Env map importance sampling by kevyuu · Pull Request #969 · Devsh-Graphics-Programming/Nabla

kevyuu · 2025-12-19T14:43:32Z

Description

Rework environment map importance sampling to vulkan and hlsl

Testing

Rework example 0 to use vulkan and hlsl
Unit Test Pull Request

TODO list:

Implement Warpmap Generation in hlsl
Implement Warpmap hierarchical map sampling in hlsl

include/nbl/builtin/hlsl/concepts/warp.hlsl

include/nbl/builtin/hlsl/warp/spherical.hlsl

include/nbl/builtin/hlsl/concepts/warp.hlsl

include/nbl/builtin/hlsl/warp/spherical.hlsl

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

…lImage class

devshgraphicsprogramming · 2026-03-25T17:03:25Z

src/nbl/video/sampling/EnvmapSampler.cpp

+					.baseArrayLayer = 0u,
+					.layerCount = 1u
+				},
+				.oldLayout = IImage::LAYOUT::TRANSFER_DST_OPTIMAL,


oldLayout be UNDEFINED, you'll be overwriting all data anyway on the destination mip

devshgraphicsprogramming · 2026-03-25T17:05:19Z

src/nbl/video/sampling/EnvmapSampler.cpp

+	core::smart_refctd_ptr<IGPUBuffer> lumaTexelBuffer;
+	const auto lumaMapLastMip = lumaMapMipLevels - 1;
+	const auto lumaMapLastMipExtent = lumaMapImage->getMipSize(lumaMapLastMip);
+	const auto lumaMapLastTexelCount = lumaMapLastMipExtent.x * lumaMapLastMipExtent.y * lumaMapLastMipExtent.z;
+	{
+		IGPUImage::SBufferCopy region = {};
+		region.imageSubresource.aspectMask = IImage::EAF_COLOR_BIT;
+		region.imageSubresource.mipLevel = lumaMapLastMip;
+		region.imageSubresource.baseArrayLayer = 0;
+		region.imageSubresource.layerCount = 1;
+		region.imageExtent = { lumaMapLastMipExtent.x, lumaMapLastMipExtent.y, lumaMapLastMipExtent.z };
+
+		IGPUBuffer::SCreationParams bufferCreationParams = {};
+		bufferCreationParams.size = lumaMapLastTexelCount * getTexelOrBlockBytesize(EF_R32_SFLOAT);
+		bufferCreationParams.usage = IGPUBuffer::EUF_TRANSFER_DST_BIT;
+		lumaTexelBuffer = logicalDevice->createBuffer(std::move(bufferCreationParams));
+		if (!lumaTexelBuffer)
+		{
+			if (auto* logger = logicalDevice->getLogger())
+				logger->log("ScreenShot: failed to create GPU texel buffer.", system::ILogger::ELL_ERROR);
+			return;
+		}
+		auto gpuTexelBufferMemReqs = lumaTexelBuffer->getMemoryReqs();
+		gpuTexelBufferMemReqs.memoryTypeBits &= logicalDevice->getPhysicalDevice()->getDownStreamingMemoryTypeBits();
+		if (!gpuTexelBufferMemReqs.memoryTypeBits)
+		{
+			if (auto* logger = logicalDevice->getLogger())
+				logger->log("ScreenShot: no down-streaming memory type for texel buffer.", system::ILogger::ELL_ERROR);
+			return;
+		}
+		auto gpuTexelBufferMem = logicalDevice->allocate(gpuTexelBufferMemReqs, lumaTexelBuffer.get());
+		if (!gpuTexelBufferMem.isValid())
+		{
+			if (auto* logger = logicalDevice->getLogger())
+				logger->log("ScreenShot: failed to allocate texel buffer memory.", system::ILogger::ELL_ERROR);
+			return;
+		}
+
+		IGPUCommandBuffer::SPipelineBarrierDependencyInfo info = {};
+		decltype(info)::image_barrier_t barrier = {};
+		info.imgBarriers = { &barrier, &barrier + 1 };
+
+		{
+			barrier.barrier.dep.srcStageMask = PIPELINE_STAGE_FLAGS::BLIT_BIT;
+			barrier.barrier.dep.srcAccessMask = ACCESS_FLAGS::TRANSFER_WRITE_BIT;
+			barrier.barrier.dep.dstStageMask = PIPELINE_STAGE_FLAGS::COPY_BIT;
+			barrier.barrier.dep.dstAccessMask = ACCESS_FLAGS::TRANSFER_READ_BIT;
+			barrier.oldLayout = IImage::LAYOUT::TRANSFER_DST_OPTIMAL;
+			barrier.newLayout = IImage::LAYOUT::TRANSFER_SRC_OPTIMAL;
+			barrier.image = lumaMapImage;
+			barrier.subresourceRange.aspectMask = IImage::EAF_COLOR_BIT;
+			barrier.subresourceRange.baseMipLevel = lumaMapMipLevels - 1;
+			barrier.subresourceRange.levelCount = 1u;
+			barrier.subresourceRange.baseArrayLayer = 0;
+			barrier.subresourceRange.layerCount = 1;
+			cmdBuf->pipelineBarrier(EDF_NONE,info);
+		}
+		cmdBuf->copyImageToBuffer(lumaMapImage,IImage::LAYOUT::TRANSFER_SRC_OPTIMAL,lumaTexelBuffer.get(),1,&region);
+	}


I don't think you need to download the average luma back to CPU, use an accessor for the average luma in your shader code

even if last mip level you compute is 2x1 or 2x2, the GPU can do a blilinearly filtered read right in the middle to get the average luma

devshgraphicsprogramming · 2026-03-25T17:08:56Z

src/nbl/video/sampling/EnvmapSampler.cpp

+			{
+				.barrier = {
+					.dep = {
+						.srcStageMask = PIPELINE_STAGE_FLAGS::BLIT_BIT,
+						.srcAccessMask = ACCESS_FLAGS::TRANSFER_READ_BIT,
+						.dstStageMask = PIPELINE_STAGE_FLAGS::ALL_COMMANDS_BITS,
+						.dstAccessMask = ACCESS_FLAGS::SHADER_READ_BITS
+					}
+				},
+				.image = lumaMapImage,
+				.subresourceRange = {
+					.aspectMask = IImage::EAF_COLOR_BIT,
+					.baseMipLevel = 0u,
+					.levelCount = lumaMapMipLevels - 1,
+					.baseArrayLayer = 0u,
+					.layerCount = 1u
+				},
+				.oldLayout = IImage::LAYOUT::TRANSFER_SRC_OPTIMAL,
+				.newLayout = IImage::LAYOUT::READ_ONLY_OPTIMAL,
+			},
+			{
+				.barrier = {
+					.dep = {
+						.srcStageMask = PIPELINE_STAGE_FLAGS::COPY_BIT,
+						.srcAccessMask = ACCESS_FLAGS::TRANSFER_READ_BIT,
+						.dstStageMask = PIPELINE_STAGE_FLAGS::ALL_COMMANDS_BITS,
+						.dstAccessMask = ACCESS_FLAGS::SHADER_READ_BITS
+					}
+				},
+				.image = lumaMapImage,
+				.subresourceRange = {
+					.aspectMask = IImage::EAF_COLOR_BIT,
+					.baseMipLevel = lumaMapMipLevels - 1,
+					.levelCount = 1,
+					.baseArrayLayer = 0u,
+					.layerCount = 1u
+				},
+				.oldLayout = IImage::LAYOUT::TRANSFER_SRC_OPTIMAL,
+				.newLayout = IImage::LAYOUT::READ_ONLY_OPTIMAL,
+			},


whats the point of the second barrier?
Barriering against a texture upload/update before this function is user's responsibility so I don't think the second barrier belongs here

