# Amberwave – ALL .env settings (reference). To get started, .env.example with a few lines is enough;
# copy what you need from here into your .env one line at a time. Almost all of it can also be set in the app.
# NEVER share the ".env" itself or put it into the update archive.
# Everything commented out uses the stated default value.

# ---- App ---------------------------------------------------------------------
# Prefix of the container names (<name>_app, <name>_db, <name>_tunnel) – change it to run a second stack next to the first
# STACK_NAME=amberwave
APP_NAME=Amberwave
# Public address (for links in emails), e.g. https://musik.example.de
APP_URL=
# Name under which the Windows client shows this server when searching the home network (empty = hostname of the container)
CLIENT_SERVERNAME=
# Key for encrypting stored credentials (SMTP password, API keys, Cloudflare). Normally leave empty:
# Amberwave generates it itself (PFAD_CACHE/.app_secret_key), shows it once at the end of setup and puts a copy
# next to the backups. Your own value, e.g. with: openssl rand -hex 32
# APP_SECRET_KEY=
# Time zone for date/time (emails, expiry dates, listening history)
# APP_ZEITZONE=Europe/Berlin
# “New version available” notice in Admin (queries the releases on GitHub, installs nothing); 0 = off
# UPDATE_HINWEIS=1
# Where the versions come from (GitHub releases API of the project); empty = no query
# AMBERWAVE_UPDATE_URL=https://api.github.com/repos/lfrank95/amberwave/releases

# ---- Host / ports / folders ------------------------------------------------
# Folder containing app/ and docker/ from the package. Required for the stack via the Unraid interface.
AMBERWAVE_ORDNER=/mnt/user/appdata/amberwave
APP_PORT=9950
# User the app reads and writes as – on Unraid like Lidarr: 99 (nobody) / 100 (users)
# Leave empty = Amberwave detects the user itself (owner of app data or music collection, otherwise 99:100) and remembers it.
# Amberwave never runs as root. Fixed values here always win; the tunnel container reads them as well.
PUID=99
PGID=100
# 0.0.0.0 = reachable across the whole LAN. If cloudflared runs directly on the host (not as a container): 127.0.0.1
# APP_BIND=0.0.0.0

# Music collection on the array (Unraid e.g. /mnt/user/musik)
MUSIK_PFAD=/mnt/user/musik
# rw (default) = the container CAN write in the music folder. Whether Amberwave MAY do so is a separate switch in the app
# (Settings → Library & automation → Music folder, newly set up: read only) or MUSIK_SCHREIBEN=1 here.
# ro = hard safety switch, music folder mounted read-only, regardless of what the app says.
# MUSIK_SCHREIBEN=0
# Restart the stack after changing this (Compose Down/Up). The Cleanup section shows whether writing works.
# MUSIK_MOUNT_MODUS=ro

# Folder for EVERYTHING Amberwave creates itself (Unraid e.g. /mnt/cache/appdata/amberwave): containing db/ (database),
# cache/ (covers, secrets) and backups/ (backups). The folders are created at startup. Not set = next to the docker-compose.yml.
APPDATA_PFAD=/mnt/cache/appdata/amberwave
# Backups to a DIFFERENT drive (recommended, e.g. the array); empty = <APPDATA_PFAD>/backups
# SICHERUNG_PFAD=/mnt/user/appdata/amberwave/backups
# Individual folders, in case it should be different – they take precedence over APPDATA_PFAD (older installations keep theirs):
# DB_DATA_PFAD=/mnt/cache/appdata/amberwave/db
# CACHE_PFAD=/mnt/cache/appdata/amberwave/cache
# BACKUP_PFAD=/mnt/user/appdata/amberwave/backups
# Move an existing installation to this scheme (on the server, as root): bash app/scripts/pfade-migrieren.sh  (dry run first)

