxy settings in downloads, update check, and cloud OCR * Show the number of forced cues in the Matroska track picker * Show the filtered file count in batch convert * Improve VobSub color isolation for letters like "i", "j" and umlauts - thx tekk42 * Fix wrong burn-in quality with the VideoToolbox encoders on new installs (macOS) * Fix Persian/RTL text and dark theme highlighting in the compare window - thx saeead * Fix keyboard focus being stolen from an open modal dialog - thx GrampaWildWilly * Fix AI review failing with engines that reject some request parameters - thx joseacuna80 * Fix blank space above the first line after End followed by Home * Fix wrong uppercasing after ASSA tags in "Fix casing" - thx ivandrofly * Update CrispASR to v0.8.28 * Update Japanese translation - thx hisui3393 * Update Korean translation - thx 12si27 * Update Traditional Chinese translation - thx love80312 * Update Turkish translation - thx bilimiyorum * Update Russian translation - thx jekovcar * Update Bulgarian translation - thx jekovcar ----------------------------------------------------------------------------------------------------- v5.2.0-beta9 (10th of August 2026) * Add image subtitle preview and VobSub export to the Matroska track chooser * Add "Train nOCR" from SE 4, and speed up nOCR matching with a result cache * Make "Set up like Subtitle Edit 4" also arrange the waveform toolbar like SE 4 * Default to the h264_videotoolbox hardware encoder for burn-in on macOS * Improve nOCR matching of small and sensitive characters (O/o/0, quotes, dashes) * Stop "play selection" on the line's last visible frame instead of overshooting * Sync the Paddle OCR language list with PaddleOCR 3.4 and the bundled models * Let the CrispEmbed OCR hardware build be re-picked after the first install * Fix target file size when burning in with VideoToolbox encoders - thx luisblop * Fix burn-in re-runs reusing the previous run's bit rate, and a stuck dialog when the bit rate was too low * Fix subtitles intermittently missing in fullscreen playback - thx Davanix * Fix secondary subtitle showing tiny or huge in the video player, and preview flicker - thx GrampaWildWilly * Fix lost window focus after applying settings with undocked windows - thx GrampaWildWilly * Fix modal dialogs dropping behind other windows - thx GrampaWildWilly * Fix "Fix names" preview races and make the preview update instantly - thx ivandrofly * Fix ASSA output writing a mismatched events Format line for files from MKV * Fix OCR scrolling away from the selected line when stopping OCR * Fix spell check source image preview showing a large transparent area * Fix 22 broken format string placeholders in the Bulgarian translation * Update CrispEmbed OCR to v0.17.8 * Update Russian translation - thx jekovcar * Update Bulgarian translation - thx jekovcar * Update Korean translation - thx 12si27 * Update Japanese translation - thx hisui3393 * Minor performance improvements in text processing - thx ivandrofly ----------------------------------------------------------------------------------------------------- v5.2.0-beta8 (9th of August 2026) * Add Apple VideoToolbox hardware encoders to "Burn-in" on macOS (and only list encoders available on the OS) * Add "Go to next empty line" shortcut (SE 4 parity) - thx Surlime * Add frame rate to the export images window (fixes wrong BDN/XML fps) * Add update check settings (stable/beta channel) and a passive update notification at startup - thx korsosattila73 * Fix rule profiles not applying the minimum/maximum duration limits * Fix rule profiles losing the custom continuation style on save/load, and one bad profile breaking the whole list * Fix dialogs, popups and Alt+letter menus hiding behind undocked tool windows - thx GrampaWildWilly * Fix lost focus after window activation loss and caret position when navigating the grid - thx St0rmXtr00per * Fix crash in ElevenLabs text to speech preview - thx cvrle77 * Fix bold "Line length" label after switching layouts - thx Davanix * Fix "Remove text for HI" reporting a trailing empty line as a change - thx Davanix * Tag all HEVC burn-in output for QuickTime/Apple compatibility (hvc1) * Minor performance improvements in OCR and list handling - thx ivandrofly ----------------------------------------------------------------------------------------------------- v5.2.0-beta7 (8th of August 2026) * WebVTT: bring back the style manager, voices and the browser preview from SE 4 * WebVTT: show styles and voices in the subtitle grid * Add advanced ASSA effects: Word flip 3D, Lower third and Cinematic title * Restore the four missing SE 4 subtitle grid double-click actions - thx PatriceBG * Add Shift+Backspace as forward-delete in the text boxes (for keyboards without a Delete key) * Point sync now also works without a video loaded - thx fraternl * Update the Netflix quality checks to the current Timed Text Style Guides * Update the auto-translate model lists (August 2026) * Update llama.cpp to b10310 and CrispEmbed OCR to v0.17.6 * Fix crashes, locale number corruption and wrong tags in the advanced ASSA effects * Fix "Set position" writing coordinates in video pixels instead of script resolution - thx Bibi-skz * Fix "Set position" preview not using the line's own style - thx Bibi-skz * Fix "Apply custom override tags" preview ignoring the file's styles, and OK without video - thx Bibi-skz * Fix ASSA styles being lost in batch convert "Split long lines" and waveform "Delete lines from video position" * Fix typing accents (dead keys) on Linux with ibus - thx Pemicope * Fix the video playing briefly when opening a file - thx rRobis * Fix the grid jumping to the next line after "play selection, then stop" - thx ghostminhtoan * Fix text to speech engines cluttering their voice folder instead of using the output folder - thx subof * Fix the chosen language being ignored by OmniVoice speech generation - thx subof * Fix seven bugs found in a review of the beta 6 changes (message box keys, batch convert, file names) * Faster "Fix common errors" select-all, "replace all" with whole word, and file compare * Update French translation - thx Need74 * Update Polish translation - thx Adam Malich * Update Korean translation - thx 12si27 * Update Japanese translation - thx hisui3393 * Update Hungarian translation - thx Zityi ----------------------------------------------------------------------------------------------------- v5.2.0-beta6 (7th of August 2026) * Add "do not break after" list and bottom-heavy percent to the auto-break settings - thx Davanix * Add an editor for the "do not break after" lists - thx Davanix * Add option to include the language code in batch speech to text file names - thx wjcarpenter * Add the CrispEmbed OCR engine to batch convert * Batch convert: choose the frame rate source for "Beautify time codes" (matching video file or fixed rate) * Make the subtitle text box section resizable - thx Ironship * Move focus between message box buttons with the arrow keys * Show Ollama OCR errors instead of silently producing empty lines - thx perkesfurmah-collab * Fix "merge as dialog" not using the configured dialog style - thx jugorakita-creator * Fix stuck ".xdp-" temp file names when saving via the Flatpak document portal - thx wjcarpenter * Fix delete not working on the first subtitle row right after startup - thx Davanix * Fix changes to the original subtitle being lost when closing with "Yes" - thx Ironship * Fix menu bar keyboard issues: focus drops, Escape highlight, F10 underlines, submenu z-order - thx GrampaWildWilly * Fix "split line at cursor" not applying the continuation style - thx Ironship * Fix command line convert adding periods before known names in "Fix common errors" - thx Ironship * Fix repeated folder names when extracting 7-Zip archives (Faster Whisper XXL) - thx Ironship * Fix the app closing completely when clicking OK in "Text to speech" - thx Ironship * Fix the AI review delay input being too narrow - thx pdjdev * Faster typing and video preview with mpv - thx hisui3393 * Fewer string allocations in hot paths - thx ivandrofly * Update Japanese translation - thx hisui3393 ----------------------------------------------------------------------------------------------------- v5.2.0-beta5 (6th of August 2026) * Add option to keep the end time when merging lines (allow overlap) - thx hisui3393 * Add "Beautify time codes" to batch convert - thx xyv1997 * Open the matching video automatically after OCR - thx fraternl * Use the container's default audio track when opening a video - thx fraternl * OCR: add the DeepSeek-OCR-2 backend (CrispEmbed v0.17.5) * Apply duration limits: the "do not go past shot change" option now works * Speech to text: select a Whisper engine on first run * Auto-translate via copy-paste: keep all ASSA tag groups in the translated text - thx FoxyLoon * Text to speech: keep the target language on cloning and detect the reference language - thx subof * ASSA: pick the most visible color for the subtitle grid when fading colors are used - thx rRobis * Copy ASSA/SSA lines to the clipboard without the file header - thx ghostminhtoan * Keep exact WebVTT cue position settings on save - thx Foxaryse * Keep multi-word "do not break" phrases together when auto-breaking - thx uckthis * Fix subtitle images changing vertical position with the text in PGS/VobSub export - thx robertmajsky * Fix SoftNI sub export always writing 25 fps instead of the project frame rate - thx rodnvs * Fix the CJK line termination check with Whisper Chinese - thx ajaska * Fix auto-trim white space merging Dutch words like "ze 's avonds" - thx fraternl * Fix "Start with uppercase letter after paragraph" treating single-letter lines as paragraph ends - thx fraternl * Fix long localized menu items being clipped by the popup width cap - thx Need74 * Fix subtitle language names in the MP4 track picker - thx 5mrpcn * Fix "Download VLC" never completing when libVLC is already installed - thx hisui3393 * Fix poor contrast of the settings icons with custom colors - thx rRobis * Fix the last change log line being covered by the horizontal scrollbar - thx ivandrofly * Fix "Return" not working as a shortcut when the subtitle grid has focus - thx Surlime * Fix the first line not being selected when opening a subtitle - thx fraternl * Faster waveform dragging and text editing - thx hisui3393 * Faster merging of selected lines, auto-break and image based export * Fewer allocations in the UI value converters * Update Turkish translation - thx bilimiyorum * Update French translation - thx Need74 * Update Traditional Chinese translation - thx love80312 ----------------------------------------------------------------------------------------------------- v5.2.0-beta4 (5th of August 2026) * Read subtitles from fragmented MP4 (DASH/CMAF): WebVTT, TTML/IMSC1 and tx3g * Offer the MP4 track picker for fragmented files with several subtitle tracks * Read stxt/sbtt text streams, tx3g styles and the VobSub palette from MP4 * Source view: find/replace, go to line, live format check and a prompt before discarding edits * Add "Go to first line"/"Go to last line" shortcuts with video sync (Ctrl+Home/End) - thx GrampaWildWilly * OCR: add the PP-OCRv6 backend (CrispEmbed v0.17.2) * AI review/assistant: add Gemma 4 E2B, EuroLLM 9B/22B and Granite 4.1 8B * Update ffmpeg downloads - Windows 9.0, macOS ARM 8.1, macOS Intel 8.0 * Update llama.cpp to b10256 * Accessibility: announce combo box and spinner value changes, and let Alt close an open drop-down - thx Shaima-Almarzooqi * Spell check: detect the subtitle language again after opening/importing a file - thx fraternl * Fix right-to-left text rendering in dialog subtitle grids - thx AliNet1974 * Fix the Settings dialog not opening with focus - thx GrampaWildWilly * Fix the subtitle grid getting stuck on the first caption after opening a file - thx GrampaWildWilly * Fix the initial video position not being 0:00 - thx GrampaWildWilly * Fix menus hidden behind the undocked audio visualizer - thx GrampaWildWilly * Fix undocked video controls popping up on mouse movement in other apps - thx GrampaWildWilly * Fix OmniVoice text to speech failing to start with the CUDA engine - thx subof * Explain native engine exit codes instead of showing a raw number - thx subof * Fix the CrispEmbed server exiting with code 127 on Linux - thx perkesfurmah-collab * Fix speech to text failing to start in the Flathub package (missing OpenBLAS) - thx DarkSwan86 * Fix Statistics and batch convert crashing with the Bulgarian UI (bad format strings) * Fix the burn-in window losing its minimum width while generating * Faster subtitle grid selection handling and time code display - thx ivandrofly * Update Polish translation - thx potplayer-fanpack * Update French translation - thx Need74 * Update Korean translation - thx 12si27 * Update Japanese translation - thx hisui3393 ----------------------------------------------------------------------------------------------------- v5.2.0-beta3 (4th of August 2026) * Extract ARIB STD-B24 captions from transport streams (Japanese ISDB, Brazilian SBTVD) * Add D-Cinema interop properties dialog - thx igitlin * ASSA: embed fonts from the font collector, in the style editor and in batch convert * OCR: auto-detect source language and let the dictionary follow the language combo - thx fraternl * Text to speech: output language for CosyVoice3/Qwen3 cross-lingual cloning - thx subof * Text to speech: OmniVoice now works on Pascal/Volta CUDA cards - thx subof * Fix common errors: draggable splitter between the fixes and the preview - thx perkesfurmah-collab * Accessibility: announce grid time codes and name the Shortcuts list items - thx Shaima-Almarzooqi * Fix video playback on Windows ARM64 (an x64 mpv was downloaded) - thx Shaima-Almarzooqi * Fix crash on macOS with AV1 video by updating the bundled dav1d - thx almog6500 * Fix "Copy text only" pasting lines out of order for a reverse selection - thx hisui3393 * Fix keyboard navigation of the menu bar with no subtitle loaded - thx GrampaWildWilly * Auto-translate via copy-paste: overwrite re-translated lines instead of appending them - thx revov * Fix window cleanup running twice when closing - thx ivandrofly * Fix "Remove formatting - font name" doing nothing - thx Ironship * Fix Enter not working in the "Replace with" box - thx Ironship * Fix Escape not closing the "Generating audio" window - thx Ironship * Fix double key handling in the Media info window - thx Ironship * Fix the indeterminate progress bar not resetting on stop - thx Ironship * Fix the output folder field staying disabled in Batch convert settings - thx Ironship * Fix Binary OCR database windows showing nOCR titles - thx Ironship * Fix the burn-in window layout: progress bar overlapping the file size box - thx Ironship * Fix the file size converter keeping the suffix of the previously loaded language - thx Ironship * Use the dark theme background in the syntax text editor - thx pdjdev * Add "Pl." (plaza) to the Spanish abbreviation list - thx fraternl * Localize many remaining hardcoded strings (downloads, export menus, engines) - thx Ironship * Update Polish translation - thx potplayer-fanpack * Update Turkish translation - thx bilimiyorum * Update Portuguese (Brazil) translation - thx igorruckert ----------------------------------------------------------------------------------------------------- v5.2.0-beta2 (3rd of August 2026) * Add type-to-search to all combo boxes - thx 9jfcc * Add "Sentence case" and make the casing options visible in the text box context menu - thx shanedk * Restore the copy-to-clipboard items in the subtitle grid context menu * Column "Paste from clipboard" accepts plain text again (text column only, like SE4) * Fix common errors: don't uppercase the word after an abbreviation like "dhr." - thx fraternl * Add abbreviation lists for 30 more languages * Fix the ASSA style editor losing data on rename and style switch - thx SiriosDev * Fix F10 not activating the menu bar: the default F10 binding is removed - thx GrampaWildWilly * Fix Alt not activating the menu bar after a dialog opened while Alt was held - thx GrampaWildWilly * Fix errors and a hang after closing the video player (mpv teardown race) - thx GrampaWildWilly * Text box: place the cursor at the line start when navigating between lines - thx lsmxpftz-cmyk * Join subtitles: the files can be rearranged again - thx Nino-kun * Source view: fix ASSA property contrast in the light theme, make empty areas clickable * Fix the Polish translation not loading (trailing commas in the JSON) and validate all in CI * Update Japanese translation - thx hisui3393 * Update Korean translation - thx 12si27 * Update Polish translation - thx potplayer-fanpack * Update Chinese Traditional translation ----------------------------------------------------------------------------------------------------- v5.2.0-beta1 (2nd of August 2026) * Auto-translate: advanced llama.cpp and Ollama engines with batch context - thx subof * Auto-translate: system prompt, parameter settings and schema-forced output for these engines * Statistics: new dashboard view with tiles, meters, checks and a CPS histogram * Add "Remove/replace Unicode characters" tool (SE4 plugin port) - thx fraternl * Add ASSA font collector * Find/replace: SE4-style regex snippet context menu * Import plain text: add "Align time codes via forced aligner" - thx subof/David * Cut video: option to also cut the subtitle * Batch convert: brightness, alpha and color adjustments for image output * Open video from URL: video/subtitle checkboxes and auto-generated subtitles * Open video from URL: show the yt-dlp update prompt as a message box * Clean up YouTube auto-generated captions on load into one plain-text cue per spoken line * Waveform: per-track audio picker with size estimates, live progress and cancel - thx shanytc * Waveform: copy the subtitle at the video position and paste at the waveform position - thx shanytc * Text to speech: custom voice models for Piper * Fix common errors: context menu with applied-rule details and a rule filter * Add default save location setting - thx gadkarisid * Add auto-break settings under Settings -> Tools - thx citizenhelene * Add more proxy settings: domain, system credentials and a bypass list * Add optional Auto-translate, Speech to text and Point sync toolbar icons - thx fraternl * Add "Show selected lines earlier/later" to the subtitle grid context menu * Add "Go to video position..." to the Video menu and a toggle-subtitles video player shortcut * Add export support for Audacity/Tenacity labels * Machine-readable ffmpeg progress for shot changes, burn-in, re-encode and transparent subtitles * Help: respect a custom help shortcut, wire F1 in more windows, and add five new help pages * Remove the MLX Whisper speech-to-text engine - thx ruonghieu * Performance: faster hot paths across the UI and core libraries - thx ivandrofly * Build asset zips from source folders at build time and ship translations compressed * Move auto-translate, speech to text and OCR to libuilogic, keeping libse a pure subtitle library * Honor ASSA override tags (position, alignment, colors, ...) in image-based export - thx Maxia1 * Honor {\pos(x,y)} in VobSub and DVD sup export * Fix red and blue being swapped in VobSub, SSA and TextST colors * Read the frame size from a standalone VobSub .idx * Keep the source position when re-exporting image subtitles * Fix undo wiping the original subtitle text column - thx abasameer * Fix Paddle OCR not starting from the GUI on macOS - thx HongyuS * Fix a batch of auto-translate bugs found in a bug hunt * Make ComboBox dropdowns follow the UI scale setting - thx goonis * Fix Google Lens OCR putting dialog dashes on their own lines - thx flashmasta * Fix inverted I/1-to-l OCR rule and extend the Dutch OCR replace list - thx fraternl * Restore total seconds/milliseconds time codes in export custom text format * Self-repair the unpacked theme folder when an icon file is missing * Point sync via another subtitle: take left times from the original timings - thx fraternl * Apply "extend to line before/after" to all selected lines - thx NiggleHub * Skip an auto-backup tick while a save is still in flight - thx ivandrofly * Download 64-bit libVLC on Windows x64 - thx StandAloof * Screen readers: the subtitle grid exposes keyboard focus via UI Automation again - thx Urs2026 * The main subtitle grid and all dialog grids now use TableView * Accessibility: announce rule violations in the grid row's accessible name, safer initial focus * Accessibility: accessible names for icon buttons, keep grid keyboard focus on Home/End jumps * Page up/down moves the selected line in the subtitle grid and dialog grids - thx SiriosDev * Flatpak: bundle OpenBLAS so the CrispASR engines can start in the sandbox - thx DarkSwan86 * Speech to text: report a missing shared library instead of silently returning an empty result * Fix "set start/end time" doing nothing when the video position was past the other time code - thx LabinSub * Modify selection: new "Hearing impaired (SDH)" rule - thx subof * Add "AI review" to the subtitle grid "Selected lines" context menu * AI review: new "delay in seconds between requests" setting for rate-limited cloud engines * Find and replace: also search the original text column in translator mode - thx radektuma * Fix the SSA storage styles context menu reading the file style list * Fix the UI freezing while text to speech seeded reference voices with ffmpeg * Update Italian translation - thx bovirus, SiriosDev * Update Korean translation - thx 12si27 * Update Polish translation - thx potplayer-fanpack ----------------------------------------------------------------------------------------------------- v5.1.0 (29th of July 2026) Subtitle Edit 5.1 is the first feature release after 5.0, collecting the work of 15 betas and 18 release candidates. A summary of the changes since v5.0.0 (full details in the pre-release entries below): * AI review (Tools > AI review) - proofread subtitles with a local LLM via llama.cpp or Ollama, with a before/after grid and an editable prompt * AI assistant in the subtitle text box - quick actions like fix spelling/grammar, reading speed and change tone * Auto-translate: new generic OpenAI-compatible engine, remote llama.cpp server option, DeepL formality picker, and many new local models (TranslateGemma, Gemma 4, Hy-MT2, Qwen 3.5/3.6, Aya Expanse) * Auto-translate: retry with backoff on rate-limited (429) LLM engines, show the API URL in error details, and fix engines that ignored a custom URL * Speech to text: new engines and models - MOSS-Transcribe-Diarize with speaker labels, OpenRouter and Alibaba Qwen3-ASR online engines, and Parakeet, Kyutai, SenseVoice, ARK-ASR and Cohere models - plus live progress for all CrispASR engines, and the input language is remembered again * Text to speech: new MOSS-TTS engine with zero-shot voice cloning and a target language dropdown, new Q6_K/Q8_0 quants, a big window overhaul, and a large bug-fix pass across the engines and the review window * OCR: new CrispEmbed local OCR engine (GLM-OCR, GOT-OCR2, Qwen3-VL), llama.cpp OCR vision models in Batch convert and seconv, a "Show only forced subtitles" filter, and histogram-based color isolation for VobSub/DVD subtitles * New formats: EBU-TT-D (read/write), IMSC 1.1 image profile export, and "DVD sup (MuxMan/Scenarist)" image-based export * macOS: multi-window support - "File -> New window" opens an independent editor window * New "file saved" prompt with file info, copy path and Play/Open - used across text to speech and the video tools * Waveform: SE4-style center/lock scrubbing (paused and while playing), configurable toolbar with seek and play buttons, drag a multi-line selection to shift time codes together * Subtitle text box: new native tag-coloring text box - IME/CJK input, right-to-left and live spell check now work with color tags enabled * Live spell check in the subtitle grid context menu, spell check window fixes, MS Word engine fixes, and new Georgian, Armenian, Luxembourgish and Faroese dictionaries * Fix common errors: reworked fix list with category chips, sortable columns and live counts - and unticked fixes are no longer applied * Shortcuts window: categories with icons, sortable columns, better search - plus many new SE4-parity shortcuts (column shortcuts, set start and go to next, play previous/next, extend, and more) * Subtitle grid: type-a-number line navigation, alternating row colors, text fit setting (clip/wrap/ellipsis), and "Insert subtitle after current line..." keeping the inserted file's own time codes * Auto-save of the open subtitle file (Options > General) * Right-to-left: mirrored grid columns, content direction in text boxes, and an ASSA karaoke RTL option * Accessibility: menu bar activation via Alt/F10, and named controls for screen readers across the app * Image-based subtitle editor: PGS position monitor, adjust color for selected lines, change speed, and HTML index export * Point sync and visual sync: sorted sync points, re-pointing, delete via menu/keyboard, and a time code box for the sync point * Split/re-balance long lines: respect the single line max length * Auto-trim white space: use the detected language, so Dutch keeps " 's" * Network: always bypass the proxy for local (loopback) URLs, and proxy password/crash fixes * Performance: non-English start-up about 3.5x faster, 10-25x faster auto line-breaking, much faster subtitle grid and waveform rendering, faster .srt/ASSA loading and MP4 parsing, and a wide allocation-cutting pass (source-generated regexes, reused StringBuilders) - thx ivandrofly * Stability: hundreds of fixes, including a 46-bug codebase sweep, an async/await bug hunt, and hardened binary subtitle parsers * Command line (seconv): llama.cpp auto-translate and OCR, image output styling options, fix-common-errors profile settings in --settings JSON, VobSub/.idx fixes, and clear errors on malformed SCC input * The libse core library is now published on NuGet * New UI languages: Arabic, Turkish, Persian, Traditional Chinese, Indonesian, Vietnamese, Greek and Thai - and all language files synced with English (3,000+ strings) * Updated engines: CrispASR v0.8.24, llama.cpp b10142 (with a CUDA 13 build on Windows), whisper.cpp v1.9.1, CrispEmbed v0.16.1, Qwen3 ASR CPP v0.1.7 and yt-dlp 2026.07.04 A big thank you to everyone who tested the pre-releases, reported issues and contributed code and translations - and a special thanks to Anthropic for sponsoring a Claude subscription :) ----------------------------------------------------------------------------------------------------- v5.1.0-rc18 (28th of July 2026) * Set sync point: add a time code box so the sync point can be typed directly (SE4 parity) * Fix CrispASR text to speech freezing the app on every generation - thx subof * Speech to text: point at the Model field when an OpenAI-compatible server rejects it - thx acc4github * Surround only the selected text with music symbols - thx fraternl * Fix common errors: flash "Analyzing..." on every re-scan - thx Davanix * Restore the SE4 order of the two split items in the text box right-click menu - thx Chouu * Fix truncated labels and narrow controls in two spell check dialogs - thx bovirus * Fix the subtitle grid scroll bar paging past the cursor when holding the trough - thx ivandrofly * Fill in the missing Dutch translations - thx Albertos22 * Update the Italian translation - thx bovirus * Improve the Polish translation of the OCR embedded subtitles dialog - thx potplayer-fanpack ----------------------------------------------------------------------------------------------------- v5.1.0-rc17 (27th of July 2026) * Update llama.cpp to b10142 and offer the CUDA 13 build on Windows * Update CrispASR to v0.8.23 - fixes garbled/hung CosyVoice3, Qwen3-TTS and Zonos audio on the Vulkan build, and voice cloning now needs to be enabled in settings - thx CrispStrobe * Add 4 Parakeet and 1 Kyutai speech-to-text model to the CrispASR engines * MOSS-TTS: add a target language dropdown, so cross-lingual voice clones are no longer heavily accented - thx subof * Fix CosyVoice3, VoxCPM2 and F5-TTS asking for a voice transcription the voice pack already ships, and fix ref-text cues being dropped when one was a substring of a word in another * Speech to text: don't fail the Faster-Whisper model download when preprocessor_config.json is missing - thx oep42 * Speech to text: make online/OpenAI-compatible server failures readable, and send the audio file last - thx acc4github * Restore type-a-number line navigation in the subtitle grid - thx SiriosDev * Fix common errors: show a green check and "Nothing to fix" when a re-scan finds nothing - thx Davanix * Fix auto-imported Matroska tracks staying "Untitled" until the first save - thx Davanix * Keep the selected audio track when going fullscreen or undocking - thx routineCode * Update the OCR fix replace list with new entries - thx diomed * Update the French translation - thx Need74 * Linux/macOS: stop the app data folder from being resolved against the working directory - thx ivandrofly ----------------------------------------------------------------------------------------------------- v5.1.0-rc16 (26th of July 2026) * Add Georgian, Armenian, Luxembourgish and Faroese spell check dictionaries * Update the Qwen3 ASR CPP engine to v0.1.7 - fixes the discrete GPU (Vulkan) crash - thx HanRU01 * Auto-hide the mouse cursor in fullscreen video - thx Davanix * Batch convert: let Delete remove the selected files - thx SiriosDev * Batch convert: keep the ASSA settings between sessions - thx SiriosDev * Fix script alignment desyncing on the second sentence in "Import plain text" align-via-speech-to-text - thx subof * Fix Alt+drag killing all shortcuts, and run alignment shortcuts in the text box - thx torvchen * MS Word spell check: check words in a range so the selected dictionary language is used, and recover a dead Word session - thx sergechaly * Fix OCR "add to user dictionary" not advancing to the next word - thx Rumbah2 * Fix false-positive ¿que -> ¿qué accenting in the Spanish OCR fix list - thx fraternl * Qwen3 ASR: report the exit code and suggest the CPU build when a GPU run crashes - thx HanRU01 * Performance: 10-25x faster auto line-breaking, faster .srt reading, waveform render and subtitle grid, and linear adjust-display-time/set-fixed-duration/remove-empty-lines ----------------------------------------------------------------------------------------------------- v5.1.0-rc15 (25th of July 2026) * Add Indonesian, Vietnamese, Greek, and Thai UI translations * Update llama.cpp to b10103 * Auto-translate via DeepSeek: turn off the new default thinking mode for instant translations again, and migrate the retired deepseek-chat/deepseek-reasoner model IDs - thx itallfelldown5241023 * Export plain text: word wrap the preview, and remember the window position and size - thx subof * Apply the saved volume to all video player windows - thx CDoggyDog62 * Shortcuts window: darker category text in light mode, wider category column and button hints - thx GrampaWildWilly * Subtitle grid menu: keep Styles/Actors at the top and move the dictionary picker down * Allow Escape to close the source view window - thx ivandrofly * Fix undo deleting all lines of the original subtitle in translation mode - thx abasameer * Fix the MS Word spell check engine checking against Word's default language instead of the selected dictionary - thx sergechaly * Fix AltGr shortcut handling racing with typing - thx Ironship * Fix out-of-range parsing of culture names with unbalanced parentheses - thx ivandrofly * Performance: micro-optimize hot UI helpers, and deserialize JSON directly from streams - thx ivandrofly * seconv: fix VobSub OCR isolation, MKV VobSub passthrough, durations and silent drops - thx albino1 * seconv: fix dump-settings mangling non-ASCII characters on Windows - thx Hlsgs * seconv: use canonical kebab-case flags in list-fce-rules output - thx Hlsgs ----------------------------------------------------------------------------------------------------- v5.1.0-rc14 (24th of July 2026) * Complete overhaul of the French translation - thx Need74 * Update the CrispEmbed OCR runtime to v0.16.1 * Improve language auto-detection for short subtitles * Waveform: hold the playhead frozen after pause, and keep the cursor centered on small paused seeks - thx Davanix * Fix Space no longer toggling play/pause after clicking the video player play/pause button - thx Davanix * Fix auto-save silently stopping after an import (converted flag was sticky) * Fix MOSS-TTS cloning random voices instead of the chosen reference voice, and apply the same fix to the other CrispASR cloning engines - thx subof * Strip subtitle formatting tags before TTS