[FFmpeg-devel] [PATCHv3] VP4 video decoder

Peter Ross pross at xvid.org
Fri May 17 13:10:30 EEST 2019


On Thu, May 16, 2019 at 10:49:41AM -0300, James Almer wrote:
> On 5/16/2019 8:35 AM, Peter Ross wrote:
> > ---
> > 
> > what's changed:
> > * reordered AV_CODEC_ID_VP4
> > * minor read_mb_values improvement (reproducible 0.5% speedup)
> > * configure/Makefile fix
> > * squished data tables to fill 80 columns.
> > 
> > UNCHECKED_BITSTREAM_READER doesn't give any performance improvement
> > on my i7, even with unrealisticly huge 4k vp4 video files.
> 
> I hope that was a typo and you tested CACHED_BITSTREAM_READER, not
> UNCHECKED_BITSTREAM_READER :p
> 
> In any case, I wouldn't bother. Ricing a VP4 decoder is hardly worth
> anyone's time.

oops, yes. and i agree.

> > @@ -2825,6 +2825,7 @@ vc1image_decoder_select="vc1_decoder"
> >  vorbis_decoder_select="mdct"
> >  vorbis_encoder_select="audio_frame_queue mdct"
> >  vp3_decoder_select="hpeldsp vp3dsp videodsp"
> > +vp4_decoder_select="hpeldsp vp3dsp videodsp"
> 
> Theora enables the vp3 decoder instead of only the related modules
> because by compiling vp3.c it's already compiling the entirety of the
> vp3 decoder, so might as well have it enabled and available. IMO for vp4
> you should do the same.
> 
> Also, big vp4 specific functions and big chunks of code within shared
> functions should ideally be wrapped inside CONFIG_VP4_DECODER checks.

thanks for spotting these. fixed and fixed.

-- Peter
(A907 E02F A6E5 0CD2 34CD 20D2 6760 79C5 AC40 DD6B)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: not available
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20190517/f92ff405/attachment.sig>


More information about the ffmpeg-devel mailing list