# ---- Database + setup ----------------------------------------------------------
# Memory MariaDB keeps for data and indexes (passed to the db container). 512M is fine up to ~50,000 tracks; 1G or more above
# DB_PUFFER=512M
# Enter nothing: on first start, open http://<server>:9950/einrichtung.php (the link is in the log of amberwave_app).
# There you create the admin account; Amberwave creates the database, user and a random password itself
# (changeable under “Advanced”) and remembers them in <CACHE_PFAD>/geheimnisse/. Settings › Database shows them later.
# Your own values here take precedence (older installations keep theirs):
# DB_ROOT_PASSWORD=
# DB_NAME=amberwave
# DB_USER=amberwave
# DB_PASSWORD=
# How long after the container starts the setup page stays open (minutes, 0 = always). After that: docker restart amberwave_app
# EINRICHTUNG_FENSTER_MINUTEN=30
# Setup code: shown in the log (the link there already contains it). Default: always required – even on the home network, because a
# Docker port forward can make requests from the internet look like they come from the home network. Set your own:
# EINRICHTUNGSCODE=
# 1 = always ask for the code, even on the home network (default 0: no code from home-network addresses;
#     requests via the Docker bridge, a proxy or tunnel always need the code)
# EINRICHTUNG_CODE_IMMER=0

# ---- Paths INSIDE the container (only change if you change the mounts) --
# PFAD_MUSIK=/musik
# PFAD_CACHE=/cache
# PFAD_BACKUPS=/backups

# ---- Scanning the library ------------------------------------------------------
# File extensions that are scanned as music
# SCAN_DATEITYPEN=flac,mp3,m4a,m4b,mp4,aac,ogg,oga,opus,wav,aif,aiff,wv,ape,alac,dsf,dff,wma,mka,mpc
# Image files in the album folder that count as cover (without extension, order = priority)
# SCAN_COVER_NAMEN=cover,folder,front,album,albumart
# Folders skipped while scanning (comma-separated, part of the path)
# SCAN_IGNORIEREN=@eaDir,.Trash,#recycle,.AppleDouble
# Automatic scan for new/changed files every … hours (0 = manual only)
# SCAN_AUTO_STUNDEN=0
# Automation after scanning (if something was new/changed), in this order: identification by sound → complete metadata
# (albums → artists → artist images) → cleanup preview. Can also be toggled in the app (Settings → Library; what is saved there takes precedence).
# AUTO_METADATEN=1
# AUTO_AUFRAEUM_VORSCHAU=1
# Nightly run: the same chain once a day at this hour (0–23, −1 = off)
# NACHTLAUF_STUNDE=3
# Daily database backup to <BACKUP_PFAD>/datenbank: hour (−1 = off), how many daily ones to keep,
# for how many additional weeks to keep one weekly backup each
# SICHERUNG_STUNDE=2
# SICHERUNG_TAGE=7
# SICHERUNG_WOCHEN=4
# Edge lengths of the cover thumbnails in pixels
# COVER_GROESSEN=200,300,400,600,1200
# In the files (tags and folder names during cleanup) write standard names instead of stylized spellings
# (“JAŸ-Z” → “JAY-Z”). Within Amberwave itself the spelling stays as in MusicBrainz.
# Can also be set via the interface. Downside: intentional stylizations like “P!nk” are normalized in the files.
# KUENSTLER_STANDARDNAMEN=0

# ---- Playback / streaming ----------------------------------------------------
# Default quality for new devices: original, opus256, opus160, opus128, aac256, aac160
# STREAM_STANDARD=opus256
# “Save offline” (albums/playlists to the phone): highest quality devices may store – opus128, opus160,
# opus256 (default, about 2 MB per minute) or original (Android app only; browsers stay at opus256). 0 turns the
# feature off. Each device picks its own quality up to this limit. Also adjustable in Settings › Offline on devices.
# OFFLINE_QUALITAET=opus256
# Maximum number of simultaneous conversions (ffmpeg processes) across all users
# STREAM_MAX_UMWANDLUNGEN=8
# Path to ffmpeg in the container
# FFMPEG_PFAD=ffmpeg
# FFPROBE_PFAD=ffprobe
# After how many percent or seconds a track counts as “listened”
# GEHOERT_AB_PROZENT=50
# GEHOERT_AB_SEKUNDEN=240
# File downloads for family members (0 = off, admin may always). Also in Settings › Offline on devices.
# DOWNLOADS_ERLAUBT=0