Also you're supposed to join barriers so the bits would just OR together

Finally destStageMasks of ALL_COMMANDS_BITS is overkill, only barrier against your own work, the barrier the user performs after should take care of this

devshgraphicsprogramming · 2026-03-25T17:09:59Z

src/nbl/video/sampling/EnvmapSampler.cpp

+				.barrier = {
+					.dep = {
+						.srcStageMask = PIPELINE_STAGE_FLAGS::NONE,
+						.srcAccessMask = ACCESS_FLAGS::NONE,
+						.dstStageMask = PIPELINE_STAGE_FLAGS::COMPUTE_SHADER_BIT,
+						.dstAccessMask = ACCESS_FLAGS::SHADER_WRITE_BITS
+					}
+				},
+				.image = warpMapImage,
+				.subresourceRange = {
+					.aspectMask = IImage::EAF_COLOR_BIT,
+					.baseMipLevel = 0,
+					.levelCount = 1,
+					.baseArrayLayer = 0u,
+					.layerCount = 1u
+				},
+				.oldLayout = IImage::LAYOUT::UNDEFINED,
+				.newLayout = IImage::LAYOUT::GENERAL,
+			}


again this barrier should be user's resposibility

your Session::getPrevBarrier(..) would basically add bits for computeWarpMap not to step on its own toes

devshgraphicsprogramming · 2026-03-25T17:16:48Z

src/nbl/video/sampling/EnvmapSampler.cpp

+	{
+		IGPUCommandBuffer::SPipelineBarrierDependencyInfo::image_barrier_t barriers[] = {
+			{
+				.barrier = {
+					.dep = {
+						.srcStageMask = PIPELINE_STAGE_FLAGS::COMPUTE_SHADER_BIT,
+						.srcAccessMask = ACCESS_FLAGS::SHADER_WRITE_BITS,
+						.dstStageMask = PIPELINE_STAGE_FLAGS::ALL_COMMANDS_BITS,
+						.dstAccessMask = ACCESS_FLAGS::SHADER_READ_BITS
+					}
+				},
+				.image = warpMapImage,
+				.subresourceRange = {
+					.aspectMask = IImage::EAF_COLOR_BIT,
+					.baseMipLevel = 0,
+					.levelCount = 1,
+					.baseArrayLayer = 0u,
+					.layerCount = 1u
+				},
+				.oldLayout = IImage::LAYOUT::GENERAL,
+				.newLayout = IImage::LAYOUT::READ_ONLY_OPTIMAL,
+			}
+		};                
+		cmdBuf->pipelineBarrier(E_DEPENDENCY_FLAGS::EDF_NONE, { .imgBarriers = barriers });
+	}
+
+	if (!cmdBuf->end())
+	{
+		if (auto* logger = logicalDevice->getLogger())
+			logger->log("ScreenShot: failed to end command buffer.", system::ILogger::ELL_ERROR);
+		return;
+	}


responsibility of the user, and they'll have a helper in the form of SSession::getNextBarrier()

devshgraphicsprogramming · 2026-03-25T21:47:49Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+// TODO: Add an option for corner sampling or centered sampling as boolean parameter
+template <typename ScalarT, typename LuminanceAccessorT 


adjust the comment, its the LuminanceAccessor that should tell you if you're supposed to corner sample (because the luma image needs to take that into account at the borders so it had to be made "that way")

devshgraphicsprogramming · 2026-03-25T21:49:57Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+  LuminanceAccessorT _map;
+  float32_t _rcpAvgLuma;
+  float32_t2 _rcpWarpSize;
+  uint16_t2 _mapSize;
+  uint16_t _mip2x1 : 15;
+  uint16_t _aspect2x1 : 1;
+
+  static HierarchicalWarpGenerator<ScalarT, LuminanceAccessorT> create(NBL_CONST_REF_ARG(LuminanceAccessorT) lumaMap, uint32_t2 mapSize, bool aspect2x1)
+  {
+    HierarchicalWarpGenerator<ScalarT, LuminanceAccessorT> result;
+    result._map = lumaMap;
+    result._mapSize = mapSize;
+    // Note: We use mapSize.y here because the currently the map aspect ratio can only be 1x1 or 2x1
+    result._mip2x1 = _static_cast<uint16_t>(findMSB(mapSize.y));
+    result._aspect2x1 = aspect2x1;
+    return result;
+  }


your create is not setting _rcpAvgLuma nor _rcpWarpSize

~~Also what's mapSize ? The size of the lumaMap or the warp ?~~ ofc its Luma size, the luma accessor is called map

also why pass aspect2x1 if you can check it from the lumaMap or warpmap size ?

_rcpWarpSize is an unused variable and should be removed

devshgraphicsprogramming · 2026-03-25T21:52:00Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+    hierarchical_image::LuminanceReadAccessor<LuminanceAccessorT, ScalarT> &&
+    concepts::Warp<PostWarpT>
+  )
+struct HierarchicalWarpSampler


just call it ComposedHierarchicalSampler, no warp in the name

devshgraphicsprogramming · 2026-03-25T21:55:16Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+  sample_type generate(domain_type xi) NBL_CONST_MEMBER_FUNC
+  {
+    const warp_sample_type warpSample = _warpGenerator.generate(xi);
+    const WarpResult<codomain_type> postWarpResult = PostWarpT::warp(warpSample.value());


you need to store a postWarp member copied/initialized in create, because while the samplers will have constant methods, they will be allowed to be stateful

devshgraphicsprogramming · 2026-03-25T21:56:14Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+    is_scalar_v<ScalarT> && 
+    hierarchical_image::LuminanceReadAccessor<LuminanceAccessorT, ScalarT> &&
+    concepts::Warp<PostWarpT>
+  )


when #1001 gets merged, this sampler will be a BackwardTractableSampler because PDF is known cheaply in both directions

and your PostWarpT will be required to meet ~~BackwardTractableSampler as well to have a backwardPdf~~ actually it will need to be BijectiveSampler because you need to call generateInverse on it to get a backwardPdf https://github.com/Devsh-Graphics-Programming/Nabla/pull/969/changes#r2991518272

devshgraphicsprogramming · 2026-03-25T21:59:41Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+  density_type backwardPdf(codomain_type codomainVal, scalar_type rcpAvgLuma) NBL_CONST_MEMBER_FUNC
+  {
+    return _map.load(codomainVal) * rcpAvgLuma;


but you already have a _rcpAvgLuma member ?

also you need to use the texelFetch signature, same as your gather, to make sure you sample level 0 and with no bilinear interpolation, so this means turning your floating point cornerSampled normalized UV into integer texel coordinate

const uint16_t2 coord = codomainVal*vector2_type(_lastTexel) + promote<vector2_type>(0.5);

devshgraphicsprogramming · 2026-03-25T22:00:11Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+  density_type forwardPdf(domain_type xi) NBL_CONST_MEMBER_FUNC
+  {
+    return generate(xi).pdf();
+  }


PR #1001 lets you pass a cache_Type from generate to forwardPdf

and you'll be able to stuff rcpPmf in that cache, so here all you'd do would just be return (_mapSize.x*_mapSize.y)/cache.rcpPmf;

Edit: taking into account https://github.com/Devsh-Graphics-Programming/Nabla/pull/969/changes#r2991807597 it would be _lastTexel.x * _lastTexel.y / cache.rcpPmf

devshgraphicsprogramming · 2026-03-25T22:00:35Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+    is_scalar_v<ScalarT> && 
+    hierarchical_image::LuminanceReadAccessor<LuminanceAccessorT, ScalarT>
+  )
+struct HierarchicalWarpGenerator


since thanks to PR #1001 warps will be indistinguishable from BackwardTractableSampler, this can be called HierarchicalSampler or HierarchicalLuminanceSampler and meet BackwardTractableSampler constraints

devshgraphicsprogramming · 2026-03-25T22:14:34Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+    result._map = lumaMap;
+    result._mapSize = mapSize;
+    // Note: We use mapSize.y here because the currently the map aspect ratio can only be 1x1 or 2x1
+    result._mip2x1 = _static_cast<uint16_t>(findMSB(mapSize.y));


_mip2x1 misleading name, _penultimateMip better

devshgraphicsprogramming · 2026-03-25T22:27:01Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+struct HierarchicalWarpSampler
+{
+  using warp_generator_type = HierarchicalWarpGenerator<ScalarT, LuminanceAccessorT>;
+  using warp_sample_type = typename warp_generator_type::sample_type;


could just assert that your PostWarpT::domain_type is same as the HierarhicalSampler::codomain_type (current warp_generator_type::codomain_type)

devshgraphicsprogramming · 2026-03-25T22:28:28Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+  using domain_type = vector2_type;
+  using codomain_type = vector3_type;


domain_type = warp_generator_type::domain_type and codomain_type = PostWarp::codomain_type you can static_assert or require that they have vector_traits<>::dimension 2 and 3 respectively

