got rudimentary kernel scaling to work

TODO: strides, padding-aware FFT, push constant padding removal

Author: devshgraphicsprogramming

examples_tests/49.ComputeFFT/image_first_fft.comp

@@ -8,8 +8,8 @@ layout(set=0, binding=0) uniform sampler2D inputImage;
 nbl_glsl_complex nbl_glsl_ext_FFT_getPaddedData(in ivec3 coordinate, in uint channel) 
 {
 	ivec2 inputImageSize = textureSize(inputImage, 0);
-	vec2 normalizedCoords = (vec2(coordinate.xy) + vec2(0.5f)) / vec2(inputImageSize);
-	vec4 texelValue= textureLod(inputImage, normalizedCoords, 0);
+	vec2 normalizedCoords = (vec2(coordinate.xy)+vec2(0.5f))/(vec2(inputImageSize)*KERNEL_SCALE);
+	vec4 texelValue = textureLod(inputImage, normalizedCoords+vec2(0.5-0.5/KERNEL_SCALE), -log2(KERNEL_SCALE));
 	return nbl_glsl_complex(texelValue[channel], 0.0f);
 }
 #define _NBL_GLSL_EXT_FFT_GET_PADDED_DATA_DEFINED_

examples_tests/49.ComputeFFT/main.cpp

@@ -25,20 +25,23 @@ constexpr uint32_t channelCountOverride = 3u;
 inline core::smart_refctd_ptr<video::IGPUSpecializedShader> createShader(
 	video::IVideoDriver* driver,
 	const FFTClass* fft,
-	const char* includeMainName)
+	const char* includeMainName,
+	float kernelScale = 1.f)
 {
 	const char* sourceFmt =
 R"===(#version 430 core
 
 #define _NBL_GLSL_WORKGROUP_SIZE_ %u
 #define _NBL_GLSL_EXT_FFT_MAX_DIM_SIZE_ %u
 #define _NBL_GLSL_EXT_FFT_HALF_STORAGE_ %u
+
+#define KERNEL_SCALE %f
  
 #include "%s"
 
 )===";
 
-	const size_t extraSize = 4u+8u+128u;
+	const size_t extraSize = 4u+8u+8u+128u;
 
 	constexpr uint32_t DEFAULT_WORK_GROUP_SIZE = 256u;
 	auto shader = core::make_smart_refctd_ptr<ICPUBuffer>(strlen(sourceFmt)+extraSize+1u);
@@ -47,6 +50,7 @@ R"===(#version 430 core
 		DEFAULT_WORK_GROUP_SIZE,
 		fft->getMaxFFTLength(),
 		fft->usesHalfFloatStorage() ? 1u:0u,
+		kernelScale,
 		includeMainName
 	);
 
@@ -235,7 +239,7 @@ int main()
 		kerImgViewInfo.format = kerImgViewInfo.image->getCreationParameters().format;
 		kerImgViewInfo.subresourceRange.aspectMask = static_cast<IImage::E_ASPECT_FLAGS>(0u);
 		kerImgViewInfo.subresourceRange.baseMipLevel = 0;
-		kerImgViewInfo.subresourceRange.levelCount = 1;
+		kerImgViewInfo.subresourceRange.levelCount = kerImgViewInfo.image->getCreationParameters().mipLevels;
 		kerImgViewInfo.subresourceRange.baseArrayLayer = 0;
 		kerImgViewInfo.subresourceRange.layerCount = 1;
 		kerImageView = driver->createGPUImageView(std::move(kerImgViewInfo));
@@ -342,15 +346,17 @@ int main()
 		);
 	}();
 
+	float bloomScale = 1.f;
 	const auto kerDim = kerImageView->getCreationParameters().image->getCreationParameters().extent;
-	const auto paddedSrcDim = [srcDim,kerDim]() -> auto
+	const auto paddedSrcDim = [srcDim,kerDim,bloomScale]() -> auto
 	{
 		auto tmp = srcDim;
-		tmp.width += kerDim.width-1u;
-		tmp.height += kerDim.height-1u;
-		tmp.depth += kerDim.depth-1u;
+		tmp.width += kerDim.width*bloomScale-1u;
+		tmp.height += kerDim.height*bloomScale-1u;
+		tmp.depth += kerDim.depth*bloomScale-1u;
 		return tmp;
 	}();
+	bloomScale = 0.5;
 	constexpr bool useHalfFloats = true;
 	// Allocate Output Buffer
 	auto fftOutputBuffer_0 = driver->createDeviceLocalGPUBufferOnDedMem(FFTClass::getOutputBufferSize(useHalfFloats,paddedSrcDim,srcNumChannels));
@@ -529,7 +535,7 @@ int main()
 		// Ker Image FFT X
 		auto fft_x = core::make_smart_refctd_ptr<FFTClass>(driver, kerDim.height, useHalfFloats);
 		{
-			auto fftPipeline_ImageInput = driver->createGPUComputePipeline(nullptr,core::smart_refctd_ptr(imageFirstFFTPipelineLayout),createShader(driver,fft_x.get(),"../image_first_fft.comp"));
+			auto fftPipeline_ImageInput = driver->createGPUComputePipeline(nullptr,core::smart_refctd_ptr(imageFirstFFTPipelineLayout),createShader(driver,fft_x.get(),"../image_first_fft.comp",bloomScale));
 			driver->bindComputePipeline(fftPipeline_ImageInput.get());
 			driver->bindDescriptorSets(EPBP_COMPUTE, imageFirstFFTPipelineLayout.get(), 0u, 1u, &fftDescriptorSet_Ker_FFT_X.get(), nullptr);
 			FFTClass::dispatchHelper(driver, imageFirstFFTPipelineLayout.get(), fftPushConstants[0], fftDispatchInfo[0]);