# ---- Metadata sources -------------------------------------------------------
# (MusicBrainz and fanart.tv active; translation and Lidarr to follow)
# Identifier for all third-party services (MusicBrainz, AcoustID, Cover Art Archive, Deezer, fanart.tv, ListenBrainz, Cloudflare,
# Let's Encrypt …): “Amberwave/<version> ( <contact> )” – without installation ID. Requests are only counted locally per service
# (Settings › Services & keys).
# DIENSTE_KONTAKT=                  # Contact for the operators; empty = project address on GitHub
# Separate identifier only for MusicBrainz (replaces the shared one) – normally empty
# MUSICBRAINZ_USER_AGENT=
# MUSICBRAINZ_URL=https://musicbrainz.org/ws/2
# fanart.tv – free project key after registering on fanart.tv
# FANART_API_KEY=
# Artist images are fetched along with “Complete metadata” (after scanning, after each album match and at night);
# FANART_ALLE_MINUTEN is no longer needed.
# FANART_URL=https://webservice.fanart.tv/v3
# “Explicit” marking (E badge) per track from Deezer – public interface without a key, runs along with the
# MusicBrainz match and with “Match metadata” (via barcode, ISRC, otherwise search with length check).
# DEEZER_AKTIV=1
# DEEZER_URL=https://api.deezer.com
# Tracks without a match are only queried again after … days
# DEEZER_WIEDERHOLEN_TAGE=30
# Cover Art Archive (cover preview during the MusicBrainz match). Only change if you use your own mirror.
# COVERART_URL=https://coverartarchive.org
# Identify albums by sound (fingerprint → AcoustID → MusicBrainz). Free application key:
# sign up at acoustid.org → “Register an application”. Can also be entered in the app under Settings.
# ACOUSTID_API_KEY=
# ERKENNUNG_AKTIV=1
# Accept automatically: regeln = also with the same track list across several editions (CD/vinyl/countries) choose by rules
# and queue for review · eindeutig = only unambiguous cases · aus = only suggest everything
# ERKENNUNG_AUTO=regeln
# When AcoustID (by sound) queries automatically: anhalt = only for albums with no clues at all (no usable artist/title),
# everything else is matched via the tags at MusicBrainz · gruendlich = also when the tags are not enough · aus = never automatically
# (Uncertain cases are left for the “Identify by sound” button)
# ERKENNUNG_MODUS=anhalt
# From which match (%) an edition counts as fitting
# ERKENNUNG_SCHWELLE=90
# Fingerprints computed simultaneously
# ERKENNUNG_PARALLEL=2
# ACOUSTID_URL=https://api.acoustid.org/v2
# Most popular tracks per artist (ListenBrainz, open, no key); they stay valid for this many days, then are fetched again
# LISTENBRAINZ_URL=https://api.listenbrainz.org
# BELIEBT_TAGE=30
# Translation of English texts into German (Claude API), empty = no translation
# CLAUDE_API_KEY=
# Lidarr for requests
# ---- Downloads (Prowlarr/Newznab + SABnzbd) – off by default, enable in the app or with USENET_AKTIV=1 ----
# USENET_AKTIV=0
# ----------------------------------------------
# Folder on the host into which SABnzbd puts the finished downloads of the music category (Unraid e.g.
# /mnt/user/downloads/complete/music). Mounted in the container at PFAD_DOWNLOADS (default /downloads).
# Amberwave copies from there to <music folder>/<DOWNLOAD_ZIEL>/<name> and scans.
DOWNLOAD_PFAD=/mnt/user/downloads/complete/music
# rw (default) = Amberwave may delete the original after import; ro = source stays untouched
# DOWNLOAD_MOUNT_MODUS=rw
# The following values can also be set in the interface (Admin → Settings → Downloads)
# PROWLARR_URL=http://192.168.1.10:9696
# PROWLARR_API_KEY=
# Prowlarr categories (Newznab): 3000 = audio overall, 3010 = MP3, 3040 = lossless
# PROWLARR_KATEGORIEN=3000
# SABNZBD_URL=http://192.168.1.10:8080
# SABNZBD_API_KEY=
# SABNZBD_KATEGORIE=music
# API path of the standard SABnzbd: /api, behind a proxy e.g. /sabnzbd/api, compatible services e.g. /api/sabnzbd
# SABNZBD_API_PFAD=/api
# Subfolder of DOWNLOAD_PFAD in which this client puts its downloads (empty = DOWNLOAD_PFAD itself)
# SABNZBD_IMPORT_ORDNER=
# Additional Newznab indexers, Prowlarr instances and SABnzbd clients: only in the interface (Settings → Downloads),
# see docs/usenet-newznab.md
# Subfolder in the music folder for imported downloads (sort in later via “Clean up files”)
# DOWNLOAD_ZIEL=Downloads
# After import: loeschen (default – delete the original only after a verified copy, requires DOWNLOAD_MOUNT_MODUS=rw) or behalten
# DOWNLOAD_NACH_IMPORT=loeschen
# Automatically sort imported albums into the scheme (/Artist/Year - Album) after the MusicBrainz match, without approval.
# Only albums with a match; unrecognized ones stay under DOWNLOAD_ZIEL and wait for you. 0 = everything stays in DOWNLOAD_ZIEL until approved.
# DOWNLOAD_EINSORTIEREN=1
# Convert WAV and AIFF losslessly to FLAC on import (audio bit-exact, verified by checksum; tags are carried over).
# FLAC is smaller and can be fully tagged. 0 = copy files unchanged.
# DOWNLOAD_WAV_ZU_FLAC=1
# Watchdog: if a download in SABnzbd does not change for this many minutes (progress, status, post-processing; pauses
# do not count), it counts as failed and is removed from the queue – if it was at 100% or in
# post-processing and the music is in the download folder, it is taken over normally. Minutes, 5–1440.
# DOWNLOAD_HAENGT_MIN=30
# If a download no longer shows up in SABnzbd at all (neither queue nor history): failed after this many minutes.
# DOWNLOAD_WEG_MIN=15
# Downloads as a separate item in the main menu (admins only; on the phone under “More”). 1 = on
# DOWNLOAD_REITER=0

