[FFmpeg-cvslog] r14684 - trunk/libavcodec/fraps.c

[Mike Melanson]

Mike Melanson wrote:
> Michael Niedermayer wrote:
>> On Sun, Aug 10, 2008 at 08:15:38PM +0200, kostya wrote:
>>> Author: kostya
>>> Date: Sun Aug 10 20:15:38 2008
>>> New Revision: 14684
>>>
>>> Log:
>>> R and B components are stored as a differences to G component in Fraps v5.
>>>
>>> This fixes roundup issue 574.
>> [...]
>>> +                U = f->data[0][0 + i*3 + j*f->linesize[0]];
>>> +                Y = f->data[0][1 + i*3 + j*f->linesize[0]];
>>> +                V = f->data[0][2 + i*3 + j*f->linesize[0]];
>>> +                R = Y + (int8_t)U;
>>> +                G = Y;
>>> +                B = Y + (int8_t)V;
>>> +                f->data[0][0 + i*3 + j*f->linesize[0]] = R;
>>> +                f->data[0][1 + i*3 + j*f->linesize[0]] = G;
>>> +                f->data[0][2 + i*3 + j*f->linesize[0]] = B;
>> f->data[0][0 + i*3 + j*f->linesize[0]] += f->data[0][1 + i*3 + j*f->linesize[0]];
>> f->data[0][2 + i*3 + j*f->linesize[0]] += f->data[0][1 + i*3 + j*f->linesize[0]];
> 
> Would this affect endian-ness? Because I can't get the Fraps v5 decoder 
> to produce the same output on little- vs. big-endian. I have disabled 
> the FATE test until this is fixed.
> 

The exact test:

http://fate.multimedia.cx/index.php?test_spec=127

$BUILD_PATH/ffmpeg -i 
$SAMPLES_PATH/fraps/fraps-v5-bouncing-balls-partial.avi -f framecrc -

This revised test is based on the file formerly entitled FPS1_garbled.avi.

-- 
	-Mike Melanson