synthesis - thx subof * Fix image format conversion issues in batch mode - PGS alpha and BDN XML input - thx Lukewarm1141 * Keep the start dash when a dialog line follows a bracket or parenthesis - thx ivandrofly * Handle non-ASCII dashes in "Remove dash in first line in non dialogs" - thx fraternl * Better encoding detection via UtfUnknown's best match - thx ivandrofly * Surface 7-zip unpack failures instead of hanging the download dialogs * Flatpak: bundle 7zr and prefer a system 7-zip on Linux * seconv: apply the .idx palette when decoding VobSub, and accept .idx files as input - thx albino1 * seconv: gate language-specific fix-common-errors rules by subtitle language, with a --fce-language override * seconv: map fix-common-errors rule IDs to their GUI labels, add dump-settings, restore --settings and --apply-min-gap docs, and fix a --help crash * seconv: fix four bugs found in a targeted bug hunt * macOS: docs and release notes now reflect that the app is signed and notarized ----------------------------------------------------------------------------------------------------- v5.1.0-rc13 (23rd of July 2026) * llama.cpp: engine settings dialog showing backend, install status, pinned release and install folder - reachable from auto-translate, AI review and the AI assistant * Show llama.cpp install status dots in the AI review and AI assistant engine combo boxes * Start-up: non-English languages load about 3.5x faster - source-generated JSON, and the translation file is no longer read twice * Build the subtitle format list off the critical path, cutting further start-up time * Speed up MP4/MOV parsing up to 2.3x on large files, and fix two sample table faults * OCR: speed up two hot loops * Fix Traditional Chinese missing from the UI language list - thx SunnyLeu * Fix message box icons never loading - they now also follow the active theme instead of always using Dark * Waveform themes: report import/export errors instead of failing silently - an unwritable export path reported success while writing nothing * Retry asset unpacking after a failed first run, instead of marking the folder up to date * Fix alternating row colors not updating when the system theme changes - thx pdjdev * Add the missing help text for OpenAI-compatible server * Remove the dead Parakeet.cpp engine * Faster SHA-256 hashing and byte parsing via one-shot HashData and BinaryPrimitives - thx ivandrofly * Fix the waveform playhead drifting slightly forward or backward just after pausing - thx Davanix * Stop and dispose dialog timers on every window-close path, fixing a leak and a rare close-time crash - thx ivandrofly * OCR: keep the preview checkerboard backdrop out of the character grid * macOS: strip a stray .gitkeep from the app bundle so code signing succeeds ----------------------------------------------------------------------------------------------------- v5.1.0-rc12 (22nd of July 2026) * Auto-translate/AI review via llama.cpp: add Gemma 4 (E4B/12B, 140+ languages), Qwen 3.6 35B-A3B (mixture-of-experts - fast on CPU) and Qwen 3.5 9B Q8_0 * Update CrispASR to v0.8.20 and offer three Linux GPU builds * Offer the translated file name when saving after auto-translate, instead of the video file name * Default exports to the source file's folder - Matroska/MP4 track pickers, main window export and image export * Reopen the original subtitle on start-up too, not just the translated one - thx radektuma * Extend to shot change: fix the previous/next commands shortening the line instead of extending it * Fix garbled import of grayscale images, e.g. from InpaintDelogo - thx Codling * OCR: checkerboard backdrop for image previews, so white-on-white pre-processing output stays visible - thx alexantr * Fix stale buffer reads on truncated segments in Blu-ray sup parsing * Convert actors: stop the preview timer when closing the window with Esc - thx ivandrofly * Sync all language files with English.json - 3,084 added/translated strings across 27 languages * Fix translations missing format placeholders (169 strings, 12 languages) * Update Japanese translation - thx hisui3393 ----------------------------------------------------------------------------------------------------- v5.1.0-rc11 (21st of July 2026) * Remove text for hearing impaired: symbol preset combos for the custom before/after fields (SE 4 parity) * Point sync: Home/End jumps to first/last line * Fix UI freeze when opening a raw PGS .sup (no "PG" headers, e.g. Matroska raw-mode extraction) - now offers OCR import with placeholder timings or advice to re-extract - thx HellbringerOnline * MKV import: show progress for PGS tracks and surface extraction errors - thx lukastribus * Fix menus and flyouts leaving keyboard shortcuts unresponsive until clicking inside the window * Extend to shot change: leave the configured gap before the shot change * Fix light title bar in non-modal windows on dark theme - thx Surlime * Remember position/size for ASSA styles/attachments windows - thx Davanix * Time code box: reject non-ASCII digits and consume full IME commits - thx lucvdv1 * Burn-in: log the ffmpeg command line, and never let ffmpeg wait on stdin * Performance: pooled buffers in Blu-ray sup and transport stream parsing, faster "fix names" in change casing - thx ivandrofly * Update Polish translation - thx potplayer-fanpack * Update Korean translation - thx 12si27 * Update Turkish translation - thx bilimiyorum * Update Italian translation - thx bovirus ----------------------------------------------------------------------------------------------------- v5.1.0-rc10 (20th of July 2026) * New "file saved" prompt - success check mark, file name and folder, size/duration info, copy path and Play/Open - now also used by text to speech and the video tools (burn-in, cut, re-encode, transparent subtitles, blank video, embedded subtitles) * Auto-translate via llama.cpp: add Hy-MT2 translation models (Tencent, 7B/1.8B - excellent for its 33 languages, no Nordic) plus larger TranslateGemma 12B and Qwen 3.5 9B quants (up to 10 GB) * Auto-translate: the URL field was ignored for winstxnhdw/nllb-api, NLLB-serve, Anthropic, Groq, OpenRouter and Nvidia - and nllb-api now tolerates a missing trailing slash in the URL - thx 3ftomi * Text to speech: fix the voice-clone transcript prompt appearing repeatedly and generation using a different voice than picked (MOSS-TTS and sibling clone engines) * Speech to text: live progress (and time remaining) for all CrispASR engines * Update CrispASR to v0.8.15 * Fix Ctrl+F/Ctrl+H (and all other shortcuts) not responding after closing a dialog such as spell check, until clicking inside the window - thx fraternl * Multiple replace: fix regex rules corrupting ASSA \N line breaks on Windows - also covered Batch Convert's replace step and regex Find & Replace - thx Davanix * Burn-in: more compact settings so the window fits smaller screens, and audio settings tidied up * Remember window position/size for fix-list tool windows * Performance: much faster subtitle grid rendering (live spell check, CPS/WPM columns) and faster ASSA file loading * Performance: core-library micro-optimizations - thx ivandrofly * Update Japanese translation - thx hisui3393 * Update Italian translation - thx bovirus * Update Polish translation - thx potplayer-fanpack ----------------------------------------------------------------------------------------------------- v5.1.0-rc9 (19th of July 2026) * Text to speech: add MOSS-TTS (CrispASR) engine with zero-shot voice cloning * OCR: new "Show only forced subtitles" filter and a Forced column - only the forced lines are OCR'ed and returned, like SE4 - thx heniowise * OCR: fix forced-filter edge cases - hidden lines now receive pre-processing/palette changes, deleting a visible line no longer wipes hidden lines' unknown words, delete is blocked while OCR runs, and the VobSub color chooser previews the correct image * Multiple replace: fix phantom empty line (and stray red mark) in the preview's Before column with regex rules on Windows - thx Davanix * Text to speech: harden the window's close path and add diagnostics for a reported crash when closing with OK - thx cvrle77 * Text to speech: an audio playback failure in the review window no longer leaves all play/regenerate buttons dead * Speech to text: update CrispASR runtime to v0.8.13 - the linux-arm64 build is also back * Speech to text: single-mode transcribe no longer picks up a leftover queued file - and no longer wipes the batch queue * Auto-translate: fix languages with underscores in the engine name never matching (e.g. Google Translate V1 Chinese) and the dialog getting stuck on "Translating..." * Embedded subtitles (mkv): the dialog no longer locks up after a failed generate * Burn-in/transparent video: cutting now burns the correct lines, keeps ASSA styles/header, and includes lines straddling the cut boundary * Batch convert: "Open containing folder" now opens the file's folder on Windows when the path contains spaces - thx cyberbrix * Fix parser edge cases: SAMI crash on malformed files (broke format auto-detection), WebVTT cues with mixed short/long timestamps being dropped, TTML fractional seconds, MicroDVD lines with two empty time codes * SeConv: auto-detect the subtitle language for MP4/MKV container tracks with no declared language (or "und") - thx ivandrofly * Update Chinese Traditional translation - thx love80312 * Update Russian translation - thx jekovcar * Update Bulgarian translation - thx jekovcar * Update Polish translation and interjections - thx potplayer-fanpack * Update Turkish translation - thx bilimiyorum ----------------------------------------------------------------------------------------------------- v5.1.0-rc8 (18th of July 2026) * Waveform: add SE4-style Previous/Play/Pause/Next text buttons to the waveform toolbar * Subtitle text box: disable ligatures - letter pairs like "fi" no longer merge into one glyph, so the caret and mouse selection treat each character separately - thx subcor * Subtitle text box: roll back rc7's right-to-left ASSA tag rendering fix - clicking inside a tag in a right-to-left line could crash; the misplaced backslash is cosmetic only and the fix will return once Avalonia can render it safely - thx muaz978 * ASSA: fix missing closing tag when setting font name on selected text - thx rRobis * Fix common errors: "Fix invalid italic tags" no longer eats the character before a dangling begin tag and now repairs more two-line stray-tag shapes * Cavena 890: fix Chinese write path so text and italics round-trip * Fix six silent text-corruption edge cases (nested closing tags, UTF-16 detection, "Fix uppercase 'i' inside words", Turkish casing after colon, Cavena 890 Hebrew italics, DVD Studio Pro bold+italic) - thx ivandrofly * Harden the binary subtitle parsers against malformed files (crashes, out-of-memory, infinite loops) - thx ivandrofly * Fix image edge cases in transparent-image detection and rectangle copying - thx ivandrofly * Network: fix proxy password decoding, a crash when no proxy is configured, and auto-translate request issues - thx ivandrofly * Performance: three core-library micro-optimizations * SeConv: malformed Scenarist SCC input no longer converts "successfully" to corrupt output - files where nothing can be decoded now fail with a clear error and a non-zero exit code, and partially damaged files convert with a warning (console and --json) - thx lthibault-stingray * Update Korean translation - thx 12si27 * Update Italian translation - thx bovirus ----------------------------------------------------------------------------------------------------- v5.1.0-rc7 (17th of July 2026) * Subtitle grid: "Insert subtitle after current line..." now works on any line, not only the last - and the inserted file keeps its own time codes like SE4 (only shifted when it would overlap the selected line), so pre-timed subtitles land at their real positions - thx UUMARTIN * Waveform: "Insert subtitle file at video position..." is now available anywhere on the timeline (was only past the last subtitle) - inserts at the right-clicked position and warns before creating overlapping lines - thx UUMARTIN * Speech to text: the input language is remembered again like SE4 - restored on window open, kept when changing engine/settings, and no longer overridden by auto detection in "Selected lines - Speech to text" - thx UUMARTIN * Subtitle text box: restore live spell check with "Color tags" enabled - red underlines and the context menu suggestions/add to dictionary/ignore all work again in the new syntax highlighting text box * Subtitle text box: fix ASSA override tags rendering with a misplaced backslash in right-to-left lines - "{\an8}" showed as "{an8\}" - thx muaz978 * Update Polish translation and add Polish names list - thx potplayer-fanpack ----------------------------------------------------------------------------------------------------- v5.1.0-rc6 (17th of July 2026) * Waveform: new option "Center video position also while paused" (Settings - Waveform/spectrogram) - restores SE4's center/lock mode: with "Mouse-wheel sets video position" on, scrubbing keeps the cursor centered so the waveform scrolls as one continuous strip, and "select current subtitle" also applies while scrubbing paused - thx Asztec00 * Open video from URL: the "Download yt-dlp" button is now only shown when an update is actually available - it was always visible, even with the latest yt-dlp installed * Open video from URL: add "Save..." button to the "Pick subtitle to download" window - saves the selected downloaded subtitle file as-is to a location of your choice * MP4 track picker: the "Export..." button now works - text tracks save as SubRip, VobSub image tracks export to Blu-ray .sup ----------------------------------------------------------------------------------------------------- v5.1.0-rc5 (17th of July 2026) * Subtitle text box: tag coloring (HTML/ASSA) now uses a new native text box - same colors as before, but centering, IME/CJK input, right-to-left and the context menu work exactly like the normal text box, and toggling right-to-left no longer turns off tag coloring * Subtitle text box: fix crash while typing an unterminated attribute quote (e.g. Export and the image/binary subtitle editor) * Command line (seconv): --multiple-replace now also accepts the multiple-replace rules exported from the GUI (.template JSON or .csv), not only the legacy XML - thx BagronkeN * Shortcuts: a user-assigned F10 shortcut is no longer overridden by the menu-bar focus toggle (the menu stays reachable via Alt) - thx Khaztaroth * Fix reading decimal values in some XML-based formats (e.g. Final Cut Pro, JacoSub) on locales with a comma decimal separator * Performance: more core-library speedups (fix common errors, time code parsing, tag/RTF/casing text handling) ----------------------------------------------------------------------------------------------------- v5.1.0-rc4 (16th of July 2026) * Live spell check: only show the subtitle grid spell check items (suggestions, add to user dictionary, ignore all) when a single line is selected * Waveform: fix "Seek silence" jumping far past the actual silence when the waveform was scrolled or zoomed - thx AdiOpi * Shortcuts: the search now accepts modifier keys in any order with optional spaces around "+" (e.g. "shift+control" or "ctrl+r"), and shortcuts always show their keys in canonical order (Control, Alt, Shift, Win) - also in the duplicate-shortcut warning - thx GrampaWildWilly * Generate transparent video: live video preview like burn-in - the styled subtitles render over the playing video while settings are tweaked - thx MbuguaDavid * Generate transparent video: fix the selected effects not being applied to the generated video, and compact the cut panel * Performance: faster shortcut key lookup and HTML tag stripping (subtitle grid, characters-per-second and per-keystroke text info) * Subtitle grid: add "Remove" to the actor context menu (also available as shortcut), so the actor can be cleared from the selected lines without creating a dummy actor - thx Khaztaroth * Waveform: the "Seek video backward/forward" toolbar buttons can now be assigned keyboard shortcuts - the assigned keys show in the button tooltips and also work in the full screen video window - thx Surlime * Shortcuts: all columns are now sortable - Active in, Category, Name and Shortcut - and sorting stays active while filtering