On Thu, 29 Jun 2017 10:25:09 +0100 Mark Thompson <[email protected]> wrote:
> This does actually work already by magic :) > > Both NV12 and P010 surfaces become UNORM R and RG planes, just with a > different size of sample underneath. Use in OpenCL then sees them > identically as planes of single-precision floating point values. > (Technically it isn't quite right for P010 because the low bits don't > necessarily do the right thing, but it's close enough for practical purposes.) You can assume the padding is 0 (or if it's not, that the source actually has higher precision). But scaling to full fixed point range instead of shifting is incorrect. I'd love to see libplacebo to be used for this - but for now it doesn't exist yet, and is OpenGL only. _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
