On 05/21/2011 04:41 AM, Reinhard Tartler wrote:

> On Wed, May 18, 2011 at 22:37:49 (CEST), Justin Ruggles wrote:
> 
>> ---
>> Updated patch for AC3 encoder channel coupling.
>> Added comments, tweaked some settings, improved the coupling coordinate
>> reuse algorithm.
>>
>> Some PEAQ comparisons:
>> http://www.flickr.com/photos/justinruggles/5734287593/in/photostream
>> http://www.flickr.com/photos/justinruggles/5734287547/in/photostream
>> http://www.flickr.com/photos/justinruggles/5734287509/in/photostream
>> http://www.flickr.com/photos/justinruggles/5734836446/in/photostream
>>
>>  doc/encoders.texi         |   33 ++
>>  libavcodec/ac3.h          |    3 +-
>>  libavcodec/ac3dec.h       |    5 -
>>  libavcodec/ac3dec_data.c  |    6 -
>>  libavcodec/ac3dec_data.h  |    1 -
>>  libavcodec/ac3enc.c       |  875 
>> +++++++++++++++++++++++++++++++++++++--------
>>  libavcodec/ac3enc_fixed.c |    2 +-
>>  libavcodec/ac3enc_float.c |    7 +-
>>  libavcodec/ac3tab.c       |    6 +
>>  libavcodec/ac3tab.h       |    1 +
>>  10 files changed, 765 insertions(+), 174 deletions(-)
> 
> Unfortunatly, this patch breaks fate:
> 
> TEST    acodec-adpcm_ima_qt
> --- /home/siretart/libav/libav/tests/ref/acodec/ac3_fixed     2011-05-21 
> 09:21:07.736146462 +0200
> +++ tests/data/regression/acodec/ac3_fixed    2011-05-21 10:40:41.676146302 
> +0200
> @@ -1,2 +1,2 @@
> -0f14801e166819dd4a58981aea36e08b *./tests/data/acodec/ac3.rm
> +d29cc71b62ae4406852f17122e2e9385 *./tests/data/acodec/ac3.rm
>  98751 ./tests/data/acodec/ac3.rm
> make: *** [fate-acodec-ac3_fixed] Error 1
> make: *** Waiting for unfinished jobs....
> 
> 
> If this change is intended, please update the checksums as well.


That's odd. The change is definitely not intended. The fixed-point
encoder isn't using channel coupling. I'll try to find where it went wrong.

Thanks,
Justin

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

Reply via email to