[FFmpeg-devel] [PATCH] avcodec/x86: add an 8-bit simple IDCT function based on the x86-64 high depth functions

[James Darnley]

On 2017-06-16 12:48, Paul B Mahol wrote:
> On 6/16/17, James Darnley <jdarnley at obe.tv> wrote:
>> On 2017-06-16 03:58, Michael Niedermayer wrote:
>>> theres something wrong with this
>>> it totally breaks this:
>>> make -j12 ffmpeg && ./ffmpeg -ss 1 -i cache:matrixbench_mpeg2.mpg -t 2 -y
>>> test.avi
>>> ./ffplay test.avi
>>>
>>> (totally as in bitstream errors not some idct artifacts)
>>
>> How did you manage to break it that much?  I don't touch avformat.  I
>> don't touch the rest of the decoder.
>>
>> I don't have that file to test with and the cut down version which I
>> think is in fate is only 0.96 seconds long.  However trying your command
>> on that file successfully decodes 1 frame from it.  I made decoding it
>> with the new functions the fate test I added.  All the other fate
>> samples in mpeg2/ decode seemingly fine with no extra messages.
> 
> File is in samples.ffmpeg.org/benchmark/testsuite1/matrixbench_mpeg2.mpg
> 
> Michael failed to write that.

TYVM.  Even with that file I get just 3 messages when I seek 1 second.

> [mpeg @ 0x399e3c0] start time for stream 0 is not set in estimate_timings_from_pts
> [mp2 @ 0x39a07a0] Header missing
> Error while decoding stream #0:2: Invalid data found when processing input

Furthermore I get those messages on master and they remain the same with
my commits.  (Stream 0:2 is the mp2 audio in this file.)