processPixelLinearQ16()

void fl::processPixelLinearQ16	(	const StreamingPipelineQ16 &	pipeline,
		u8	r,
		u8	g,
		u8	b,
		i32(&)	light[3] )

Pre-response linear emitter light for a power prepass.

Source decoding, gamut mapping and the physical solve are exactly the encoder's stages; brightness/response inversion intentionally follow this boundary.

Definition at line 263 of file pipeline.cpp.hpp.

                                                                {
    // One semantic conversion per channel: code straight to linear light.
    // Nothing here writes an RGB8 intermediate, which is what B3 asks and
    // what ci/tests/test_no_rgb8_intermediate.py enforces structurally.
    const u16 linear_r = decodeTransferU16(pipeline.transfer, r);
    const u16 linear_g = decodeTransferU16(pipeline.transfer, g);
    const u16 linear_b = decodeTransferU16(pipeline.transfer, b);
 
    i32 xyz[3];
    linearRgbToXyzQ16(pipeline.source, linear_r, linear_g, linear_b, xyz);
 
    if (pipeline.gamut_policy == GamutPolicy::Clamp) {
        // The caller asked for the cheap answer. Solve and clip, which is
        // `map_clip` from the P7 study -- about 20 dE2000 from the reference
        // where the mapper is at 0.15, and offered only because the binding
        // offers it.
        solveRgbDrivesQ16(pipeline.gamut.solve, xyz, drives);
        for (int i = 0; i < 3; ++i) {
            if (drives[i] < 0) {
                drives[i] = 0;
            } else if (drives[i] > 65536) {
                drives[i] = 65536;
            }
        }
    } else {
        // The mapper owns clamping, so the stages above hand it the honest
        // target even when that is outside the hull.
        mapAndSolveDrivesQ16(pipeline.gamut, xyz, drives);
    }
 
}

References Clamp, decodeTransferU16(), FL_NO_EXCEPT, g, fl::StreamingPipelineQ16::gamut, fl::StreamingPipelineQ16::gamut_policy, linearRgbToXyzQ16(), mapAndSolveDrivesQ16(), fl::GamutMapQ16::solve, solveRgbDrivesQ16(), fl::StreamingPipelineQ16::source, and fl::StreamingPipelineQ16::transfer.

Referenced by processPixelQ16().

Here is the call graph for this function:

Here is the caller graph for this function: