{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Projective sampling integrators\n",
"==============================="
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Overview \n",
"\n",
"In this tutorial, we will optimize a curve such that its shadows match some target reference. We will do this by only observing the shadows and making use of the the projective sampling integrators.\n",
"\n",
"Gradients introduced by indirect visibilty effects, such as shadows or reflections are difficult to capture. In previous tutorials, we mentioned how discontinuities require special care and that Mitsuba 3 uses a \"projective sampling\"-based method to handle these. Here, we will dig deeper into the integrators that implement this method:\n",
"\n",
"- [direct_projective
][1]\n",
"- [prb_projective
][2]\n",
"\n",
"More details about this projective sampling method can be found in the paper \"[Projective Sampling for Differentiable Rendering of Geometry][3]\".\n",
"\n",
"
bsplinecurve
][1] plugin such that its two shadows \"draw\" a moon and a star.\n",
"\n",
"We provide you with a pre-built scene in a XML file, we just need to load it. Let's also render it.\n",
"\n",
"[1]: https://mitsuba.readthedocs.io/en/latest/src/generated/plugins_shapes.html#b-spline-curve-bsplinecurve"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
""
],
"text/plain": [
"Bitmap[\n",
" pixel_format = rgb,\n",
" component_format = float32,\n",
" size = [256, 128],\n",
" srgb_gamma = 0,\n",
" struct = Struct<12>[\n",
" float32 R; // @0, premultiplied alpha\n",
" float32 G; // @4, premultiplied alpha\n",
" float32 B; // @8, premultiplied alpha\n",
" ],\n",
" data = [ 384 KiB of image data ]\n",
"]"
]
},
"execution_count": 2,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"scene = mi.load_file('../scenes/shadow_art.xml')\n",
"image_scene = mi.render(scene, spp=4096)\n",
"mi.Bitmap(image_scene)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"At the bottom, you can see our curve shape that we've wrapped on itself to create a circle. In the top left and top right, there are the two shadows that were casted by the curve onto two separate walls. We can't see the emitters from this point of view, as they are both slightly behind the sensor's viewpoint."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"For the purposes of this tutorial, we will only be considering the shadows themselves during the optimization. We should therefore build two sensors that respectively look straight at the left and right wall. Neither sensor should see the curve directly. For convenience, Mitsuba ships with a [batch
][1] sensor, which renders multiple views at once. There is a performance benefit to this approach: the just-in-time compiler only needs to trace the rendering process once instead of multiple times (once for each viewpoint).\n",
"\n",
"With the [batch
][1] sensor, the rendered output is a single image of all viewpoints stitched together horizontally.\n",
"\n",
"[1]: https://mitsuba.readthedocs.io/en/latest/src/generated/plugins_sensors.html#batch-sensor-batch"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
""
],
"text/plain": [
"Bitmap[\n",
" pixel_format = rgb,\n",
" component_format = float32,\n",
" size = [256, 128],\n",
" srgb_gamma = 0,\n",
" struct = Struct<12>[\n",
" float32 R; // @0, premultiplied alpha\n",
" float32 G; // @4, premultiplied alpha\n",
" float32 B; // @8, premultiplied alpha\n",
" ],\n",
" data = [ 384 KiB of image data ]\n",
"]"
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"dist = 5\n",
"batch_sensor = mi.load_dict({\n",
" 'type': 'batch',\n",
" 'sensor1': {\n",
" 'type': 'perspective',\n",
" 'fov': 100,\n",
" 'to_world': mi.ScalarTransform4f.look_at(\n",
" origin=[0, 0, -dist + 1],\n",
" target=[0, 0, -dist],\n",
" up=[0, 1, 0]\n",
" ),\n",
" },\n",
" 'sensor2': {\n",
" 'type': 'perspective',\n",
" 'fov': 100,\n",
" 'to_world': mi.ScalarTransform4f.look_at(\n",
" origin=[-dist + 1, 0, 0],\n",
" target=[-dist, 0, 0],\n",
" up=[0, 1, 0]\n",
" ),\n",
" },\n",
" 'film': {\n",
" 'type': 'hdrfilm',\n",
" 'width': 128 * 2,\n",
" 'height': 128,\n",
" 'sample_border': True,\n",
" 'filter': { 'type': 'box' }\n",
" },\n",
" 'sampler': {\n",
" 'type': 'independent',\n",
" 'sample_count': 256\n",
" }\n",
"})\n",
"\n",
"image_primal = mi.render(scene, sensor=batch_sensor, spp=4096)\n",
"mi.Bitmap(image_primal)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Reference image"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We can now load the reference image (by [Ziyi Zhang][1]). As mentionned previously, our goal is to reconstruct a star and moon outline with the shadows. Keep in mind that although the image shows both outlines next to each other, they will be on separate walls in our scene.\n",
"\n",
"Note that if you want use your own reference image, you should carefully match the pixel values of the initial images (wall and shadow values).\n",
"\n",
"[1]: http://rgl.epfl.ch/people/ziyi"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
""
],
"text/plain": [
"Bitmap[\n",
" pixel_format = y,\n",
" component_format = float32,\n",
" size = [256, 128],\n",
" srgb_gamma = 0,\n",
" struct = Struct<4>[\n",
" float32 Y; // @0, premultiplied alpha\n",
" ],\n",
" metadata = {\n",
" generatedBy => \"Mitsuba version 3.4.0\",\n",
" pixelAspectRatio => 1,\n",
" screenWindowWidth => 1\n",
" },\n",
" data = [ 128 KiB of image data ]\n",
"]"
]
},
"execution_count": 4,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"bitmap_ref = mi.Bitmap('../scenes/references/starmoon.exr')\n",
"image_ref = mi.TensorXf(bitmap_ref)\n",
"\n",
"# Reshape into [height, widht, 1]\n",
"resolution = bitmap_ref.size()\n",
"image_ref = mi.TensorXf(image_ref.array, shape=[resolution[1], resolution[0], 1])\n",
"\n",
"bitmap_ref"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Projective sampling integrator\n",
"\n",
"In this section, we will go over some of the options/parameters that are availble in the projective sampling integrators. We will see how they impact the quality of the gradients."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Finite differences"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We assess the quality of our gradients by comparing them against a ground truth measurement using finite differences. As was mentioned in previous tutorials, visualizing forward gradients is a nice method to get a better understanding of how some change in a parameter will change the rendered output.\n",
"\n",
"To capture a forward-mode gradient image, we must apply some change to our scene. Here, we arbitraily chose to look at a rotation aound the Y axis (vertical axis) of the [bsplinecurve
][1] object\n",
"\n",
"[1]: https://mitsuba.readthedocs.io/en/latest/src/generated/plugins_shapes.html#b-spline-curve-bsplinecurve"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [],
"source": [
"key = 'curve.control_points'\n",
"params = mi.traverse(scene)\n",
"initial_control_points = mi.Float(params[key])\n",
"\n",
"def apply_y_rotation(params, value):\n",
" control_points = dr.unravel(mi.Point4f, params[key])\n",
" radii = control_points[3]\n",
" points = mi.Point3f(control_points.x, control_points.y, control_points.z)\n",
" \n",
" rotation = mi.Transform4f.rotate([0, 1, 0], value)\n",
" rotated_points = rotation @ points\n",
" new_control_points = mi.Point4f(rotated_points.x, rotated_points.y, rotated_points.z, radii)\n",
" \n",
" params[key] = dr.ravel(new_control_points)\n",
" params.update()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We render the finite differences gradient estimate in multiple rounds with different seeds to avoid exceeding the maximum number of samples in a single kernel. In total we are using approximately 4.3 billion samples."
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"32/32\r"
]
}
],
"source": [
"epsilon = 1e-3\n",
"fd_spp = 2048\n",
"fd_repeat = 32 if 'PYTEST_CURRENT_TEST' not in os.environ else 1\n",
"res = batch_sensor.film().size()\n",
"img1 = dr.zeros(mi.TensorXf, (res[1], res[0], 3))\n",
"img2 = dr.zeros(mi.TensorXf, (res[1], res[0], 3))\n",
"\n",
"for it in range(fd_repeat):\n",
" apply_y_rotation(params, -epsilon)\n",
" img1 += mi.render(scene, sensor=batch_sensor, spp=fd_spp, seed=it)\n",
" params[key] = initial_control_points # Undo translation\n",
" params.update()\n",
" \n",
" apply_y_rotation(params, +epsilon)\n",
" img2 += mi.render(scene, sensor=batch_sensor, spp=fd_spp, seed=it)\n",
" params[key] = initial_control_points # Undo translation\n",
" params.update()\n",
" \n",
" print(f\"{it+1}/{fd_repeat}\", end='\\r')\n",
" \n",
"img_fd = (img2 - img1) / (epsilon*2) / fd_repeat"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We can now visualize the sum of all 3 channels' gradients. In the plot below, warmer colors indicate an increase in pixel value and colder values a decrease.\n",
"\n",
"We can see how the shadow on the left is becoming thinner, whereas the one on the right is becoming larger. This is expected as the circle will become more perpendicular to the wall in one case and more parallel in the other."
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"