by category or search - thx GrampaWildWilly * Drag and drop now works on Linux (Avalonia UI 12.1.0 fix) * Settings: colored section icons in the menu and section headers, matching the Shortcuts window look * Batch convert: show the status column as colored badges - green for converted, red for errors, gray for cancelled * Word lists: colored panel icons and matching count badges - also fixes the near-unreadable badge text in light theme * Batch convert: add llama.cpp OCR engine with local OCR vision models (GLM-OCR, LightOnOCR, PaddleOCR-VL) and automatic engine/model download - thx go2tom42 * Command line (seconv): add llama.cpp OCR engine (--ocr-engine:llamacpp with --ocr-model/--ocr-url), starting the local server automatically - thx go2tom42 * Waveform: fix the video player ending up both docked and undocked after closing the waveform toolbar settings while the audio visualizer was undocked - thx GrampaWildWilly * Update CrispASR to v0.8.11 * Update llama.cpp to b10035 ----------------------------------------------------------------------------------------------------- v5.1.0-rc3 (15th of July 2026) * Export: add "DVD sup (MuxMan/Scenarist)" image-based export - the classic DVD-Video subpicture .sup that DVD authoring tools import * Spell check: start at the current line instead of asking "Continue spell check from current line?" - the question is now only asked when an earlier spell check was left unfinished * Names: load the region specific names list - the Portuguese lists (pt_PT_names.xml and pt_BR_names.xml, ~4800 names each) were never read, as only a neutral pt_names.xml was looked for * Names: add former countries like Yugoslavia, Czechoslovakia and Persia to the English, Dutch, German, French, Spanish and Portuguese names lists - thx fraternl * Spell check: fix the spell check window opening blank - no word, no text and no suggestions - unless the first line was selected * Live spell check: add spell check items to the subtitle grid context menu - suggestions, add to user dictionary and ignore all * Live spell check: picking a dictionary now also updates the subtitle grid, and no longer hides the spell check items in the text box context menu * Live spell check: fix cancelling "Pick spell check dictionary" switching the dictionary anyway, and the list always opening on English instead of the dictionary last used * Live spell check: fix "Ignore all" being forgotten again as soon as the settings window was closed * Live spell check: fix the English dictionary being used for other languages - thx muaz978 * Subtitle grid: fix an arrow key after shift+arrow jumping back to the anchor line instead of continuing from the moving end of the selection * Undo: fix automatic change detection reading the subtitles from a timer thread, which could silently drop an undo entry or record a state that never existed * Scroll bar: shift+click on the trough jumps to that position - thx muaz978 * Shortcuts: normal weight for the shortcut name column - thx muaz978 * Sync: add "Visual sync" to the subtitle grid context menu under "Selected lines", so segments that are off by different amounts can be synced one at a time - thx Hikari2w2 * OCR: "Save as" for an OCR'd subtitle opens in the source file's folder again, instead of e.g. Documents or the last-used folder - thx Turcote/fraternl * Replace: fix regex replace of a line break (search \n) leaving the break in place with a space after it on Windows-style (CRLF) text - thx witchpiggie ----------------------------------------------------------------------------------------------------- v5.1.0-rc2 (14th of July 2026) * Auto-translate: update llama.cpp to b9993 * Auto-translate: fix chat template for self-supplied .gguf models in the llama.cpp engine - thx saneryiyiersansi * Fix common errors: "Fix common OCR errors" no longer applies fixes you unticked in the list - thx fraternl * Fix common errors: unknown-word guessing (which splits words) is now off by default, with its own check box - thx fraternl * Fix common errors: clearer step 2 selection - Select all/none toggle, live selected-counts on the Apply button and category chips - thx muaz978 * Fix common errors: make Apply/Action/Before/After columns sortable in the fix list - thx subcor * seconv: warn about unknown keys in the --settings JSON instead of silently ignoring them - thx shanecoopt * seconv: --apply-min-gap can now be used without a value, taking the gap from minimumMillisecondsBetweenLines - thx shanecoopt * seconv: fix minimumMillisecondsBetweenLines of 0 giving a 1 ms gap when bridging gaps - thx shanecoopt * Check for updates: no longer offers an older version as an update * OCR: update CrispEmbed to v0.15.0 - fixes garbled math-OCR output, faster detection/decoding * Save the language file with LF line endings on every OS, so re-saving it no longer produces a whole-file diff * Update Korean translation - thx 12si27 * Update Japanese translation - thx hisui3393 ----------------------------------------------------------------------------------------------------- v5.1.0-rc1 (13th of July 2026) * Speech to text: add MOSS-Transcribe-Diarize engine (CrispASR) - transcription with speaker labels like "(Speaker 1)" in a single pass - thx Oplay66 * Speech to text: add Arabic and Japanese models to the CrispASR Cohere engine - thx muaz978 * Speech to text: update CrispASR to v0.8.10 - fixes long-audio repetition/empty output, much faster long audio on CUDA * Add find-text history dropdown, text-length sorting, and sort shortcuts - thx mrklhsnbrg * Open video from URL: suggest the video's real title as download file name instead of e.g. "watch.mkv" * Update yt-dlp to 2026.07.04 * Auto-translate: update model suggestion lists across engines * Shortcuts: tighten category tiles, add space below, and keep the scrollbar off the shortcut key column * Performance: allocation-free nOCR pixel matching (faster batch image-based OCR) and skip redundant waveform buffer sorting * Performance: remove constant background allocations in change detection and character-count lookup * ElevenLabs: fix SSML break tags being ignored/spoken by sending enable_ssml_parsing - thx cvrle77 * ElevenLabs: fix v3 stability/language settings being silently ignored, show engine error details in test voice, and sort the turbo v2.5 language list * ElevenLabs: revert apply_text_normalization=off and log request bodies to the tools log - thx cvrle77 * Fix alignment shortcut not toggling, OCR adding {\an2} tags to every line, and add Ctrl+I italic in the OCR text box - thx qvimse * Fix common OCR errors changing valid words like Dutch "Al" to "AI" * Fix vertical scroll bar overlapping the outermost text column in RTL/translation mode - thx muaz978 * Fix scroll bar trough click not paging in grids - thx muaz978 * Fix common errors: "Select all"/"Invert selection" now respect the active category chip - thx subcor * Fix download retry corrupting the file when the server does not support resume * Fix shortcuts editor issues: stale key chips after edit/reset, cancel not reverting imports/reset, numpad keys dropped, duplicate detection missing Ctrl/Control aliases, and modifier-only bindings not resettable * Fix auto-backup racing UI edits (snapshot now taken on the UI thread) * Fix ffmpeg path being ignored for waveform extraction on macOS * Fix gap column showing "0,000" on the last line * Fix AI review default shortcut colliding with Multiple replace * Fix "Check for updates" always reporting it was unable to check - thx nms42 * Fix point sync via other subtitle not loading a re-selected subtitle file - thx fraternl * Fix truncated "Number" and "Show" columns in the auto-translate grid - thx ivandrofly * Add Persian translation - thx rmtjokar * Update Korean translation - thx 12si27 * Update Polish translation - thx Adam Malich * Update Chinese Simplified translation - thx wuwufei * Update Italian translation - thx bovirus * Update Turkish translation - thx bilimiyorum * Update Russian translation - thx jekovcar * Update Bulgarian translation - thx jekovcar * Update Portuguese (Brazil) translation - thx igorruckert ----------------------------------------------------------------------------------------------------- v5.1.0-beta15 (12th of July 2026) * Shortcuts: add categories with icons (File, Video, Sync, Translate, Search, Tools, AI, ...) with filter tiles + "Active in" column - thx hisui3393 * OCR: add CrispEmbed engine - local ggml-based OCR (same family as CrispASR) with GLM-OCR, GOT-OCR2, and Qwen3-VL-2B backends * OCR: add PaddleOCR-VL 1.6 (109 languages) to the llama.cpp model list * Command line (seconv): add auto-translate with llama.cpp, starting the local server automatically - thx go2tom42 * Rename "FCP/png" export to "Final Cut Pro + image" * Speech to text: offer re-download when the engine executable has gone missing - thx zezartiti * Speech to text: fix whisper.cpp (CPU) engine not starting - strip the "Release" subfolder when unpacking - thx zezartiti * Command line (seconv): fix garbled PGS/DVB-sub OCR by binarizing before Tesseract - thx albino1 * Fix thin outlines (width 1-2) disappearing on horizontal edges in image-based exports like BDN XML/Blu-ray SUP - thx pepelugil * Fix ASSA/SSA styles editor not switching to "Styles in file" when clicking the already-selected style row * Update CrispASR Qwen3 model list (re-add qwen3-asr-1.7b q4_k) * Fix Turkish translation not loading (broken JSON) * Add new word replacements to the Croatian OCR fix list - thx diomed * Update Windows installer languages for Inno Setup 6.7.3 - thx Blackspirits * Update Japanese translation - thx hisui3393 * Update Italian translation - thx bovirus * Update Russian translation - thx jekovcar * Update Bulgarian translation - thx jekovcar ----------------------------------------------------------------------------------------------------- v5.1.0-beta14 (10th of July 2026) * Add PGS position monitor to the image-based subtitle editor - a "Position map" - thx Victory61 * Auto-translate: add generic OpenAI-compatible engine (custom URL/model/API key) * AI assistant: engine settings in the window, strip think blocks, and show the model's reasoning behind an info button * Add AI assistant to the subtitle text box context menu * Add shortcut to toggle subtitle grid formatting mode * Text to speech: ElevenLabs pause cues work again - thx cvrle77 * Text to speech: fix Google voices with 3-letter language codes (Mandarin/Cantonese/Filipino) failing to generate audio * Text to speech: the ElevenLabs "Speaker Boost" setting is now actually sent to the API * Fix waveform border drag not registering on trackpads (macOS) - thx okpukhalska-dot * Fix malformed ASSA tags when splitting a line with two or more active tags (e.g. bold + color) * Fix TTML/DFXP export losing an enclosing style after a nested tag (e.g. bold inside italic) * Fix subtitle format edge cases: MicroDVD "{y:u,i}" tag, uppercase \uXXXX in JSON * OCR: better separation of vertically touching text lines (two line-splitter fixes) * Fix ASSA/SSA properties video resolution and dark-theme picker color - thx kit132 * Re-apply the layout direction live when the interface language changes - thx muaz978 * Update CrispASR Qwen3 and parakeet-ja model lists * Add Turkish UI translation - thx bilimiyorum * Update Italian translation - thx bovirus * Update Korean translation - thx 12si27 * Update Danish translation ----------------------------------------------------------------------------------------------------- v5.1.0-beta13 (9th of July 2026) * Add an AI assistant button to the subtitle text box - quick actions (fix spelling/grammar, reading speed, change tone) * OCR: add histogram-based colour isolation for VobSub/DVD subtitles - rebuilds each subpicture as crisp black-on-white - thx tekk42 * Restore the last editing position when reopening a subtitle file - thx hisui3393 * Command line (seconv): image output styling for Blu-ray sup/VobSub/BDN-XML etc. - new options - thx Fredrik * Command line (seconv): fix double line spacing in image-based output, and do not warn that --resolution is ignored for image targets (it sets the canvas size) * MLX Whisper (macOS): forward advanced command line options (e.g. --condition-on-previous-text False) - thx muaz978, asiaminor77 * Fix exported Blu-ray sup being rejected by BDSup2Sub ("Subpicture too large") when a subtitle line was empty - thx Victory61 * Beautify time codes: fix subtitle blocks missing from the change preview waveform - thx pdjdev * Right-to-left: do not mirror the video and waveform - thx muaz978 * Add Arabic UI translation - thx muaz978 * Update Japanese translation * Update Korean translation - thx 12si27 * Update Italian translation - thx bovirus * Update Russian translation - thx jekovcar * Update Bulgarian translation - thx jekovcar * Update Danish and Swedish translations (machine-translation cleanup + missing strings) ----------------------------------------------------------------------------------------------------- v5.1.0-beta12 (8th of July 2026) * macOS: add multi-window support - "File -> New window" opens an independent editor window, with a per-window menu bar and Dock integration - thx muaz978 * Add "Set start and go to next" shortcut (Subtitle Edit 4 parity) - thx NiggleHub * Text to speech: the Review window's OK now applies its subtitle changes to the main window (Cancel discards) - thx cvrle77 * MLX Whisper (macOS): chunked decoding with punctuation persistence, beam search, and subtitle-standard cues for better long-recording quality - thx muaz978 * Speech to text: stream Python engine (Whisper CTranslate2 / MLX Whisper) output live instead of only after the process finishes - thx muaz978 * Speech to text: show live segments for the Whisper CTranslate2 engine (previously looked frozen) - thx muaz978 * OCR: make word un-compositing fully optional - the "OCR: use word split list" toggle now disables all merged-word splitting - thx fraternl * Waveform: refresh the context menu after changing the UI language (no restart needed) - thx pdjdev * Fix "Fix common OCR errors" splitting valid closed compounds in compounding languages (e.g. Dutch "modderwarboel") - thx fraternl * Fix app closing anyway when "Save as" was cancelled during the close prompt (losing unsaved work) * Fix a SubtitleEdit process staying alive in the background after closing the main window (which locked its folder) - thx KaDeeKe * Fix black video preview in the burn-in and visual sync windows until the window was resized - thx MbuguaDavid * Show the MKV progress window for multi-track files (not just single-track) - thx lukastribus * Export image-based (Blu-ray sup): keep the profile's Resolution on reopen - thx HolgerKuehn * Speech to text: fix a Latin period being appended after non-Latin punctuation (e.g. Arabic/Urdu) in the "add periods" step - thx muaz978 * MLX Whisper (macOS): detect pipx/venv/conda installs and show a clearer "not found" message - thx asiaminor77 * Online speech to text: log failures, add a DashScope region hint, timeout handling, and smaller OpenRouter chunks - thx caiweihan, acc4github * Whisper CPP: fix incorrect model file sizes shown in the model list - thx asiaminor77 * Multiple replace: import replace rules directly from a Subtitle Edit 4 Settings.xml * ASSA karaoke (advanced effects): add a right-to-left option - thx rzeczywistoscia992-web * Right-to-left: apply content direction to the translation text boxes and mirror the edit grid columns (Subtitle Edit 4 parity) - thx muaz978 * Waveform: freeze the cursor on all pause paths and ease small position corrections after pause/stop instead of snapping * Fix Apply and Find text buttons missing in "Point sync via other subtitle" * Fix a message box appearing behind an undocked audio visualizer * Fix broken French and Swedish spell-check dictionary download URLs * Update CrispASR to v0.8.9 * Update Japanese translation * Update Korean translation - thx 12si27 * Update Italian translation - thx bovirus * Update Turkish translation - thx bilimiyorum * Update Russian translation - thx jekovcar * Update Bulgarian translation - thx jekovcar ----------------------------------------------------------------------------------------------------- v5.1.0-beta11 (4th of July 2026) * Fix settings not saving when the UI language had a duplicate/empty translation (e.g. Portuguese Brazil) - thx rosilucia-hub * Fix ASSA "Set position" and "Image color picker" showing