# LIDARR_URL=
# LIDARR_API_KEY=

# ---- Email (web.de) ----------------------------------------------------------
# Can also be maintained in the app under Settings → Email (which then takes precedence).
# SMTP_HOST=smtp.web.de
# SMTP_PORT=587
# SMTP_VERSCHLUESSELUNG=tls
# SMTP_BENUTZER=
# SMTP_PASSWORT=
# SMTP_ABSENDER=
# SMTP_ABSENDER_NAME=Amberwave
# SMTP_TLS_PRUEFEN=1
# Days before an account expires on which the reminder is sent
# MAIL_ABLAUF_TAGE=30

# ---- Accounts & security -----------------------------------------------------
# Registration: freigabe (open, admin approves) or aus
# REGISTRIERUNG_MODUS=freigabe
# REGISTRIERUNG_MAX_PRO_STUNDE=3
# LOGIN_MAX_FEHLVERSUCHE=5
# LOGIN_SPERRE_BASIS_SEKUNDEN=60
# LOGIN_MAX_FEHLVERSUCHE_PRO_IP_STUNDE=30
# LOGIN_MAX_FEHLVERSUCHE_PRO_NAME_STUNDE=20
# PASSWORT_RESET_MAX_PRO_IP_STUNDE=5
# PASSWORT_RESET_MAX_PRO_MAIL_STUNDE=3
# PASSWORD_RESET_GUELTIG_MINUTEN=60
# REMEMBER_ME_TAGE=30
# PASSWORT_MIN_LAENGE=8
# SITZUNG_STUNDEN=12
# Devices per user (0 = unlimited), admins unlimited by default
# MAX_GERAETE_PRO_NUTZER=3
# MAX_GERAETE_ADMIN=0
# Validity of device keys for apps (e.g. a future Windows player) in days, 0 = unlimited
# API_TOKEN_TAGE=365
# SICHERHEIT_CSP=1
# SICHERHEIT_HSTS=1
# IP address(es) of the tunnel/proxy whose headers (CF-Connecting-IP) are trusted, e.g. 172.17.0.11 – recommended
# PROXY_IPS=
# auto = trust proxy headers from private senders, but without PROXY_IPS only with a public address in them; ja/nein forces it
# PROXY_VERTRAUEN=auto

# ---- Sorting -------------------------------------------------------------------
# Artists are sorted by first name last name (“Michael Jackson” under M).
# 1 = a leading article does not count (“The Beatles” under B), 0 = the article counts too (under T)
# SORTIERUNG_OHNE_ARTIKEL=1

