dotRowQ16()

i32 fl::anonymous_namespace{source_xyz.cpp.hpp}::dotRowQ16	(	const i32(&)	row[3],
		u16	r,
		u16	g,
		u16	b )

One matrix row against the pixel.

The accumulator is i64 because the three Q16xQ16 products are Q32: at full scale each term is ~4.3e9, which overflows i32 on its own.

Definition at line 46 of file source_xyz.cpp.hpp.

                                                                     {
    const i64 acc = static_cast<i64>(row[0]) * static_cast<i64>(r)
                  + static_cast<i64>(row[1]) * static_cast<i64>(g)
                  + static_cast<i64>(row[2]) * static_cast<i64>(b);
    // Round to nearest rather than truncating: truncation biases every
    // pixel downward, and the bias accumulates across the later stages.
    return static_cast<i32>((acc + 32768) >> 16);
}

References dotRowQ16(), fl::FL_NO_EXCEPT, and fl::g.

Referenced by dotRowQ16().

Here is the call graph for this function:

Here is the caller graph for this function: