[FFmpeg-devel] [PATCH] Altivec versions of h264's weight/biweight functions

Guillaume POIRIER poirierg
Tue Jan 6 23:30:35 CET 2009


Hello,

On Tue, Jan 6, 2009 at 6:32 AM, David Conrad <lessen42 at gmail.com> wrote:
> On Jan 5, 2009, at 6:49 PM, Guillaume POIRIER wrote:
>> On Tue, Jan 6, 2009 at 12:29 AM, Jason Garrett-Glaser
>> <darkshikari at gmail.com> wrote:
>>> On Mon, Jan 5, 2009 at 6:25 PM, Guillaume POIRIER <poirierg at gmail.com>
>>> wrote:
>>>> I applied only the biweight part since my video sample didn't use
>>>> those new weight, and I wanted to make damn sure everything worked as
>>>> expected before committing anything.
>>>>
>>>> The remaining part of your patch that's left to be applied is in
>>>> attachment.
>>>>
>>>> Meanwhile, if someone wants to share a sample that uses "weight"
>>>> functions, go right ahead. :)
>>>
>>> NeroAVC.mp4 on samples (vcodecs/h264).
>>
>> Thanks for the sample. The CRC differs after and before the patch.
>>
>> I used the command:
>> ./ffmpeg -i ~/NeroAVC.mp4 -an -f framecrc - > /tmp/log
>>
>> Did I missuse FFmpeg, or is the new routine not bitexact?
>
> It was the same signed/unsigned issue, here's it fixed.

Thanks, it works now. Applied.

Guillaume
-- 
Only a very small fraction of our DNA does anything; the rest is all
comments and ifdefs.

Douglas William Jerrold  - "The only athletic sport I ever mastered
was backgammon."




More information about the ffmpeg-devel mailing list