a blank video background when the "HH:MM:SS:FF" time code format was enabled - thx bichitoxxx * Restore the WebVTT "X-TIMESTAMP-MAP" setting (Options -> Settings -> Subtitle Formats) to optionally ignore the header offset that shifts all time codes on load - thx anakinsleftleg * Fix Ctrl+End / Ctrl+Home on the subtitle grid not scrolling the last/first line fully into view - thx KaDeeKe * Open video file dialog: add a "Video and audio files" filter as the default so audio files (e.g. mp3) can be picked without switching the filter * Waveform: add a customizable "<< seek >>" panel to the toolbar - thx BlazDT * Waveform: add "Configure toolbar items..." to the More menu * Open a .mkv without subtitles as a video file - thx KaDeeKe * Text to speech: apply Review window text edits back to the subtitle - thx cvrle77 * Text to speech: fix empty language list in the Review window - thx cvrle77 * Update Portuguese (Brazil) translation - thx rosilucia-hub * Update Korean translation - thx 12si27 * Update Italian translation - thx bovirus ----------------------------------------------------------------------------------------------------- v5.1.0-beta10 (4th of July 2026) * Fix online speech-to-text (OpenRouter / Alibaba Qwen3-ASR) settings layout - the engine's settings rows no longer overlap/jumble together * OpenRouter speech-to-text: split the transcript into multiple timed sentences instead of one long line, and run post-processing using the engine's language hint ----------------------------------------------------------------------------------------------------- v5.1.0-beta9 (4th of July 2026) * Add OpenRouter and Alibaba Qwen3-ASR online speech-to-text engines to Audio to text - OpenRouter reaches Whisper/GPT-4o-transcribe/Groq/Chirp with one key; Alibaba Qwen3-ASR (DashScope) uploads the audio and transcribes asynchronously with sentence timestamps * Google Lens Sharp: fix scrambled word order on multi-line subtitles - the text chunks Google returns interleaved between the two lines are now re-ordered top-to-bottom/left-to-right using their bounding boxes - thx fraternl * OCR: fix the Dictionary combo box becoming empty after downloading/re-downloading a spell check dictionary * Remove redundant "Repeat previous line"/"Repeat next line" shortcuts - use "Play previous (and loop)"/"Play next (and loop)" instead (they also handle multiple selections and no selection) - thx abc16361 ----------------------------------------------------------------------------------------------------- v5.1.0-beta8 (3rd of July 2026) * Fix "Fix common OCR errors" doing nothing when no Hunspell dictionary is installed for the language - the OCR replace list (e.g. spa_OCRFixReplaceList.xml) is now applied on its own again - thx perkesfurmah-collab * Fix UI freeze when switching to layout 12 (no video pane) with a video loaded * Text to speech: review window - Space no longer clicks OK on key release either (Avalonia raises a focused button's click on Space key-up, independent of the tunnel-stage fix in beta7) - thx cvrle77 * Text to speech: fix the Import/Export/merge-output file dialogs and the TTS window dropping behind the main window on Windows - thx cvrle77 * Text to speech: Import -> OK now shows progress (bar, status text, Cancel) while merging instead of looking idle - thx cvrle77 * Text to speech: Export now writes audio clips into a "wav" subfolder instead of next to SubtitleEditTts.json; imported sessions no longer use the export folder as scratch space - thx cvrle77 ----------------------------------------------------------------------------------------------------- v5.1.0-beta7 (3rd of July 2026) * Add subtitle grid text fit setting: clip / wrap / ellipsis (Settings -> Appearance) - thx muaz978 * Text to speech: pimp up the window UI - context line (line count, video file name), section headers, engine description, voice count badge, checkbox hints, progress percent/ETA, and a primary accent Generate button * Text to speech: fix Import not doing anything on OK - it now merges/saves/adds to video just like Generate - thx cvrle77 * Text to speech: review window - Space no longer triggers OK, R regenerates the selected line, and playing a line selects it - thx cvrle77 * Text to speech: mask the API key by default with a reveal button (also applied to Auto-translate's API key) - thx cvrle77 * Fix a codebase-wide sweep of 46 confirmed bugs (crashes, wrong behavior, and dead error-handling branches across import/export, translation, spell check, ASSA/SSA styles, fix common errors, and more) * Update Russian translation - thx jekovcar * Update Turkish translation - thx bilimiyorum ----------------------------------------------------------------------------------------------------- v5.1.0-beta6 (3rd of July 2026) * Add SE4's "Column" list view shortcuts (delete text, delete text and shift up, insert text, paste, text up/down) - thx SiaL8er * Text to speech: fix stuck/silent generation - failures now show an error, cancel stops the running ffmpeg, and rate-limited ElevenLabs requests retry with backoff - thx cvrle77 * Text to speech: fix export/import round-trip (import file filter, portable relative paths, missing-file handling) - thx cvrle77 * Text to speech: large bug-fix pass across the engines and the review window (60+ fixes - voice-list refreshes no longer wipe cached voices, Azure/Piper voice refresh, Murf audio corruption, per-actor cast fixes, review playback/regenerate state fixes, the voice instruction is now applied when generating, Qwen3 model choice is persisted, and more) * Text to speech: play/stop the review window's selected line with the play/pause shortcut (Space) - thx cvrle77 * Text to speech: the review window's Regenerate now offers the engine/model download prompts * Accessibility: fix Alt not activating the menu bar (e.g. with NVDA), and name splitters, waveform and Settings controls - thx Shaima-Almarzooqi * Make tool-dialog buttons consistent and fix "Remove text for HI" undo - thx mjuhasz * Fix seconv OCR (Tesseract/PaddleOCR) garbling non-ASCII output on Windows - thx Gorkycreator * Update Bulgarian translation - thx jekovcar * Update Czech translation - thx Matěj * Update Japanese translation - thx hisui3393 * Update Korean translation - thx 12si27 * Update Russian translation - thx jekovcar * Update Turkish translation - thx bilimiyorum ----------------------------------------------------------------------------------------------------- v5.1.0-beta5 (2nd of July 2026) * Add AI review (Tools > AI review) - proofread subtitles with a local LLM via llama.cpp or Ollama, with a before/after grid and editable prompt * Add llama.cpp translation to Batch convert (auto-detect/start/download) * Add interjection lists for German, Italian, Swedish, Norwegian and Dutch * Add copy/paste support to the time code boxes (a bare number is milliseconds) - thx RobertoHN * Waveform: show the duration inside the new-selection rectangle while dragging - thx geroyannis * Open the source view with the caret at the selected subtitle - thx pdjdev * Settings: "Color text if more than N lines" now tracks "Max number of lines" live - thx subcor * Accessibility: fix the text-editor Tab focus trap, name status-bar icons, and label Settings/Shortcuts/AutoTranslate/FixCommonErrors/SpeechToText controls - thx kylesskim-sys * Polish Fix common errors, Netflix errors, Split/re-balance long lines, Compare and Restore auto-backup (filter chips, colored tags, word diffs, live counts, icons) * Fix ChatGPT auto-translate failing with "This is not a chat model" (bad default model) - thx EspritIT * Speech to text: add smaller model quants for FunASR nano/MLT-nano and Parakeet Japanese, and fix dead MADLAD download links * OCR: select the just-downloaded spell-check dictionary on any UI language * Spell check: clean up dictionary names with variant suffixes (e.g. de_DE_frami) * Fix keyboard shortcuts in "Adjust all times" not working, now moving 10/100/500 ms - or 1 frame / 10 frames / 1 second in frame mode - thx KaDeeKe * Fix up/down time code and duration stepping in frame mode not wrapping frames correctly - thx KaDeeKe * Fix invalid PTS/DTS in exported Blu-ray sup files (broke SupMover negative delay) - thx GCRaistlin/MonoS * Fix batch convert forgetting settings when the window is closed via X/Escape - thx MonsterSe7en * Fix batch convert llama.cpp auto-start never running * Fix crash when pasting an out-of-range value into a time code box * Fix stale RTL text-box selection anchor after mouse/native selection changes - thx muaz978 * Fix change detection not including the original subtitle file name * Fix macOS legacy UI font defaults migration - thx mjuhasz * Fix UI tests overwriting user settings - thx mjuhasz * Add discard assignment to DataGridCheckboxMultiSelect instantiations - thx ivandrofly * Update Polish translation - thx potplayer-fanpack * Update Italian translation - thx bovirus * Update Korean translation - thx 12si27 ----------------------------------------------------------------------------------------------------- v5.1.0-beta4 (1st of July 2026) * Add CSV import/export to Multiple replace - thx faon-92 * Add a re-usable "Apply" button to Multiple replace (SE4 parity) - thx KaDeeKe * Add a "New" button to the Fix common errors rule profiles window - thx MbuguaDavid * Spell check: show a completion summary (changed/skipped/correct/names/added) with a "do not show again" option (SE4 parity) - thx cyberbrix * Fix common errors: restore the keyboard flow so Return advances and applies fixes (SE4 parity) - thx KaDeeKe * Focus the subtitle grid after extracting a subtitle from an mkv, so Ctrl+S works right away - thx KaDeeKe * Fix icon-button hints not showing in dialogs on macOS * Fix Google Lens OCR including server retry messages in the OCR text - thx LurkingNinja * Fix bare Alt not activating the main menu bar on Windows * Fix small custom-millisecond video seeks being unreliable on a single key press - thx AndryOut * Fix Option+Arrow word navigation in the color-tag / source-view text editors on macOS - thx mjuhasz * Fix Escape closing the Binary edit window when a dialog was opened from its context menu - thx mjuhasz * Fix the status dot and description in the "Get dictionaries" window - thx ivandrofly * Fix garbled Korean/CJK text in the Linux Flatpak build (bundle Noto Sans CJK) * Fix caret misplacement on macOS by defaulting the UI font to Helvetica Neue - thx mjuhasz * Fix the waveform cursor lagging/jittering when pausing playback - thx AndryOut ----------------------------------------------------------------------------------------------------- v5.1.0-beta3 (30th of June 2026) * Add DeepL formality picker (SE4 parity) - thx fraternl * Add ASSA "Replace style with..." (SE4 parity) * Add "Apply" button to "Remove text for hearing impaired" - thx fraternl * Remove text for hearing impaired: show count of lines found - thx bptrsn * Add editable whitelist for "Remove if all uppercase" - thx muaz978 * Add a text box for the grid single-line separator (Settings) - thx bichitoxxx * Show the grid single-line separator in a distinct color * Auto-translate (llama.cpp): add Qwen 3.5 and Aya Expanse models * Enable Qwen3 ASR CPP speech-to-text on macOS * Add CrispASR ARK-ASR speech-to-text backend * Update CrispASR to v0.8.6 * Update llama.cpp to b9833 * Waveform: block overlap when drag-creating a new selection - thx muaz978 * Spell check: recognize hyphenated names from the names list - thx samdoesknow * Fix spell check flagging words wrapped in single quotes (e.g. 'Een) - thx fraternl * Fix dialogs (Settings etc.) opening behind undocked video/audio windows - thx GrampaWildWilly * Fix invalid JSON in AI translate request bodies - thx korsosattila73 * Fix "go to next/previous subtitle from video position" skipping lines - thx muaz978 * Fix stuck shortcut keys when the main window loses focus - thx muaz978 * Fix regex \r\n / \n not matching line breaks in Find and Multiple replace - thx KaDeeKe * Fix "Save as" not converting source-format native tags when changing format - thx wjcarpenter * Fix Visual sync not honoring the selected audio track - thx zoltan-antal * Fix menu: Escape needing a 3rd press and stale menu state after Alt+Tab - thx kylesskim-sys * Fix CrispASR Parakeet ignoring the selected language - thx sergechaly * Fix Qwen3 ASR (CPP) failing with "'0x0D' is invalid within a JSON string" - thx ProFire * Fix unsaved-changes prompt treating a dismissed dialog (close/Escape) as a choice - thx pdjdev * Fix Blu-ray sup export crash on Linux (HarfBuzz symbol clash) - thx CannaraX * Fix umlaut diacritics (Ä/Ö/Ü) clipped in the OCR window on macOS - thx Bastians-Bits * Fix VLC not loading from a path with non-ASCII characters (e.g. Chinese) - thx StandAloof * Use nameof for rule IDs in FixCommonErrorsRunner - thx ivandrofly * Update Turkish translation - thx bilimiyorum * Update Japanese translation - thx hisui3393 * Update Italian translation - thx bovirus * Update Korean translation - thx 12si27 ----------------------------------------------------------------------------------------------------- v5.1.0-beta2 (28th of June 2026) * Update CrispASR to v0.8.5 * Add ASSA style categories - group storage styles (e.g. per project) in the Styles window - thx bichitoxxx * ASSA styles: keep the "Set style" menu in style order instead of sorting alphabetically - thx bichitoxxx * Batch convert: honor the selected custom text format and add the image-based formats in the format picker * Add "Remove text for hearing impaired" and "Save as..." to the grid "Selected lines" menu * OCR: auto-select the spell-check dictionary from the OCR language (SE4 parity) - thx fraternl * OCR: use user OCR-fix replace lists (e.g. xxx_OCRFixReplaceList_User.xml) - thx fraternl * seconv: support "Fix common OCR errors" with bundled English dictionaries * Binary edit: correctness, resource-safety and selection fixes * Make icon-button tooltips honor the "Show hints" appearance setting * Fix text box selection shortcuts (e.g. "Selection to lowercase") being inactive and hidden - thx abc16361 * Fix EBU STL Save/Save As writing an invalid 14-byte file - thx jancisefcik * Fix seconv failing to start due to a SkiaSharp native/managed version mismatch - thx pcspeak * Update Portuguese (Brazil) translation - thx igorruckert * Update Turkish translation - thx bilimiyorum * Update Korean translation - thx 12si27 * Fix "Join subtitles" by time codes not sorting the joined lines by time - thx han8mr8 * seconv: run operations in the given order, repeating each occurrence (e.g. "--fix-common-errors" twice runs two passes) - thx Rouzax * seconv: bundle the Linux native libs (libSkiaSharp/libHarfBuzzSharp) so Fix common errors no longer crashes - thx Rouzax * seconv: expose the Fix-common-errors profile values (min gap, CPS, max lines, dialog/continuation style, ...) in --settings JSON - thx Rouzax * Add option to not auto-generate the waveform when opening a video (click the empty waveform to generate) - thx nugemon * "Set up like Subtitle Edit 4" now shows grid line breaks as "