# Lock the interface against zooming (Ctrl+mouse wheel, Ctrl +/−, two-finger zoom, double tap) – it then always fits the window.
# 0 = zooming allowed (e.g. for people who need larger text).
# ZOOM_SPERREN=1

# ---- Remote control -----------------------------------------------------------
# Devices of the same account control each other (admins: all devices). For this, every open app keeps a back channel
# (Server-Sent Events) to the server, which occupies a PHP process – the limit protects Apache (MaxRequestWorkers).
# Beyond that, devices poll every 2 s (somewhat sluggish, but works).
# FERN_AN=1
# FERN_MAX_VERBINDUNGEN=40
# Seconds without a sign of life until a device counts as offline
# FERN_OFFLINE_SEKUNDEN=15
# Length of a back channel, after which the device rebuilds it (proxies/Cloudflare cut long connections)
# FERN_KANAL_SEKUNDEN=30
# How often the channel checks for new commands (milliseconds) – smaller = faster, more database queries
# FERN_TAKT_MS=250

# ---- Home network (automatic quality) ----------------------------------------
# At home, devices on “Automatic” get the original file, on the go a suitable Opus quality.
# Networks that count as home (CIDR, comma-separated)
# LOKALE_NETZE=10.0.0.0/8,172.16.0.0/12,192.168.0.0/16,127.0.0.0/8,fc00::/7,fe80::/10,::1/128
# Devices at home that come in through the Cloudflare Tunnel are recognized by the app by your public address.
# It is detected automatically (every 15 min.); with a fixed IP you can enter it here.
# HEIM_IP=
# HEIM_IP_ERKENNEN=1
# HEIM_IP_URLS=https://1.1.1.1/cdn-cgi/trace,https://[2606:4700:4700::1111]/cdn-cgi/trace

# Theme radios (80s, 90s …, genres) on the home page: this many songs are needed for a decade/genre to get its own button
# RADIO_THEMA_MIN=25
# Artist radio: share of the artist themselves in percent (favorites first). 100 = only the artist (default); smaller = the rest are matching tracks by others (30–100)
# RADIO_KUENSTLER_ANTEIL=100
# Genre radios, how strict: family radio (pop, rock …) takes songs whose strongest family it is (from RADIO_FAMILIE_MIN) or which
# reach the share RADIO_FAMILIE_SICHER; single-genre radio takes album genres among the two strongest from RADIO_GENRE_MIN.
# Pure representatives play first, borderline cases less often. Too few songs → automatically looser.
# RADIO_FAMILIE_MIN=0.40
# RADIO_FAMILIE_SICHER=0.55
# RADIO_GENRE_MIN=0.35

# Genre families (clicking “Pop”, “Electronic” … finds the whole family): from what percentage an album counts toward the family (5–90)
# GENRE_FAMILIE_SCHWELLE=25

# --- Sound analysis (acoustic similarity: sound radio, signal in radio) ---
# Runs in the background with the built-in method (Python 3 + numpy in the image). Can also be toggled in the app.
# KLANG_AKTIV=1
# KLANG_PYTHON=python3
# Optional extended model (Discogs-EffNet by MTG, CC BY-NC-SA 4.0 – non-commercial only). NOT included:
# the admin installs it under Admin → Library → Tools with a button, the server then downloads it directly from the vendor.
# KLANG_PIP_PAKET=essentia-tensorflow
# KLANG_MODELL_URL=https://essentia.upf.edu/models/feature-extractors/discogs-effnet/discogs-effnet-bs64-1.pb
# Style recognition (400 Discogs styles per track, refines genre families and radio) – head and labels for the model:
# KLANG_STIL_URL=https://essentia.upf.edu/models/classification-heads/genre_discogs400/genre_discogs400-discogs-effnet-1.pb
# KLANG_STIL_JSON_URL=https://essentia.upf.edu/models/classification-heads/genre_discogs400/genre_discogs400-discogs-effnet-1.json
# KLANG_LIZENZ_URL=https://essentia.upf.edu/models/LICENSE

