[FFmpeg-cvslog] r9447 - trunk/libavcodec/wmv2.c

[Panagiotis Issaris]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

Ivan Kalvachev wrote:
> 2007/6/27, Michael Niedermayer <michaelni at gmx.at>:
>> Hi
>>
>> On Wed, Jun 27, 2007 at 04:27:48PM +0200, iive wrote:
>>> Author: iive
>>> Date: Wed Jun 27 16:27:47 2007
>>> New Revision: 9447
>>>
>>> Log:
>>> Identify the loop_filter flag as such and remove the unused variables in the context.
>> you are not wmv2 maintainer, you didnt send a patch,
>> you didnt provide any sample which shows artifacts without this
>> and i doubt it is correct as is, wmv2 likely doesnt use the h.263
>> loop filter
> 
> Even using wrong deblocker on decoding gives much better results than
> not using any.
> 
> Encoding with loop_filter set to 1, as it was until now, whole not
> using any filtering at all is obviously wrong. I see encoding with
> enabled "loop_filter" is marked as problematic even for h263, so I
> don't see it as urgent problem.
> And yes,  sorry, it changes the regression, however the encoded file
> should differ in only 1 bit. The decoded output should stay the same.

Should the attached patch be applied to fix the regression tests?



> I'll see if the person who provided me with samples would allow me to
> upload them.
> I think that the 703 in samples.mphq have the loopfilter enabled,
> however it have other problems (especially when playing it with win32
> loader)

With friendly regards,
Takis
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFGiLZ69kOxLuzz4CkRAicQAJ4xf66TazYALUJ8kq0mCURAlMVo9gCcDlRt
nhFt4LSDqFL5WNxWmZ8VNus=
=7JEJ
-----END PGP SIGNATURE-----
-------------- next part --------------
A non-text attachment was scrubbed...
Name: pi-20070702T102303-ffmpeg-update_regression_tests.diff
Type: text/x-patch
Size: 590 bytes
Desc: not available
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-cvslog/attachments/20070702/3812e743/attachment.bin>