Hi,

On Fri, Jun 10, 2011 at 12:10 PM, Jason Garrett-Glaser <[email protected]> wrote:
> On Fri, Jun 10, 2011 at 9:04 AM, Ronald S. Bultje <[email protected]> wrote:
>> On Thu, Jun 9, 2011 at 9:55 PM, Jason Garrett-Glaser <[email protected]> wrote:
>>> Updated with a fix to high bit depth wpred.  Ronald has been sent FATE 
>>> samples.
>> [..]
>>> diff --git a/libavutil/pixfmt.h b/libavutil/pixfmt.h
>> [..]
>>> @@ -137,10 +137,14 @@ enum PixelFormat {
>>>      PIX_FMT_BGR48LE,   ///< packed RGB 16:16:16, 48bpp, 16B, 16G, 16R, the 
>>> 2-byte value for each R/G/B component is stored as little-endian
>>>      PIX_FMT_YUV420P9BE, ///< planar YUV 4:2:0, 13.5bpp, (1 Cr & Cb sample 
>>> per 2x2 Y samples), big-endian
>>>      PIX_FMT_YUV420P9LE, ///< planar YUV 4:2:0, 13.5bpp, (1 Cr & Cb sample 
>>> per 2x2 Y samples), little-endian
>>> +    PIX_FMT_YUV444P9BE, ///< planar YUV 4:4:4, 27bpp, (1 Cr & Cb sample 
>>> per 1x1 Y samples), little-endian
>>> +    PIX_FMT_YUV444P9LE, ///< planar YUV 4:4:4, 27bpp, (1 Cr & Cb sample 
>>> per 1x1 Y samples), big-endian
>>>      PIX_FMT_YUV420P10BE,///< planar YUV 4:2:0, 15bpp, (1 Cr & Cb sample 
>>> per 2x2 Y samples), big-endian
>>>      PIX_FMT_YUV420P10LE,///< planar YUV 4:2:0, 15bpp, (1 Cr & Cb sample 
>>> per 2x2 Y samples), little-endian
>>
>> This breaks ABI, you can just add them at the bottom (I know, ordering
>> would be nice, but ABI is also important for users.)
>>
>> For the sws changes, there's currently a line "        if (dstFormat
>> == PIX_FMT_YUV420P9BE || dstFormat == PIX_FMT_YUV420P9LE) {" in
>> swscale.c, which should likely be replaced with "if
>> (av_pix_fmt_descriptors[dstFormat].comp[0].depth_min1 == 8) {" or
>> something along those lines, can you include that in your sws patch?
>>
>> Rest looks OK to me. The h264 patches are big, it would've been nice
>> to split out the "split luma code into its own function" so that the
>> functional patches would've been smaller, but I guess I can live with
>> it. Looking at the code, this should not slow down 8bit or 10bit 420,
>> it'd be nice to confirm that before we apply this...
>
> Can you just apply the swscale portion with the changes you want?

Yes, I was going to do that anyway. ;-).

Ronald
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to