Skip to content

Latest commit

 

History

History
87 lines (59 loc) · 3.99 KB

luminance-preserving-mapper.md

File metadata and controls

87 lines (59 loc) · 3.99 KB

FidelityFX Luminance Preserving Mapper

FidelityFX Luminance Preserving Mapper (or LPM) is a tone and gamut mapping solution for high dynamic range (HDR) and wide gamut content. FidelityFX LPM tone maps the luma of the red-green-blue (RGB) pixel instead of the color itself, but it ensures that tonemap(luma(RGB)) would be very similar to luma(tonemap(RGB)), preserving the luminance of the pixel.

For details on the techniques that underpin the FidelityFX LPM effect you can refer to the respective technique documentation.

FidelityFX LPM off

NoLPM

FidelityFX LPM on

LPM

Requirements

  • Windows
  • DirectX(R)12
  • Vulkan(R)

UI elements

The sample contains various UI elements to help you explore the techniques it demonstrates. The table below summarises the UI elements and what they control within the sample.

Element name Value Description
IBLFactor 0..1 Contribution of image-based lighting to the lit scene.
Soft Gap 0..0.5 Controls how much to feather region in out-of-gamut mapping. Range from 0 to a little over zero, 0 == clip.
HDR Max 1..FLT_MAX Brightest linear input RGB pixel value.
LPM Exposure 1..2^n=FLT_MAX Set n to how many times you need to double 18% mid level on the input.
Contrast 0..1 Input range {0.0 (no extra contrast) to 1.0 (maximum contrast)}.
Shoulder Contrast 1..1.5 Shoulder shaping, 1.0 = no change (fast path)
SaturationRGB -1..1 A per channel adjustment, use < 0 decrease, 0 == no change, > 0 increase.
crosstalkRGB 0..1 One channel must be 1.0, the rest can be <= 1.0 but not zero. Lengthens color path to white by walking across gamut.

Setting up the LPM compute shader

The FidelityFX LPM compute shader takes as input the linear color buffer produced by the geometry rendering passes, including the opaque, transparent, and sky box geometry, and also any post processing passes - but before any GUI elements are rendered.

The FidelityFX LPM shader also has some uniform parameters which are passed into the LPM constant setup call described above. They then get passed into the LPM filter call where the tone and gamut mapping happens. Both are defined in ffx_lpm.h.

The ffx_lpm.h header file is designed to be included from both C++ and HLSL/GLSL via #define macros (see below).

C++:

#define FFX_CPU
#include <gpu/ffx_core.h>
#include <gpu/ffx_lpm.h>

HLSL:

#define FFX_GPU
#define FFX_HLSL
#include "ffx_core.h"
#include "ffx_lpm.h"

GLSL:

#define FFX_GPU
#define FFX_GLSL
#include "ffx_core.h"
#include "ffx_lpm.h"

Note that ffx_lpm.h is dependent on ffx_core.h, so both must be included.

Shader options

The FidelityFX LPM shader or more specifically the LpmFilter function has options for using 16-bit floating point (FP16) math versus 32-bit floating point (FP32) math. In general, FP16 math is more efficient than FP32 math on AMD hardware and for FidelityFX LPM it can enable processing two pixels per thread giving double the throughput.

For best performance, the sample compiles these options into the multiple shader permutations and chooses the specific permutations of the shader at runtime depending on the sample's current configuration (see UI Elements section above). In all there are two permutations of the shader:

  • FP16-LpmFilter
  • FP32-LpmFilter

Sample controls and configurations

For sample controls, configuration, and FidelityFX Cauldron Framework UI element details, see Running the samples.

See also