" on a single row, like SE4 - thx bichitoxxx * Redesign the "Change speed" dialog to match the "Adjust all times" dialog - thx mjuhasz * Fix "Change speed" radio buttons disabled/not respected in the main window - thx mjuhasz * Binary edit: wire up "Change speed", respect the selected-lines radio, and always apply "Change frame rate" to all lines - thx mjuhasz * Binary edit: keep the grid scroll position after "Adjust all times" and "Change speed" - thx mjuhasz * Binary edit: add "Adjust color" for selected lines in the image-based subtitle editor (Tools > Selected lines) - thx mjuhasz * Spell check: show the source image for the current line - auto-attached after OCR, or loaded via the image button left of Done (Blu-ray sup, VobSub, BDN XML, transport stream, Matroska) - thx fraternl * Add two more configurable "move video position custom" forward/back jumps (custom 3 and 4) * Main and binary-edit menu: Alt/F10 toggle menu activation and Escape deactivates it (Windows standard) - thx GrampaWildWilly * Update Polish translation - thx Adam Malich * Update Japanese translation - thx hisui3393 * Update Italian translation - thx bovirus * Update Bulgarian translation - thx jekovcar * Update Russian translation - thx jekovcar * Use asynchronous JSON serialization in the plugin runner - thx ivandrofly * Fix video playback speed reverting to 1.0x after changing settings or layout - thx Hakunamatata67 ----------------------------------------------------------------------------------------------------- v5.1.0-beta1 (26th June 2026) * Add auto-save of the open subtitle file (Options > General) - thx muaz978 * Add remote/external llama.cpp server (URL) option to auto-translate - thx muaz978 * Add HTML index export to binary edit - thx mjuhasz * Add "Toggle translation mode" command/shortcut (SE4 parity) * Add "Play next (and stop)" and "Play next (and loop)" shortcuts - thx nguyenphuc2312 * Add "Play previous (and stop)" and "Play previous (and loop)" shortcuts * Add "Treat words ending in 'in'' as 'ing'" spell check setting - thx RedSoxFan04 * Add Undo button to spell check (SE4 parity) * Add grid context menu "insert subtitle after current line" on the last line * Add grid alternating-row coloring option to Appearance settings * Add a couple of "Extend" shortcuts from SE4 - thx OmrSi * Bridge more SE4 shortcuts in the SE4 settings importer * Open original subtitle into an empty subtitle as a blank translation * Drag a multi-line selection in the waveform to shift all time codes together - thx wuwufei * Spell check can continue from the current line and wrap around (SE4 parity) - thx p1nkyy * F10 focuses the main menu for keyboard/screen-reader access - thx kylesskim-sys * Restore horizontal grid lines in the waveform * Improve Dutch casing rules ('s/'t contractions and IJ digraph) * Batch convert: pluralize the remove-files confirmation for multi-select * Use ffmpeg from the system PATH (v7+) when the download fails * Update CrispASR to v0.8.4 * Update whisper.cpp to v1.9.1 * Fix "Save as" to ASSA using Arial instead of the configured default style - thx korsosattila73 * Fix proxy settings not being saved - thx baoyu0 * Fix Split tool dropping the last part; default output to the source folder * Fix SSA toolbar hints showing "Advanced Sub Station Alpha" * Fix missing window title on the ASSA/SSA style picker * Fix found/selected grid row not always being fully scrolled into view - thx subcor * Fix Bold shortcut bolding the whole line instead of the selected words - thx Khaztaroth * Fix profile editor not showing/saving the chars/sec and words/min values - thx Khaztaroth * Make "Merge continuation lines" work with CJK text (treat like English) - thx acc4github * Change the default "Auto-break text" shortcut from Ctrl+R to Ctrl+Alt+B * Align the CJK merge shortcut label with SE4 wording * Localize "Set up like Subtitle Edit 4" dialog strings * Add an assignable "Waveform insert new selection" shortcut (SE4 parity) - thx geroyannis * Add a Tesseract OCR engine-mode setting (SE4 parity) * Add accessible names to the waveform toolbar controls for screen readers * Burn-in: default the target file size to the source video size, and add per-file target size in batch mode * Extend Dutch casing contractions to apply after a colon and a mid-paragraph period * Fix OCR word colors and subtitle-image background contrast in light mode * Shorten and resize a long video file name in the player so it no longer overlaps the position info * Add "Play from just before text" video shortcut (SE4 parity) - thx GordonLeekt * Add "Move text after cursor position to next subtitle and go to next" (and a "...and play" variant) shortcuts (SE4 parity) - thx kadrimarzouki * Add "Set start and keep duration" shortcut (SE4 parity) * Add "Set end, add new and go to new" shortcuts (SE4 parity) - thx Estevarena * Add "Auto detect" language for the standard Whisper engines - thx l0uev4-lab * Add an "Add custom model" GUI for the Whisper engines - thx l0uev4-lab * Make the waveform "Toolbar items" settings row findable in search * Binary edit: F10 focuses the menu and menu keys no longer close the window - thx kylesskim-sys * Improve CJK UI text rendering and font fallbacks on Linux * Optimize waveform/spectrogram drawing * Clean up speech-to-text temp files via a per-run subfolder - thx l0uev4-lab * Fix 24 fps being read as 23.976 (and 30 as 29.97) - thx Larxen * Fix Find Next/Previous shortcuts while the Find/Replace window is open - thx mjuhasz * Fix "Surround with" placing symbols outside formatting tags - thx mjuhasz * Fix EBU STL export writing a blank Display Standard Code * Fix multi-character custom tags in "Remove text for hearing impaired" - thx l0uev4-lab * Fix Tesseract OCR blanking colored (non-white) text, and surface OCR errors instead of blank lines * Fix Ollama OCR runaway repetition / hang * Fix ASSA "Set position" frame grab and rotation, and the blank preview - thx bichitoxxx * Add Traditional Chinese (zh-Hant) translation - thx love80312 * Add Hebrew and Swedish whisper.cpp models for download - thx darnn * Update Italian translation - thx bovirus * Update Korean translation - thx 12si27 * Update Russian translation - thx jekovcar * Update Bulgarian translation - thx jekovcar * Update Polish translation - thx potplayer-fanpack ----------------------------------------------------------------------------------------------------- v5.0.0 (22nd June 2026) Subtitle Edit 5 is a major new release and a big step for the project. For the first time, Subtitle Edit runs natively on Windows, macOS, and Linux from a single, modern, cross-platform codebase. The builds are self-contained, so no separate .NET installation is required, and on macOS and Linux the needed media components (mpv/ffmpeg) are bundled in. Please read before upgrading: Subtitle Edit 5 is a new application, not just an update of Subtitle Edit 4. It has been rebuilt from the ground up to be cross-platform, so: - It is not 100% the same app. The look, layout, and some workflows have changed. Some things are in different places, and a few behave differently than in SE4. - Not every SE4 feature exists in SE5 yet. SE5 covers all the core editing, conversion, sync, video playback, OCR, and online services, but some of the more specialized SE4 tools are not available yet. Features will continue to be added. If you rely on a specific SE4 feature that is missing, please keep SE4 installed alongside SE5. The easiest way to run both side by side is to use the Portable versions of SE4 and SE5, which keep their settings separate and do not interfere with each other. Which version should I use? - Subtitle Edit 5: recommended for most users on Windows 10 (22H2) or newer, macOS 12+, and Linux. - Subtitle Edit 4: please continue to use SE4 if you are on an older Windows version (Windows 7/8), or on older / slower computers where SE5 may not run well. SE4 remains available and is the right choice in those cases. To run SE4 and SE5 at the same time, use the Portable versions - you can try SE5 while keeping SE4 as a fallback. Feedback is welcome on the issue tracker - it directly helps prioritize what comes next. Thank you to everyone who has supported and contributed to Subtitle Edit over the years :) ----------------------------------------------------------------------------------------------------- v5.0.0-rc4 (11th of June 2026) * Add Zonos TTS text-to-speech engine (CrispASR) * Add VoxCPM2 text-to-speech engine (CrispASR) * Add model-download button with per-model download size to Text to speech * Add settings window for the Piper text-to-speech engine * Add Parakeet RNNT models and SenseVoice-Small speech recognition (CrispASR) * Add Music Symbol settings to Fix Common Errors * Add live video preview to the burn-in window - thx MbuguaDavid * Add "Beautify time codes..." to the subtitle grid context menu * Add CJK unbreak actions * Add French installer translation - thx Need74 * Add forced-flag support and dirty-state tracking to binary edit - thx mjuhasz * Show a progress overlay while adding files in Batch Convert * Use default ASSA storage style for new/converted subtitles * Make bookmark add/toggle shortcuts work from the text editor - thx mjuhasz * Use the system font (SF Pro) as the default UI font on macOS - thx mjuhasz * Send consent attestation when voice cloning (CrispASR/Chatterbox) * Persist burn-in encoding settings between sessions * Update CrispASR to v0.7.1 * Update Polish translation - thx potplayer-fanpack * Update Russian translation - thx jekovcar * Update Turkish translation - thx bilimiyorum * Fix #11357: also ignore Ctrl+Left/Right word-jump navigation while a text box is focused - thx GrampaWildWilly * Fix #11422: "Check for updates" download button opened two browser tabs - thx nms42 * Fix #11479: OCR hang/out-of-memory on a stray "<" - thx wulf357 * Fix #11505: macOS menu bar reverting to English after restart - thx Mimikaki * Fix #11515: crash when deleting; log unhandled UI-thread exceptions - thx OldMan-Jakob * Fix #11529: hang when Modify Selection returns a large number of lines - thx ChocOranger * Fix OCR Change All not replacing current item in BinaryImageCompare mode * Fix OCR unknown-word prompt flow (Skip All / Skip Once) - thx mjuhasz * Fix shift-click and Shift+arrow selection in subtitle/batch-convert/OCR grids - thx mjuhasz * Fix Shift+End scrolling to top after a large-range selection - thx mjuhasz * Fix toggle focus between subtitle grid and waveform - thx mjuhasz * Fix grid coloring to honor too-many-lines and add independent CPS/WPM toggles * Color Show/Hide cells red on overlap - thx mjuhasz * Color Duration red for high CPS only when the CPS column is hidden - thx mjuhasz * Improve Fix Common Errors inline editor and grid consistency - thx mjuhasz * Fix continuation action sorting in Fix Common Errors - thx mjuhasz * Fix wrong "After" preview (doubled text) in Fix Common Errors - thx mjuhasz * Fix custom continuation style not reaching the engine; store it per profile * Fix corrupted ellipsis character in the custom continuation style dropdown - thx mjuhasz * Fix several burn-in bugs (output folder, cut precision, logo label) * Fix transparent-video output-folder and subtitle-detect bugs * Fix min-duration for the last subtitle in Apply duration limits * Fix Option+Arrow word navigation in the edit box on macOS - thx mjuhasz * Fix bookmark icon misaligning subtitle grid row numbers - thx mjuhasz * Change default bookmark color to dark amber for light-theme readability - thx mjuhasz * Fix cleared shortcuts being re-added from defaults on next startup - thx mjuhasz * Fix incomplete UI refresh when the system theme switches dark to light - thx mjuhasz * Fix Modify Selection length rules to use the longest visible line - thx mjuhasz * Fix video state when opening image-based subtitle formats - thx mjuhasz * Align binary-edit video player behavior with the main window - thx mjuhasz * Fix vertical text alignment in diff grid cells - thx mjuhasz * Unify the error log filename to error-log.txt * Statistics: fix gap-average calculation - thx ivandrofly ----------------------------------------------------------------------------------------------------- v5.0.0-rc3 (5th June 2026) * Add multi-select to Batch Convert file grid (#11407) - thx sharazy * Add multi-select to Fix Common Errors fix list - thx mjuhasz * Add Unbreak, Auto-break, Split/rebalance and Evenly distribute to the Selected Lines menu * Add CosyVoice text-to-speech engines (CrispASR) * Add hovered-position tooltip to the video player slider * Add download-status dots to the Speech to text forced aligner list + auto-select after download * Add "Skip step 1" setting to Fix Common Errors - thx mjuhasz * Restore /video: command-line argument for opening a video with the subtitle * Update WhisperCpp to v1.8.6 * Update CrispASR to v0.6.12 * Add Hungarian translation - thx Zityi * Update Turkish translation - thx bilimiyorum * Update Russian translation - thx jekovcar * Update Bulgarian translation - thx jekovcar * Update Czech translation - thx Matěj * Update Korean translation - thx 12si27 * Fix #10940: sync Fix Common Errors language dropdown to the Language field * Fix #11355: startup crash on Debian 13 (no system fonts) * Fix #11357: bare Left/Right arrow keys hijacked while a text box is focused * Fix #11379: CPS line-length strategy not updating the grid until restart * Fix #11388: Replace & find next silently skipping the current match * Fix #11392: honor user video shortcuts in the fullscreen video window * Fix #11393: restore focus to the subtitle grid after exiting fullscreen video * Fix waveform Alt+drag producing overlapping subtitles - thx mjuhasz * Fix missing renumber after inserting subtitles from the waveform * Fix spell check Change / Change all enabled before the word is edited * Fix OCR / spell check window font handling - thx mjuhasz * Fix macOS native menu toggle item labels - thx mjuhasz * Fix regex Replace & Find Next newline handling - thx mjuhasz * Reduce redundant Matroska cluster scanning (#11374, performance) ----------------------------------------------------------------------------------------------------- v5.0.0-rc2 (3rd June 2026) * Update WhisperCpp to v1.8.5 * Add engine update button to Auto-translate window * Add "Text to speech..." to subtitle grid context menu (#10941) * Add SSA Properties and Attachments windows + toolbar/menu items * Add Space/Home/End keyboard navigation in tool-window data grids - thx mjuhasz * Redesign Fix Common Errors Step 2 button flow and add fixes-applied counter - thx mjuhasz * Auto-translate: remember line-merge strategy per engine * Cache Regex and HTML stripping in three hot paths (performance) * Update Inno Setup installer Italian translation - thx bovirus * Update Italian translation - thx bovirus * Update Polish translation - thx potplayer-fanpack * Update Japanese translation - thx hisui3393 * Update Turkish translation - thx bilimiyorum * Update Portuguese translation - thx Blackspirits * Update Korean translation - thx 12si27 * Include LICENSE in release packages (#11278) * Fix #11275: "Adjust timings" always re-checked in Speech-to-text post-processing * Fix OpenAI-compatible STT for "Speech to text selected lines" - thx dkakaie * Fix #11307: highlight Duration column on CPS-too-high * Fix #11305: align waveform window with restored video position * Fix #11314: TextSplitResult crash on headless Linux * Fix operator-precedence and locale-code bugs * Fix image-export alignment and override-position bugs * Fix #11280: tighter undo capture + restore redo for unrecorded changes * Fix UndoRedoManager deadlock vector, Dispose race, and clone-timestamp bug * Fix ColorTextTooLong incorrectly triggering on CPS violations - thx mjuhasz * Fix #11327: window overflow with long video file paths * Fix UI scale not applied at startup and freeze on scale change * Fix OCR window progress bar / status text overlap * Fix progress bar / status text overlap in video generation windows * Fix SCC import: reject raw CEA-608 data rows and decode PAC line breaks * Fix #9803: SCC import italics and byte-misaligned musical notes * Fix #11342: SSA style editor *Default style recognition and style preview * Fix color picker: hex is 6 chars (RRGGBB) when there is no alpha channel * Fix SSA / ASSA style window preview box height and font scaling * Fix #11308: move last/first word up/down no longer leaves subtitles stuck at 3 lines * Fix #11308: merge selected lines auto-breaks the result to respect the 2-line limit * Fix #11308: undo/redo keeps the grid at the cursor row instead of jumping to the snapshot's saved row * Fix Google TTS voice listing crash when a parsed voice name is shorter than 5 characters ----------------------------------------------------------------------------------------------------- v5.0.0-rc1 (30th May 2026) * Update docs for RC1 * Add native macOS menu bar (NativeMenu) - thx mjuhasz * Add wrap-around to Find and Replace (Find Next/Previous + Replace) - thx mjuhasz * Add setting to hide the Plugins menu (default off) * Add SE 4 "Move start/end one frame (keep gap if close)" shortcuts * Add waveform "Set video position when moving start/end" (SE 4 parity) * Add SE 4 per-paragraph waveform footer + zero-padded ruler labels * Add waveform "Extract audio" format and sample rate settings (no more forced 16 kHz) * Read word-level timestamps from OpenAI-compatible speech to text (e.g. Grok) * Extend Shift+selection to Page Up/Down and Home/End in subtitle grid - thx mjuhasz * Float undocked video/waveform windows while SE main window is active * Show undocked video/waveform as independent top-level windows (own Alt+Tab entry) * Split-line: use full MinGap, proportional duration, and auto-break halves over max line length * Merge