# --- Spatial sound (upmix) ---
# Package with room impulse responses for the reverb; the server downloads it on button press (Admin → System)
RAUMKLANG_IR_URL=https://www.voxengo.com/files/impulses/IMreverbs.zip
# Spatial sound – center (LCR via M/S) and bass management (defaults for new devices, switchable per device in the spatial sound menu):
# Crossover frequency of the bass management in Hz (Linkwitz-Riley 24 dB/oct.): 60–250 in 10 Hz steps or 0 = bass stays in front, LFE empty
RAUMKLANG_LFE_TRENNUNG=80
# LFE level in dB: -10 (the receiver raises the LFE by 10 dB) or 0
RAUMKLANG_LFE_PEGEL=-10
# Center share 0–100: 0 = L/R only, 50 = classic matrix (center 0 dB, front −3 dB), 100 = the whole middle in the center
RAUMKLANG_CENTER_ANTEIL=50
# Rear delay for new devices in ms (0–20): Ls/Rs delayed so that nothing drifts to the back (precedence effect), adjustable per device in the effects rack
# RAUMKLANG_REAR_DELAY=0
# AmberSpace strength “Wide” (sliders 0–100; intensity and level: 50 = like the preset, ±12 dB over the whole range).
# Default: intensity 34 (−3.8 dB), surround level 75 (+6.0 dB), effects 26 %, width 22 %.
# “Subtle” and “Intense” = Wide scaled by ∓/± PERCENT (intensity and level as amplitude, effects and width directly)
# RAUMKLANG_MITTEL_INTENSITAET=34
# RAUMKLANG_MITTEL_PEGEL=75
# RAUMKLANG_MITTEL_EFFEKTE=26
# RAUMKLANG_MITTEL_BREITE=22
# RAUMKLANG_STAERKE_PROZENT=20
# Upload your own impulse responses: all users (1) or admins only (0); only an admin may ever delete
# RAUMKLANG_UPLOAD_ALLE=1
# Restart the spatial sound module fresh on every track selection, every skip and every play (1 = yes, 0 = let it run through)
RAUMKLANG_NEUSTART=1

# --- Direct access on the home network (like Plex) ---
# Set up under Admin → System. Everything here is optional – the interface can set it just as well.
# Port on which Amberwave is reachable on the home network via HTTPS (open it in the stack, NO router port forward needed)
LOKAL_HTTPS_PORT=8443
# Cloudflare API token with the permission “Zone → DNS → Edit” (template “Edit zone DNS”);
# for the Cloudflare Tunnel additionally “Account → Cloudflare Tunnel → Edit”
CLOUDFLARE_API_TOKEN=
# Cloudflare Tunnel – Amberwave sets it up under Settings › Access & network.
# TUNNEL_IM_STACK=1: separate cloudflared container “tunnel” in the stack (Amberwave creates the tunnel and gives it the token)
# TUNNEL_IM_STACK=0: you already have a cloudflared on the server or in the network – Amberwave only adds its rule to that
#                    tunnel (the tunnel must be configured in the Cloudflare dashboard, not via config.yml) and you
#                    specify in the interface how this cloudflared reaches the app (e.g. http://192.168.1.10:9950)
TUNNEL_IM_STACK=1
# Public name (otherwise enter in the interface), target of the forwarding (default: in the stack http://app:80,
# otherwise the home network address of the app), name of the cloudflared container in the stack, image tag
# TUNNEL_HOST=musik.example.de
# TUNNEL_ZIEL=
# TUNNEL_CONTAINER=amberwave_tunnel
# TUNNEL_IMAGE_TAG=latest
# Name for direct access (points to the LAN IP via DNS) and the LAN IP of the server
DIREKT_HOST=
DIREKT_IP=
# Optional: email for Let's Encrypt (notices before expiry in case automatic renewal fails)
ACME_EMAIL=
# Renew the certificate automatically (1 = yes) and how many days before expiry
DIREKT_AUTO_ERNEUERN=1
DIREKT_ERNEUERN_TAGE=30
# Instead of the automatic mode, specify your own local HTTPS address (e.g. via Nginx Proxy Manager)
DIREKT_URL=

