NV_post_convert_rounding

EGL_NV_post_convert_rounding

Bryan Eyler, NVIDIA
Daniel Kartch, NVIDIA
Greg Roth, NVIDIA
Mark Vojkovich, NVIDIA
Nicolai de Haan Brogger, NVIDIA
Peter Pipkorn, NVIDIA

Nicolai de Haan Brogger, NVIDIA Corporation (nicolaid 'at' nvidia.com)

Complete

Last Modified 17 Oct 2012
Version 2

EGL Extension #44

Requires EGL 1.0.

This extension is written against the wording of the EGL 1.4
Specification.

This extension defines the conversions for posting operations
when the destination's number of components or component sizes do
not match the color buffer. This extension supports posting a 24 bit
(888) color buffer to a 16 bit (565) destination buffer, posting a
16 bit (565) color buffer to a 24 bit (888) destination buffer, and
posting a component that is present in the source buffer, but not
present in the destination buffer.    

None

In Chapter 3.9.3, replace paragraph 3 with:
    "    For each color component, if the bit depth of the color
     buffer being posted is 24 (888) and the destination buffer is
     16 (565), the lower order bits of the color buffer are
     truncated.

     If the bit depth of the destination buffer is 24 (888) and the
     color buffer being posted is 16 bit (565), a RGB gain and
     rounding operation is applied to the color buffer values prior
     to posting. The destination buffer will contain the rounded
     (nearest) and clamped result of the vector product of [1.03125,
     1.015625, 1.03125] with the RGB values of the color buffer.

     For cases where a component is present in the color buffer but
     the matching component is not present in the destination
     buffer, that component will be dropped."

In Chapter 3.9.3, remove paragraph 5.