two subtitles: keep ASSA styles and split tracks by layer * Remove merge-continuation-lines prompt after speech to text * Route Compare time display through the full time-code converter - thx mjuhasz * Change a few default settings (video player on Windows + write tools log) * Update Italian translation - thx bovirus * Update Korean translation - thx 12si27 * Update Polish translation * Fix Compare dialog previous/next difference buttons and selection sync - thx mjuhasz * Fix Compare dialog showing 0 and 00:00:00.000 for placeholder rows - thx mjuhasz * Fix selected DataGridRow hover darkening / hover highlight issues - thx mjuhasz * Fix Shift+arrow selection in subtitle grid - thx mjuhasz * Fix Shift+PageDown skipping the last partially visible row - thx mjuhasz * Fix subtitle grid page size returning fewer rows than visible - thx mjuhasz * Make Find/Replace topmost only while the main window has focus * Fix Cmd+A selecting all rows instead of text when the edit box is focused (macOS) - thx mjuhasz * Fix "Merge continuation lines" finding nothing on mixed-language (CJK + Latin) files ----------------------------------------------------------------------------------------------------- v5.0.0-beta33 (28th May 2026) * Add IndexTTS TTS engine (Bilibili / IndexTeam, voice cloning ~870 MB) via CrispASR * Add CEA-708 (DTVCC) caption decoding from H.264 SEI cc_data in MP4 * Add Beautify time codes profile editor + integrated beautify tool * Add Merge continuation lines tool + Speech-to-text prompt * Add Snap-all-times-to-frames tool * Add Toggle dialog dashes / Merge as dialog / Frame-move commands * Add ShotChanges set-cue-to-shot-change green-zone commands (SE 4 parity) - thx JDTR75 * Add ShotChanges extend/snap shortcuts (SE 4 parity) - thx JDTR75 * Add "Set duration to max CPS" shortcut for selected lines * Add waveform of original video audio in TTS review window * Add burn-in / RemoveTextForHi / VisualSync / BeautifyTimeCodes toolbar icons * Expand SE 4 shortcut importer (15 more legacy mappings) + Sort by number / Video toggle brightness (mpv) commands * TTS: unbreak lines centrally before handing text to engines * Update CrispASR to v0.6.11 * Frame-align waveform gridlines and timeline labels in frame mode * Keep AdjustAllTimes window open across New/Open * Show "Allow overlap" checkbox on Waveform settings page * Show min/max warning background in SecondsUpDown duration field * Strip dialog dashes when splitting at text-box position * Update Italian translation - thx bovirus * Update Korean translation - thx 12si27 * Update Russian translation - thx jekovcar * Update Bulgarian translation - thx jekovcar * Update Portuguese (Brazil) translation - thx igorruckert * Update Finnish translation - thx Encrust8368 * Update Polish translation - thx potplayer-fanpack * Fix Find Previous shortcut not working after first use - thx mjuhasz * Fix CJK IME input issue on Find/Replace auto-focus - thx pdjdev * Fix bookmark persistence after line mutations and undo/redo - thx mjuhasz * Detect bookmark changes in undo/redo change detection - thx mjuhasz * Restore subtitle grid focus after toolbar/panel layout change - thx mjuhasz * Fix dialogs reopening due to stale shortcut key state after close - thx mjuhasz * Restore focus to subtitle grid / text box when Find/Replace closes; focus edit text box after Find Next/Previous - thx mjuhasz * Focus subtitle grid on startup so keyboard shortcuts work immediately - thx mjuhasz * Fix FindNext crash from null ScrollIntoView and stale find state - thx mjuhasz * Fix Replace dialog crashes (shorter/empty replacement, empty text editor) and stale Count after Replace All - thx mjuhasz * Fix Alt+Tab not returning to main window when video/waveform are undocked * Beautify window: move profile button to bottom + per-side fallback in change notes * Fix #11181: Whisper parameters textbox scrollbar overlapping text - thx yawoo * Fix #11176: BurnIn dialog unable to close after adding a logo - thx bichitoxxx ----------------------------------------------------------------------------------------------------- v5.0.0-beta32 (25th May 2026) * Add TTS Cast: per-actor voice / model / instruction mapping in Text-to-speech (for ASSA) * Add "Import from SE 4..." to the Shortcuts editor (Settings.xml / SE_Shortcuts.xml) * Add slot-based "Set actor 1..10" + "Set new actor..." commands for ASS/SSA workflows * Add "Go to previous bookmark", "Selection to lowercase/uppercase", and "Google it" (selected text) commands * Add CrispASR Granite "plus" model (granite-speech-4.1-2b-plus, 4 quants) * Chunk OpenAI-compatible STT uploads > 25 MB with silence-aware splits * Add audio upload format choice (MP3/M4A/WebM/WAV) for OpenAI-compatible STT - thx harjoc * Update Czech translation - thx Matěj * Update Italian translation - thx bovirus * Update Polish translation - thx potplayer-fanpack * Fix CrispASR Mega advanced-settings crash on Speech-to-text - thx rikimtasu * Fix CEA-608 ordering when reading H.264 captions from MP4 with B-frames * Fix STT response_format mismatch when sending timestamp_granularities - thx harjoc * Fix Mac → Windows shortcut import mapping Cmd to WIN instead of Ctrl - thx Larxen * Fix DeepSeek auto-translate prompt JSON-escaping - thx halaxy * Fix h264_qsv / hevc_qsv preset list (drop unsupported ultrafast / superfast) - thx bichitoxxx ----------------------------------------------------------------------------------------------------- v5.0.0-beta31 (24th May 2026) * Add scrub video position by mouse wheel over the video surface * Add OCR fallback database picker to the OCR window * Add waveform Seek Silence back/forward shortcuts * Update CrispASR to v0.6.10 - required for the new Mega backend * Manage Qwen3 TTS (CrispASR) model downloads in SE * Reload settings-page action labels and shortcut display names after a UI language change * Make OpenAI-compatible STT 'stream' opt-in (default off) * Update Korean translation - thx 12si27 * Update Russian translation - thx jekovcar * Update Bulgarian translation - thx jekovcar * Update Polish translation - thx potplayer-fanpack * Race-proof the auto-translate cancel guard and suppress the cancel-error dialog * Tighten Re-encode video window progress layout * Fix freeze on Settings OK after toggling "Show stop/fullscreen button" - thx bichitoxxx * Fix invisible status dots in Qwen3 TTS (CrispASR) settings * Fix Save-As dropping extension after a language suffix * Add Crisp ASR "Mega" backend (Qwen3-ASR-1.7B + robustness LoRA) for noisy / degraded speech * Add WhisperX-style wav2vec2 forced-aligner zoo to the Speech-to-text dialog * Update llama.cpp * Work on plugins ----------------------------------------------------------------------------------------------------- v5.0.0-beta30 (22th May 2026) * Add Qwen3 TTS 1.7B VoiceDesign model with natural-language voice instructions - thx subof * Add voice-design keyword picker to Text-to-speech for OmniVoice TTS - thx subof * Update CrispASR to v0.6.9 - restores Windows CPU and CPU-Legacy builds * Update OmniVoice TTS to OmniVoice-2026-05-22 - fixes the macOS/Linux dyld "Library not loaded" error * Add "green/orange/gray" dot for engines/models for STT/TTS/Auto-translate * Add snap-to-frames option for waveform drag * Update Japanese translation - thx hisui3393 * Add frame-based min-gap setting - thx p33t3r * Fix MP4 tx3g import returning only one subtitle when samples share a chunk * Update Italian translation - thx bovirus ----------------------------------------------------------------------------------------------------- v5.0.0-beta29 (20th May 2026) * Add Engine settings dialog for Qwen3 TTS, Kokoro TTS, and Chatterbox TTS * Bundle macOS DMG with mpv and ffmpeg * Add "Fill selected lines with clipboard text" to the main and OCR grid (Ctrl+Shift+V) * Add "Three letter (ISO 639-2/B)" language code option to save-as - thx LurkingNinja * Add "Try to use source encoding" option to Batch Convert - thx Hlsgs * Add Windows CUDA variant for Qwen3 TTS * Add Macedonian translation - thx kirepp-cmd * Find and offer a matching subtitle when a video file is dropped - thx GrampaWildWilly * Redesign llama.cpp OCR settings dialog and localize engine status labels * Get spell check dictionaries from the maintained LibreOffice repository and redesign the dialog * Select first row after OCR completes - thx LurkingNinja * Trim long paths in the Reopen menu and tighten menu spacing * Use a consistent, thicker progress bar across download and progress windows * seconv: reject unknown --encoding values and add --encoding:source * Update CrispASR to v0.6.8 - adds funasr, fun-asr-mlt-nano, and sensevoice (FunAudioLLM) backends; voxcpm2 TTS ~5-10x speedup * Update Bulgarian translation - thx jekovcar * Update Russian translation - thx jekovcar * Update Portuguese (Brazil) translation - thx igorruckert * Fix Text-to-speech async and Piper engine bugs * Fix Edge TTS crash on bad rate format and retry transient failures - thx KevinHa59 * Fix shortcut window not receiving focus * Fix Cancel button being clipped in the llama.cpp and speech-to-text engine download windows * Fix Alt-key access shortcuts not firing on buttons - underscore in label now drives the access key - thx LurkingNinja ----------------------------------------------------------------------------------------------------- v5.0.0-beta28 (17th May 2026) * Add drag-and-drop area to TTS voice settings for importing a voice * Add DeepSeek auto-translate engine * Add llama.cpp model picker, download, and Start/Stop server to OCR window * Add Ollama model picker for OCR window * Add nOCR database picker for OCR window * Add Ollama OCR model picker for Batch Convert * Add opt-in last-ditch nOCR pass for Batch Convert * Add nOCR fallback for BinaryOCR in Batch Convert * Add BinaryOCR fallback for nOCR in Batch Convert * Add copy-to-clipboard icon button for the speech-to-text console log * Add Linux ARM64 (aarch64) support to release builds and downloads * Add Linux x86_64 CUDA option to CrispASR * Redesign auto-translate window for better usability * Show llama.cpp model size for not-yet-installed entries * Localize llama.cpp OCR settings dialog and validation strings * Improve OCR cancellation handling and add llama.cpp request timeout * Register Alt+Space Windows system menu via Window class handler * Keep speech-to-text window open after Cancel * Clean up Import Plain Text and harden script alignment for long videos * Guard Ollama OCR against wrong (non-vision) model choices * Flash a check icon after copying ASSA draw code to the clipboard * Update llama.cpp to b9174 * Update CrispASR to v0.6.7 * Update Bulgarian translation - thx jekovcar * Update Russian translation - thx jekovcar * Fix Chatterbox TTS Turbo crash and silent default-voice fallback * Fix MP3 playback stopping before the end of the file * Fix batch Split/Break long lines adding stray line breaks * Fix auto-translate OK button and prompt before re-downloading llama.cpp model * Fix auto-translate prompt persistence and translated filename * Fix Netflix Quality Check ignoring per-check Apply checkboxes * Fix subtitle format changing to source format on save after auto-translate * Update Polish translation - thx potplayer-fanpack * Fix "Import plain text" align-button label to "Speech to text" - thx David ----------------------------------------------------------------------------------------------------- v5.0.0-beta27 (15th May 2026) * Save Settings.json in human-readable indented format - thx schabau * Auto-scroll speech-to-text console log to bottom - thx GrampaWildWilly * Update Bulgarian translation - thx jekovcar * Update Russian translation - thx jekovcar * Update Polish translation - thx potplayer-fanpack * Fix speech-to-text engine combo not reflecting selection - thx humble-b * Fix numeric keypad keys recorded as wrong shortcut - thx GrampaWildWilly ----------------------------------------------------------------------------------------------------- v5.0.0-beta26 (15th May 2026) * Add server-managed llama.cpp auto-translate engine * Add CrispASR MADLAD auto-translate engine - also available in Batch Convert * Add support for OpenAI-compatible speech-to-text servers * Add update detection for the llama.cpp engine * Add saving/restoring of Fix common errors window size - thx ironhussar * Add Russian translation - thx jekovcar * Move CrispASR to a shared data folder * Show install status and size in speech-to-text engine picker - thx David * Show media info window faster + make shortcut work in fullscreen - thx GrampaWildWilly * Update CrispASR to v0.6.6 - Granite engine now uses granite-4.1 * Update Bulgarian translation - thx jekovcar * Update Japanese translation - thx hisui3393 * Update Polish translation - thx potplayer-fanpack * Update Portuguese (Brazil) translation - thx igorruckert * Fix main window freeze after closing Settings with OK - thx David/bereld * Fix llama.cpp engine download on macOS/Linux - missing symlinked libraries * Fix llama-server startup with TranslateGemma models * Fix "Clear original" and original-changed flag after translation - thx wjcarpenter * Fix Mistral prompt text box missing in translate settings * Fix mp3 position jumping back to last seek at end of file * Fix removing color from non-WebVTT formats leaving tags * Hide extend-to-line items in subtitle grid header context menu ----------------------------------------------------------------------------------------------------- v5.0.0-beta25 (13th May 2026) * Add Finnish translation - thx userx77 * Add Mistral AI to AutoTranslate engine list * Add option for "Center-in-subtitle-grid" - thx GrampaWildWilly * Add green to WebVTT default color classes - thx wjcarpenter * Consolidate Whisper/speech-to-text logs into tools-log (incl. active-setting) * Allow editing subtitle text in Fix common errors step 2 * Honor selected audio track for waveform "Extract audio" and "Speech-to-text" * Improve "Speech to text for new selection" language selection - thx routineCode * Update CrispASR to v0.6.2 * Update Bulgarian translations - thx jekovcar * Update Czech translation - thx Matěj * Update Polish translation - thx potplayer-fanpack * Update Simplified Chinese translation - thx wuwufei * Fix duplicate waveform generation when opening a recent video * Fix minor issue with the right-click menu on mac - thx surfincanoy * Fix mp3 playback position jumping back at EOF * Fix PaddleOCR CUDA extraction on Linux - thx po5 * Fix Toggle Full Screen shortcut - thx TospeedArts * Fix UI freeze on Settings Apply/OK from file-type association save - thx David/bereld ----------------------------------------------------------------------------------------------------- v5.0.0-beta24 (10th May 2026) * Fix log text box height in speech-to-text - thx rikimtasu * Add "Merge two subtitles" tool * Add "Import CSV/XLSX with custom columns" window - thx Dubbing-Nerd * Add multi-mode to "Add to names list" window + report imported count - thx kt * Add OmniVoice TTS CUDA support * Add "Extract audio..." to waveform context menu * Improve video position sliders and waveform * Improvements for seconv (command line conversion tool) * Allow importing styles from Aegisub .sty files - thx maxz1717 * Add a few names to English name list * Update Bulgarian translations - thx jekovcar * Update Portuguese (Brazil) translation - thx igorruckert * Improve CSV import for unknown format detection - thx Dubbing-Nerd * Fix waveform issue - thx benshgit * Fix for additional video file extensions - thx nms42 * Fix OmniVoice for mac ----------------------------------------------------------------------------------------------------- v5.0.0-beta23 (9th May 2026) * Update Italian translation - thx bovirus * Add Czech translation - thx Matěj * Add forced aligner picker for CrispASR * Show install status and size in forced aligner pickers * Prompt for variant when re-downloading CrispASR * Update CrispASR to v0.6.0 * Add change-style action to Batch-convert * Remember active functions in Batch-convert * Remember selected rules in Fix Netflix errors * Add tooltips for video player - thx Monstercate * Show friendly voice names in Kokoro TTS - thx David * Set max waveform zoom to 500 * Update libmpv to 2026-04-21 * Reset fullscreen auto-hide timer on each user activity * Drop right padding on undocked video player * Preserve PAC justification on italic lines * Fix merge-lines in export-plain-text to merge all text * Fix mp3 file