# --- Activity (Admin → Activity) ---
# Hide paused playbacks after this many minutes
AKTIVITAET_PAUSE_MINUTEN=10

# --- Loudness (EBU R128) ---
# Background analysis (1 = on): measures each track and each album as a whole; the players use this to equalize volume
LAUTHEIT_AKTIV=1
# Simultaneous measurements (ffmpeg, low priority). Approx. 50–100× real time per measurement.
LAUTHEIT_PARALLEL=2
# Default target value in LUFS (each device can choose a different one in its profile)
LAUTHEIT_ZIEL=-16
# Highest true peak after boosting in dBTP (there is no limiter – quiet tracks are only raised as far as this allows)
LAUTHEIT_DECKE=-1

# --- Clean up files (Admin → Library; always preview first, then approval) ---
# Scheme for folders and file names. Placeholders: {albumkuenstler} {album} {originaljahr} {jahr} {edition} {land} {format}
# {katalognummer} {disc} {nr} {titel}. Text in the brackets only appears if the value is not empty:
# “{ [edition]}” → “ [Special Edition]”, “{disc-}” → “2-” (only with multiple CDs).
# {edition} = version name of the edition; without a name and when there are several editions: year country format.
AUFRAEUMEN_ORDNER={albumkuenstler}/{jahr} - {album}{ (edition)}
AUFRAEUMEN_DATEI={CD disc/}{nr} - {titel}
# Allowed characters in names: windows = everything that works on Windows and over SMB (/ \ : * ? " < > | are
# replaced, invisible characters removed, accents normalized, CON/AUX/NUL … defused; umlauts stay)
# ascii = strictly ASCII only (ä → ae, é → e, emojis are dropped) – e.g. for car radios and old players
AUFRAEUMEN_ZEICHEN=windows
# Permissions for everything Amberwave creates or rewrites in the music folder (octal). Default like Unraid's “New Permissions”:
# folders 777, files 666 – so that Windows (SMB) and other containers can open everything
AUFRAEUMEN_RECHTE_ORDNER=777
AUFRAEUMEN_RECHTE_DATEI=666
# “CD 1” folder also for albums with only one disc (1); 0 = CD folder only for multiple discs
AUFRAEUMEN_CD_IMMER=0
# sauber = old/foreign tags are removed (except AUFRAEUMEN_BEHALTEN), ergaenzen = foreign ones stay
AUFRAEUMEN_TAGS=sauber
AUFRAEUMEN_BEHALTEN=LYRICS,UNSYNCEDLYRICS,ISRC
# (ITUNESADVISORY – the “explicit” marking – is always kept)
# Write contributors (producer, composer, musicians, guests …) into the files? 0 = no (default; most players
# can't do anything with them, they are still kept in Amberwave), 1 = yes
AUFRAEUMEN_MITWIRKENDE=0
# What goes into the trash as junk (patterns, case-insensitive)
AUFRAEUMEN_MUELL=*.nfo,*.sfv,*.m3u,*.m3u8,*.log,*.cue,*.md5,*.ffp,*.url,*.accurip,Thumbs.db,desktop.ini,.DS_Store,._*
# Embed cover in the files (1 = yes) and at what size (pixels, longest side of the available step)
AUFRAEUMEN_COVER=1
AUFRAEUMEN_COVER_GROESSE=1200
# Trash (folder .amberwave-papierkorb in the music folder): keep for this many days
AUFRAEUMEN_PAPIERKORB_TAGE=30
# Albums from which a track was played in the last N minutes are only cleaned up by Amberwave on the next run (0 = off)
AUFRAEUMEN_SCHONZEIT_MIN=10
# How long an approval is valid (hours). Whatever could not be moved by then (album was playing, scan was running)
# goes back to “open” and must be approved again – nothing starts silently days later.
# AUFRAEUMEN_FREIGABE_STUNDEN=12
# Length limits for names: per file/folder name in bytes (encrypted folders e.g. on Synology: 140) and for the whole
# path below the music folder in characters (Windows programs manage about 260 in total). Anything too long is shortened sensibly.
AUFRAEUMEN_NAME_MAX=180
AUFRAEUMEN_PFAD_MAX=200

# --- WAV export (playlist as one long file, e.g. for the mixing console) ---
# Keep finished exports for this many hours, then delete automatically
EXPORT_AUFHEBEN_STUNDEN=48
# At most this many exports per user (older ones are deleted on the next export) and tracks per export
EXPORT_MAX_PRO_NUTZER=3
EXPORT_MAX_TITEL=500
# Split files after this many minutes (0 = automatically just under 4 GB, approx. 6 hours)
EXPORT_TEIL_MAX_MINUTEN=0