devshgraphicsprogramming · 2026-03-25T22:32:48Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+  sample_type generate(domain_type xi) NBL_CONST_MEMBER_FUNC
+  {
+    const warp_sample_type warpSample = _warpGenerator.generate(xi);
+    const WarpResult<codomain_type> postWarpResult = PostWarpT::warp(warpSample.value());
+    return sample_type::create(postWarpResult.dst, postWarpResult.density * warpSample.pdf());
+  }
+
+  density_type forwardPdf(domain_type xi) NBL_CONST_MEMBER_FUNC


when #1001 comes, you'll need to put into this struct's cache:

the _warpGenerator.generate's cache members

the hierarchical output sample (since its hidden and intermediate value needed to call postWarp.forwardPdf)

postWarp's cache

devshgraphicsprogramming · 2026-03-25T22:38:22Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+  density_type backwardPdf(codomain_type codomainVal) NBL_CONST_MEMBER_FUNC
+  {
+    return PostWarpT::backwardPdf(codomainVal, _rcpAvgLuma) * _warpGenerator.backwardPdf(codomainVal);
+  }


does your test even compile, its the _warpGenerator (hierarhical sampler) that currently cares about _rcpAvgLuma not the post-warp

anyway neither should care.

Anyway the function is wrong, because you'd need to call uv = _postWarp.generateInverse(codomainVal) to get the uv to feed into _warpGenerator.backwardPdf(uv);

devshgraphicsprogramming · 2026-03-25T22:44:56Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+template <typename ScalarT, typename LuminanceAccessorT, typename PostWarpT 
+  NBL_PRIMARY_REQUIRES(
+    is_scalar_v<ScalarT> && 
+    hierarchical_image::LuminanceReadAccessor<LuminanceAccessorT, ScalarT> &&


LuminanceAccessorT should tell you the PostWarpT since the luma map had to be made with luminance weighed by 1/forwardPdf of the post-warp, so the warp mustl be strictly paired with the luma map

and then in turn the PostWarpT also tells you your scalar_type (its own typedef)

So only one tempalte parameter needed

devshgraphicsprogramming · 2026-03-25T22:52:15Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+template <typename ScalarT, typename LuminanceAccessorT, typename HierarchicalSamplerT, typename PostWarpT 
+  NBL_PRIMARY_REQUIRES(is_scalar_v<ScalarT> &&
+    concepts::accessors::GenericReadAccessor<LuminanceAccessorT, ScalarT, float32_t2> &&
+    hierarchical_image::WarpAccessor<HierarchicalSamplerT, ScalarT> &&
+    concepts::Warp<PostWarpT>)


I think you can pretty much just template on LuminanceAccessorT for same reasons I outlined for the ComposedHierarchicalSampler you don't need PostWarpT or ScalarT

~~and I think there isn't much freedom in HierarchicalSamplerT , you can construct it from the other 3~~

wait, but your HierarchicalSamplerT isn't a HierarchicalWarpGenerator at alll, such a misleading name, its a WarpmapAccessorT

devshgraphicsprogramming · 2026-03-25T22:52:39Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+  using domain_type = vector2_type;
+  using codomain_type = vector3_type;


same as with the ComposedHierarchicalSampler

devshgraphicsprogramming · 2026-03-25T22:53:44Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+  using vector4_type = vector<ScalarT, 4>;
+  using domain_type = vector2_type;
+  using codomain_type = vector3_type;
+  using weight_type = scalar_type;


after PR1001 you'll also need a density_type

devshgraphicsprogramming · 2026-03-26T00:12:14Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+    if (p.x == 0)
+      xi.x = xi.x * scalar_type(0.5) + scalar_type(0.5);
+    if (p.y == 0)
+      xi.y = xi.y * scalar_type(0.5) + scalar_type(0.5);
+    if (p.x == _mapSize.x - 1)
+      xi.x = xi.x * scalar_type(0.5);
+    if (p.y == _mapSize.y - 1)
+      xi.y = xi.y * scalar_type(0.5);
+
+    const vector2_type directionUV = (vector2_type(p.x, p.y) + xi) / _mapSize;


there's a bit of a problem with what you're returning (corner sampled) because the PostWarp is resolution agnostic, it doesn't know about corner sampling and expects a [0,1]^2 Unit Square to map to the Sphere where the boundaries touch

So I'd divide by _mapSize - promote(1) and subtract 0.5 from the p+xi before that so you get a normalized coordinate within the corner sampling

This also means that you return _lastTexel.x * _lastTexel.y /rcpPmf as the PDF instead because of how the UV domain gets shrunk

devshgraphicsprogramming · 2026-03-26T00:23:23Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+  LuminanceAccessorT _map;
+  float32_t _rcpAvgLuma;
+  float32_t2 _rcpWarpSize;
+  uint16_t2 _mapSize;


you shold really store _lastTexel (_mapSize-1) instead of mapSize

devshgraphicsprogramming · 2026-03-26T00:36:13Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+    WarpmapSampler<ScalarT, LuminanceAccessorT, HierarchicalSamplerT, PostWarpT> result;
+    result._lumaMap = lumaMap;
+    result._warpMap = warpMap;
+    result._effectiveWarpArea = (warpSize.x - 1) * (warpSize.y - 1);


warpMap is using a custom concept anyway, add a resolution() getter to it

also store _lastWarpTexel instead of _effectiveWarpArea

devshgraphicsprogramming · 2026-03-26T00:37:55Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+
+template <typename ScalarT, typename LuminanceAccessorT, typename HierarchicalSamplerT, typename PostWarpT 
+  NBL_PRIMARY_REQUIRES(is_scalar_v<ScalarT> &&
+    concepts::accessors::GenericReadAccessor<LuminanceAccessorT, ScalarT, float32_t2> &&


use LuminanceRead accessor, and add a getAvgValue method to the concept

devshgraphicsprogramming · 2026-03-26T00:40:02Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+    const vector2_type interpolant;
+    matrix<scalar_type, 4, 2> uvs; 
+    _warpMap.gatherUv(xi, uvs, interpolant);


you know the warp size, make it your job to compute the interpolant and apply corner sampling, so you give the gatherUv(coord,uvs) method a coord thats in [0.5,WarpMapSize-0.5]^2

Lesss code for user to write and get wrong, see https://github.com/Devsh-Graphics-Programming/Nabla-Examples-and-Tests/pull/257/changes#r2991829432

devshgraphicsprogramming · 2026-03-26T00:45:40Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+    const scalar_type detInterpolJacobian = determinant(matrix<scalar_type, 2, 2>(
+      lerp(xDiffs[0], xDiffs[1], interpolant.y), // first column dFdx
+      yDiff // second column dFdy
+    )) * _effectiveWarpArea;
+
+    const scalar_type pdf = abs(warpResult.density / detInterpolJacobian);


postWarp density, xDiffs, yDiff and interpolant.y go in the cache between generate and forwardPdf

also the uv put into the post warp (for the forward weight function)

devshgraphicsprogramming · 2026-03-26T00:50:23Z

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl

+  weight_type forwardWeight(domain_type xi) NBL_CONST_MEMBER_FUNC
+  {
+    return generate(xi).value();
+  }
+
+  weight_type backwardWeight(codomain_type direction) NBL_CONST_MEMBER_FUNC
+  {
+    vector2_type envmapUv = PostWarpT::inverseWarp(direction);
+    scalar_type luma;
+    _lumaMap.get(envmapUv, luma);
+    return luma * _rcpAvgLuma * PostWarpT::backwardDensity(direction);
+  }


no the forward and backward weights need to be identical, so:

forwardWeight must return lumaMap.get(generateCache.uv) * _rcpAvgLuma * generateCache.postWarpPdf

backwardWeight must do what it does now*

except that right now it doesn't apply corner sampling to envmapUv which it must do

kevyuu added 9 commits November 11, 2025 15:50

Fix small bug in GenericDataAccessor definition

b99ae6e

First draft of Warpmap Generation workgroup implementation

b9537ea

Add warp concept

a737173

Add spherical warp

64349db

Remove envmap accessors.hlsl

e44fcf4

Hierarchical image sampling implementation

9b29dfd

Merge branch 'master' into env_map_importance_sampling

06ae3e4

Merge branch 'master' into env_map_importance_sampling

16ecb52

Remove envmap.hlsl

8d682b9

devshgraphicsprogramming reviewed Dec 21, 2025

View reviewed changes

include/nbl/builtin/hlsl/concepts/warp.hlsl Outdated Show resolved Hide resolved

devshgraphicsprogramming reviewed Dec 21, 2025

View reviewed changes

include/nbl/builtin/hlsl/concepts/warp.hlsl Outdated Show resolved Hide resolved

devshgraphicsprogramming reviewed Dec 21, 2025

View reviewed changes

include/nbl/builtin/hlsl/concepts/warp.hlsl Outdated Show resolved Hide resolved

devshgraphicsprogramming reviewed Dec 21, 2025

View reviewed changes

include/nbl/builtin/hlsl/warp/spherical.hlsl Outdated Show resolved Hide resolved

devshgraphicsprogramming reviewed Dec 21, 2025

View reviewed changes

include/nbl/builtin/hlsl/warp/spherical.hlsl Outdated Show resolved Hide resolved

devshgraphicsprogramming reviewed Dec 22, 2025

View reviewed changes

include/nbl/builtin/hlsl/concepts/warp.hlsl Outdated Show resolved Hide resolved

devshgraphicsprogramming reviewed Dec 22, 2025

View reviewed changes

include/nbl/builtin/hlsl/warp/spherical.hlsl Outdated Show resolved Hide resolved

devshgraphicsprogramming reviewed Dec 22, 2025

View reviewed changes

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl Outdated Show resolved Hide resolved

devshgraphicsprogramming reviewed Dec 22, 2025

View reviewed changes

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl Outdated Show resolved Hide resolved

devshgraphicsprogramming reviewed Dec 22, 2025

View reviewed changes

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl Outdated Show resolved Hide resolved

devshgraphicsprogramming reviewed Dec 22, 2025

View reviewed changes

include/nbl/builtin/hlsl/sampling/hierarchical_image.hlsl Outdated Show resolved Hide resolved

kevyuu added 10 commits December 22, 2025 23:59

Move to sampling namespace and implement backward density

890f7c6

Remove private, public from hierarchical_image

f99c63b

Refactor hierarchical image to keep accessor and common data as member

3ff2791

Refactor hierarchical image to separate binarySearch from Hierarchica…

76ef536

…lImage class

Fix Spherical warp indentation

ef773fd

Add some comment why we add xi to the sample uvs

b9467fe

Merge branch 'master' into env_map_importance_sampling

3682604

WIP

ac1e2f3

Rename uv to coord for LuminanceAccessor concepts

baca1cf

Fix hierarchical_image.hlsl

f12b797

devshgraphicsprogramming reviewed Mar 25, 2026

View reviewed changes

devshgraphicsprogramming reviewed Mar 26, 2026

View reviewed changes

Merge branch 'master' into env_map_importance_sampling

fe91d9a

		// TODO: Add an option for corner sampling or centered sampling as boolean parameter
		template <typename ScalarT, typename LuminanceAccessorT

		using domain_type = vector2_type;
		using codomain_type = vector3_type;

Conversation

kevyuu commented Dec 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Testing

TODO list:

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

devshgraphicsprogramming Mar 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

devshgraphicsprogramming Mar 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

devshgraphicsprogramming Mar 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

devshgraphicsprogramming Mar 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

devshgraphicsprogramming Mar 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

devshgraphicsprogramming Mar 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

devshgraphicsprogramming Mar 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

kevyuu commented Dec 19, 2025 •

edited

Loading

devshgraphicsprogramming Mar 25, 2026 •

edited

Loading

devshgraphicsprogramming Mar 25, 2026 •

edited

Loading

devshgraphicsprogramming Mar 25, 2026 •

edited

Loading

devshgraphicsprogramming Mar 25, 2026 •

edited

Loading

devshgraphicsprogramming Mar 25, 2026 •

edited

Loading

devshgraphicsprogramming Mar 25, 2026 •

edited

Loading

devshgraphicsprogramming Mar 25, 2026 •

edited

Loading

devshgraphicsprogramming Mar 25, 2026 •

edited

Loading

devshgraphicsprogramming Mar 26, 2026 •

edited

Loading