not stopping at end (regression) * Fix crash when loading idx file - thx Codling * Fix saving interjections for non-English languages - thx Pemicope * Update Polish translation - thx potplayer-fanpack * Don't invert waveform zoom direction with "Invert mouse wheel" - thx bereld * Add OmniVoice TTS * Update Japanese translation - thx hisui3393 * Add Swedish translation * Improve csv/xlsx/ods import - thx Dubbing-Nerd * Improve save-as-file-name after translate - thx wjcarpenter ----------------------------------------------------------------------------------------------------- v5.0.0-beta22 (5th May 2026) * Update Italian translation - thx bovirus * Update CrispASR to v0.5.7 * Add "Close translantion" file menu item - thx emwgee * Add NVIDIA API for autotranslate - thx subof * Fix for Piper Linux espeak symlink - thx Ironship * Handle original subtitle save failures better - thx Ironship * Fix crash when using "Generate video with subtitles" - thx AndroidTS007 * Fix remove color from WebVTT - thx wjcarpenter * Add more functions to batch-convert * Update Polish translation - thx potplayer-fanpack * Fix crash for binary ocr in batch-convert * Fix download/unpack title - thx sam12345 * Fix change-speed - thx JelmdtcSings ----------------------------------------------------------------------------------------------------- v5.0.0-beta21 (3rd May 2026) * Update Simplified Chinese localization - thx Monstercate * Try to improve vulkan loading for Qwen3TTS/Vulkan * Try to fix full screen video controls at startup - thx GrampaWildWilly * Fix for video full screen player right margin - thx GrampaWildWilly * Fix grid-auto-fix-columns for large font sizes - thx JDJD777 * Remove blur when zooming nocr images * Allow 0 lines for nOCR auto-draw - thx Zoltán * Testing different nOCR drawing algorithms * Minor improvements to speech-to-text with 1 file + improve video file check when opening subtitles * Minor UI fixes/improvements for OCR in batch convert * Fix for click on video player (regression) - thx JDJD777 * Fix ASSA toolbar not visible when default format is ASSA - thx farisruhi * Add Bulgarian translation - thx jekovcar * Improve alignment detection for OCR * Binary edit - allow import of image subs from mkv * Seveal fixes for nOCR + subtitle text import/export * Add "/batchconvertui" / "--batchconvertui" command line options to open batch convert UI only * Update CrispASR to v0.5.5 + add Granite 4.1 models (base/plus/nar) + add Kyutai STT backend (en/fr) * Minor fixes for undo/redo - thx rRobis * Fix for speech-to-text encoding UI log * Fix remove color from WebVTT - thx wjcarpenter ----------------------------------------------------------------------------------------------------- v5.0.0-beta20 (29th April 2026) * Add menu toggle for waveform toolbar - thx Ironship * Add auto-language for CrispASR glm/parakeet/Qwen3 * Append missing extension on save as - thx Ironship * Allow speech-to-text ETA to increase - thx Ironship * Preserve line breaks in compare HTML export - thx Ironship * Update Polish translation - thx potplayer-fanpack * Update Korean translation - thx Hackjjang * Update CrispASR download links * Update Kokoro TTS engine to v0.1.1 * Try to make toolbar more accessible - thx kylesskim-sys * Fix Flatpak release bundle publishing * Fix repeated video surface click toggles - thx Ironship * Fix regex multiline anchors with CRLF text - thx Ironship * Fix adjust duration percent calculation - thx Ironship * Fix TT family properties export - thx Ironship * Feature complete "seconv" (command line conversion tool) * Atomic save and shared lock for NOcrDb - thx Zoltán * Add "Merge selected lines bilingual" shortcut - thx jackqk * Stamp executable with version from Se.cs in publish steps - thx Zoltán ----------------------------------------------------------------------------------------------------- v5.0.0-beta19 (26th April 2026) * Rename "Whisper" folder to "SpeechToText" * Rename "TTS" folder to "TextToSpeech" * Improve nOcr quality in Batch Convert * Auto-detect pixels-is-space and language for BinaryOcr/nOcr in Batch Convert * Add BinaryOcr engine in Batch Convert * Delete temp files between batch speech-to-text items - thx xjlin0/cookesan * Add speech-to-text for new waveform selection - thx Ivo * Remember size/pos for translate window - thx Milo * More user friendly single-file-batch-speech-to-text - thx GrampaWildWilly * Check for large file in open to avoid crash - thx GrampaWildWilly * Update CrispASR urls * Update Italian translation - thx bovirus * Improve "Center text in subtitle text box" - thx enji1000 * Improve position detect for OCR * Improve Qwen3 TTS (Qwen3-TTS-12Hz-1.7B model + Windows-CPU-only-build) * Update Jananese translation - thx hisui3393 * Add Kokoro TTS ----------------------------------------------------------------------------------------------------- v5.0.0-beta18 (24th April 2026) * Improve auto-translate error messages - thx hisui3393 * Add a few ignore words for spell check * Fix renumber after splitting subtitle - thx rRobis * Improve Qwen3 TTS * Add Polish translation - thx potplayer-fanpack * Update Korean translation * Download silero vad for most speech-to-text engines * Check libmpv/libvlc at load - thx Aesthermortis * Fix spectrogram remains visible after disabling "Generate spectrogram" - thx Aesthermortis * Fix Crisp ASR Qwen3 + Granite * Add Crisp ASR Omni * More work on Crisp ASR * Pick Crisp ASR cpu/vulkan/cuda engine for Windows * Fix WebVTT STYLE italic/bold/color loss on conversion (Apple TV) - thx OtaStrom * Fix for context menu on mac in several windows ----------------------------------------------------------------------------------------------------- v5.0.0-beta17 (21st April 2026) * Add Korean translation - thx Hackjjang * Add Japanese translation - thx hisui3393 * Update Italian translation - thx bovirus * Add new language tags - thx potplayer-fanpack * Add setting for single-letter shortcuts in text box - thx rRobis/darnn * Fix subtitle file delimiter problem for seconv - thx rRobis * Update Chrome-lens-standalone to v3.4.0 - thx timminator * Add "alignment" as preview video sub setting - thx Mimikaki * Dialog fix for GoogleLensSharp - thx ArturAlekseev * Improve one-line detection for GoogleLens OCR - thx darnn * Fix possibly wrong data in time code control - thx JDJD777 * Add TT family properties menu - thx ryzen88 * Fix split line issue - thx benji1000 * Add waveform import/export/theme - thx benji1000 * Add customizable waveform shot change color - thx benji1000 * Black background for libmpv id (regression from earlier beta) - thx rRobis * Fix ASSA Style not changing in preview after switch style - thx rRobis * Fix text box info value reset after making new file - thx rRobis * Add some support for reading embedded lyrics from audio files - thx MG240 * Allow video - Cut, from/to mp3/wav * Fix video not playing when Loaded from URL (Mac) - thx Keysz * Update Crisp ASR to the latest version + add more models - thx Oplay66 * Improve live spell check - thx benji1000 * Try to Add custom model loading for Whisper - thx Adelio860 * Add separate parameter settings foreach speech-to-text engine ----------------------------------------------------------------------------------------------------- v5.0.0-beta16 (15th April 2026) * Update to Avalonia UI V12 * Update Italian translation - thx bovirus * Add "edit/export" from ocr window - thx donjondejota-art * Add new language tags - thx potplayer-fanpack * Minor improvements for main layout * Fix IsSpelledCorrect splitting hyphenated words by space instead of dash - thx ivandrofly * Fix Word spell check - thx moob158 * Fix grid sync when playing without waveform - thx vsemozhetbyt * Fix display preview on video panel for audio files - thx benshgit * Add Crisp ASR: Parakeet, Canary, Cohere * Fix bugs in NextWordInDoNotBreakList - thx ivandrofly * Add check box for custom start/end in remove text for HI - thx aaroncledge * Fix crash in OCR change allways - thx Ronaldvr * Some dictionary improvements for OCR * Add new setting for open-file-on-start - thx shotfirer * Allow more single-click-keys in text box - thx darnn * Fix Adjust-all-times in edit-image-based subs - thx AlexPaynes * Fix for EdgeTTS on mac ----------------------------------------------------------------------------------------------------- v5.0.0-beta15 (12th April 2026) * Update Italian translation - thx bovirus * Fix for import source-view text with BOM - thx MG240 * Fix OCR user dictionary add-actions - thx Hellbringer * Fix prompting for unknown words in Tesseract OCR - thx Ryu481 * Fix ASSA style issue - thx esprit-mi * Fix language tags + update Polish spell check directory - thx potplayer-fanpack * Add more Qwen3 ASR model options - thx Ironship * Fix issue with undo - thx Dravic * Add empty waveform when video has no audio * Add Windows-system-menu to many windows (alt+space) * Add total adjustment text in Sync - Adjust all times - thx kvalle22 * Try to add Word spell check - thx moob158 ----------------------------------------------------------------------------------------------------- v5.0.0-beta14 (8th April 2026) * Add Romanian translation - thx zildan * Update Italian translation - thx bovirus * Update Portuguese (Brazil) translation - thx igorruckert * Fix click on libmpv-wid to toggle play/pause - thx vsemozhetbyt * VLC now shows subtitle * Fix italic in text box for edit image ocr db * Fix for mpv refresh for full screen player * Fix: Allow large text files in batch-convert - thx phannhanhn201 * Add toggle select sub while playing to video menu (+shortcut) - thx vsemozhetbyt * Add "Extend only" to adjust durations - thx Dravic * Fix for nOCR load - thx HellbringerOnline * Fix for nOCR expanded view - thx HellbringerOnline * Fix toolbar separator visibility not updating when setting changes - thx ivandrofly ----------------------------------------------------------------------------------------------------- v5.0.0-beta13 (2nd April 2026) * Add Turkish translation - thx Hayri * Update Italian translation - thx bovirus * Add Inno Setup Portuguese.isl - thx Blackspirits * Update pt_PT_se.xml with new words - thx Blackspirits * Fix for binedit not setting correct position - thx KingGainer999 * Fix a few binedit issues * Fix extend to line after - thx laccka * Go to line: clamp out-of-range value to max subtitle count - thx ivandrofly * Fix a few batch convert issues * Improve multiple-replace context menu - thx ivandrofly * Add check-for-updates menu item - thx ivandrofly * Fix re-calc duration issue * Fix split issue * Fix issue with toggle tags in text box for ASSA * Add parameters for ASSA advanced karaoke effect - thx peruchali * Fix mpv not updating after change (regression from beta12) * Fix some missing entries in the language file - thx potplayer-fanpack * Add experimental Qwen3 TTS ----------------------------------------------------------------------------------------------------- v5.0.0-beta12 (30th March 2026) * Add Ukrainian translation - thx iohomenetyou * Add Portuguese (Brazil) translation - thx igorruckert * Update Italian translation - thx bovirus * Add Mistral TTS engine * Add Qwen3 ASR engine * Nice performance fixes - thx Ironship * Fix ASSA save storage style (in some cases) - thx writetome1/Ironship * Add setting for waveform-auto-focus - thx abc16361 * WebVTT now auto-merges lines with same time codes - thx xiaomotalktech/Ironship * Fixes for fix-names - thx ivandrofly * Fix duration cell not turning red on subtitle overlap - thx papioski1111/ivandrofly * Fix for "Convert frame" reverse conversion (and other improvements) - thx larsk2 * Disable space key for waveform buttons (use "Enter" if you want to use keyboard) - thx darnn * Check for subtitles in ".m4a + .m4b" files - thx Battletoadz * Fix for ErrorColor update after Options - Setting * Fix mpv native resize flicker on Windows - thx Ironship * Prompt for delete rules/groups in Multiple Replace - thx ivandrofly ----------------------------------------------------------------------------------------------------- v5.0.0-beta11 (26th March 2026) * Fix pixel-width error column background color - thx rRobis * Fix Remove Text for Hearing Impaired not working in batch convert - thx MKBB-85/ivandrofly * Mac: Option+Backspace deletes the previous word - thx derzz * Add Edge-TTS engine + TTS audio improvements - thx Ironship * Fix burn-in logo preview hidden under video on Windows with libmpv wid - thx Ironship * Improve mpv rendering on Linux - thx Ironship * Add Linux flatpak installer - thx Ironship * Add more ASSA advanced effect * Allow "manual sync" in visual sync - thx Bas * Fix for "count" in find-window - thx ivandrofly * Fix for missing space in italic text in IMSC 1.1 - thx Ironship * Update whisper.cpp to 1.8.4 * Add ASSA override tag history * Fix for Baidu Translate - thx Ironship * Fix: prevent portable from creating AppData\Subtitle Edit\Dictionaries - thx Ironship * Improve Find-window - thx ivandrofly * Update yt-dlp * Add more ASSA background boxes * Add keypad horizontal scrolling to the audio visualizer - thx derzz * Fix for merge lines selected item/original - thx derzz * Fix gap is shown in ms, not in frames - thx accessallareas * Fix build for mac with SIP disabled - thx Ironship/ilikepeaches ----------------------------------------------------------------------------------------------------- v5.0.0-beta10 (22th March 2026) * Multiple replace: Add find rule window (Ctrl+F) * Multiple replace: Add rule context menu + make "..." buttons optional * Multiple replace: Click on applied rules, now selects the rule * OCR window: Ctrl+/Ctrl- now zoom images in/out * Fix PaddleOCR crash - thx max2000777 * Add ASSA advanced effect: word spacing - thx rRobis * Add ASSA advanced effect: TV close (transition) * Add pixel-width column - thx rRobis * Fix numeric up/down null safety - thx rRobis * Fix for Grid-lines-all option * Add export to Cheetah (.cap) * Add tool Renumber - thx ivandrofly * Fix WPM info update in grid * Fix translate-via-copy-paste - thx truubo * Fix content alignment for image export - thx oxie93 * Allow toggle on/off "Start time" column - thx abc16361 * Add color chang coloring for Fix-common-error-fixes ----------------------------------------------------------------------------------------------------- v5.0.0-beta9 (17th March 2026) * Work on installer - thx bovirus * More customization possible for waveform toolbar buttons * Reload toolbar after UI language change - thx bovirus * Fix: Burn-In dialog crash on Linux - thx Ironship * Build macOS: Finish implementing signing/notarization - thx jdpurcell * Build: Update to V5 github actions * Italian language update - thx bovirus * Implement syntax-color-if-pixels-is-too-wide - thx Louis * Fix for WebVTT in mp4 - thx xiaomotalktech * Expose start and end line waveform colors in options - thx rRobis * Add margin for subtitle view preview - thx darnn * Fixes for Speech to text selected lines incl. option for default prompt settings - thx rRobis * Multiple-replace: Coloring, expland/collapse, remember expanded state, show rule matches for fixes - thx rRobis/TR-9970X ----------------------------------------------------------------------------------------------------- v5.0.0-beta8 (14th March 2026) * Use language from installer at first use - thx bovirus * Option for showing change log in installer * Translate about box - thx bovirus * Minor spectrogram fixes (you can toggle style live via a shortcut) * Add classic/legacy icon theme - thx Cyberyoda1411 * Add NOcr CaseFixer warmup * Rename Purfview-Whisper-Faster to Purfview-Faster-Whisper-XXL * Fix TTML Rosetta alignment - thx james * Add new ASSA advanced effects * Add Video -> More -> Open seconds subtitle - thx matmaggi * Fixes for ASSA styles font handling * Add "fixed duration" to import plain text * Allow some sorting of waveform toolbar buttons * Double click in edit-binary-sub-grid goes to video position ----------------------------------------------------------------------------------------------------- v5.0.0-beta7 (13th March 2026) * Work on installer - thx bovirus * Fix inclusion of Italian and Portuguese translations - thx bovirus * Fix for interjections - thx Hunique * Fix for ASSA font names - thx Hunique * Work on Video -> More -> Open seconds subtitle (not complete) - thx matmaggi * Enable gen-transparent-video with no subtitles (start in batch mode) * Work on OCR fix/replace lists