# ---- Warmth (tube saturation in the player) ---
# Default for new devices: aus | hauch | warm | glut
WAERME_VORGABE=aus
# Reference loudness of the tube: at this loudness (LUFS) Hauch/Warm/Glut sound as tuned; other tracks are driven
# correspondingly harder or softer (±12 dB, auto-gain keeps the level). Higher (e.g. -14) = more tame overall, lower = stronger
WAERME_BEZUG_LUFS=-18
# Tube type for new devices: ax7 (12AX7, hot) | at7 (12AT7, punch) | au7 (12AU7, smooth) | sn7 (6SN7, broad)
WAERME_ROEHRE=ax7
# Tube stage topology: eintakt (lots of H2) | gegentakt (even harmonics cancel, more H3; double processing load in HQ mode)
WAERME_SCHALTUNG=eintakt

# ---- Tape (tape machine in the player, after the tube) ---
# Default for new devices: aus | fresh | worn | old. Reference loudness as WAERME_BEZUG_LUFS (own value possible)
BAND_VORGABE=aus
# BAND_BEZUG_LUFS=-18
# Tape type for new devices: sm900 (modern high-output tape, lots of headroom) | 456 (the classic, saturates earlier) | sm468 (soft, saturates early at 7.5 ips)
# BAND_SORTE=sm900
# Equalization standard: nab (USA, 50 µs) | ccir (Europe/IEC, 35 µs at 15 ips, 70 µs at 7.5 ips) – determines how strongly the highs are boosted before the tape
# BAND_NORM=nab

# ---- Import unsorted music (sorting assistant, Admin → Library → Unsorted) ---
# Inbox folder relative to the music folder: put unsorted files here (nested arbitrarily). It is never scanned normally –
# only the assistant identifies each file by sound (AcoustID key required), suggests albums and sorts them in after your approval.
# SORTIEREN_ORDNER=_unsortiert
# Staging area for new albums until they are linked to MusicBrainz and sorted into AUFRAEUMEN_ORDNER
# SORTIEREN_ZWISCHEN=_sortiert
# From this AcoustID match score (%) a track counts as recognized (below that: “unrecognized”, manual search possible)
# SORTIEREN_SCHWELLE=80
# Preselection for duplicate files (same recording, worse quality): papierkorb | liegen
# SORTIEREN_DUBLETTEN=papierkorb

# ---- Effects rack: own presets (per account, on all devices) ---
# Maximum number of own presets per effects device (Glow, Iron, Silk, Space, EQ, whole rack) and account
# PRESETS_MAX=100

# ---- Transformer (AmberIron in the player, between tube and tape) ---
# Default for new devices: aus | hauch | satt | schwer, core: nickel | stahl | eisen. Reference loudness as WAERME_BEZUG_LUFS
TRAFO_VORGABE=aus
# TRAFO_KERN=stahl
# TRAFO_BEZUG_LUFS=-18

# ---------- Home network devices (DLNA: AV receivers, TVs, streamers) ----------
# On/off (1/0)
# HEIMNETZ_AN=1
# Address at which devices on the home network reach the server. Empty = detect itself (from how the app is called on the
# home network, otherwise direct access + APP_PORT). Only set if detection gets it wrong.
# HEIMNETZ_URL=http://192.168.1.10:9950
# Where to search (CIDR, comma-separated). Empty = the private networks of the server address (/24). In its own Docker network
# (bridge) multicast does not arrive – so Amberwave queries every address directly (unicast SSDP).
# HEIMNETZ_NETZE=192.168.1.0/24
# Effects renderer for “With effects” (default: bin/aw-render in the app folder)
# AW_RENDER_PFAD=
# Background service for the devices exits after this many minutes without playback
# HEIMNETZ_AGENT_RUHE=10
# How often the service polls the devices for status (ms)
# HEIMNETZ_ABFRAGE_MS=1000
# App ID of Amberwave's own screen for Google Cast (Custom Receiver in the Google Cast SDK Developer Console,
# address: <direct access or APP_URL>/cast/empfaenger.html). Empty = Google's default receiver.
# CAST_APP_ID=