MSU MPEG-4 AVC/H.264 2011 Video Codec Comparison - CALL FOR CODECS

[Dyomich]

Dear video codec developers,

Moscow State University Graphics & Multimedia Laboratory
starts next seventh annual H.264 video codecs comparison.

New this year features:

• Videoconference Encoding Analysis
  Due to the fact that videocoference encoding has some specific
  that is quite different to transcoding tasks we plan to perform
  special analysis for videoconference encoding.
  
• New Metrics
  We will use new objective quality metrics MS-SSIM and 3-SSIM
  in addition to classical SSIM and PSNR.
  
• GPU Analysis
  This year we want to add GPU-encoders analysis to the comparison.
  For this analysis we plan to use top GPU, e.g.
  NVIDIA GeForce GTX 580.
  
• HRD Compliance
  We consider the possibility to perform HRD compliance test for
  video codecs that support this feature.
  
• Transcoding Analysis
  Also we want to propose developers to take part in special
  additional testing mode - transcoding analysis.
  Developer sends us a pair (MPEG-2 decoder + H.264 encoder)
  and we will test it for transcoding use-case we will take
  MPEG-2 source video sequecnes to encode
  them with H.264 encoder.

Important Dates
February, 1 - Deadline for receipt of a H.264 codec with presets
February, 14 - Deadline for settling technical problems with codec
March, 15 - Report's draft will be sent to all participants
March, 22 - Deadline for reception of comments to the draft
April, 7 - Comparison report release

The full text of Call for Codecs is available at
http://compression.ru/video/codec_co...codecs_11.html

Variants of Participation
There are two variants for companies to participate in our comparison:

1. Participation for free. All results of your codec will be
   published, except special cases of measurements
   problems due to codec instability.
   
2. Private participation. A special report will be prepared only
   for your company. This report contains:

   • Your codec results and all material from the free version
   • Special additional analysis of your codec

If you are interested in the private participation,
please contact us for details.

Useful Links

• Call for codecs-2011
  http://compression.ru/video/codec_co...codecs_11.html
  
• Sixth Annual MSU MPEG-4 AVC/H.264 Video Codec Comparison
  http://www.compression.ru/video/code...ison/h264_2010
  
• Options Analysis of MPEG-4 AVC/H.264 Codec x264
  http://compression.ru/video/codec_co...sis_08_en.html

-----
Best regards,
Dr. Dmitriy Kulikov,
Moscow State University (MSU)
Graphics&Media Lab
Videocodec Testing Team
videocodec-testing@graphics.cs.msu.ru

[kieranrk]

Originally Posted by Dyomich

[*] HRD Compliance
We consider the possibility to perform HRD compliance test for
video codecs that support this feature.

I would argue there isn't such thing as "HRD compliance". It's more of a case of whether a particular analyzer doesn't complain that a file is not compliant.

(In my opinion Pengvado was right not to allow a HRD patch in x264 for many years since compliance is a totally arbitrary concept)

I think you should check VBV compliance instead.

[IgorC]

http://compression.ru/video/codec_comparison/h264_2011/

[CruNcher]

Is this a joke they didn't compared Nvidias GPU Encoder ? i mean even a ex MSU guy is working on that geez, and didn't they said if we dont have at least 3 GPU Encoder we wont compare GPU Encoder @ all ?

Though nothing new http://forum.doom9.org/showthread.ph...94#post1477694

[nm]

Originally Posted by CruNcher

Is this a joke they didn't compared Nvidias GPU Encoder ? i mean even a ex MSU guy is working on that geez, and didn't they said if we dont have at least 3 GPU Encoder we wont compare GPU Encoder @ all ?

Maybe they tested it but didn't get permission to publish the results. Wouldn't be surprising, considering how the MainConcept CUDA encoder fared.

[mandarinka]

It's in the pdf. Page 99 and further (in the free to download pdf).

[nm]

Originally Posted by mandarinka

It's in the pdf. Page 99 and further (in the free to download pdf).

MainConcept's CUDA encoder is there, not others. CruNcher wanted to see nvcuvenc results.

[me7]

Too bad QuickTime wasn't included, I'd like to see how it fares compared to a recent build of x264

[Blue_MiSfit]

If I was Apple, I sure as hell wouldn't want anyone doing proper testing of my H.264 encoder!

[CruNcher]

Originally Posted by Blue_MiSfit

If I was Apple, I sure as hell wouldn't want anyone doing proper testing of my H.264 encoder!

HiHi

Originally Posted by nm

MainConcept's CUDA encoder is there, not others. CruNcher wanted to see nvcuvenc results.

Yes that was what i meant, sad that there is also nowhere a reference about which Cuda Encoder was tested from Mainconcept (settings,version)
just a -> MainConcept H.264 (CUDA based encoder) <- that's not much if they said Mainconcept Provided us with the latest High Profile Encoder it would be clear but this way it could be also just the Main Profile encoder, without adaptive b-frames, no 8x8dct (FreXt), no hadamard, no weighted prediction ?

[henryho_hk]

I am afraid XviD encoding parameters (if anyone cares) are not well selected.

In my own SSIM tests of SD size, XviD 1.3(.1)'s "-qpel" is always counter-productive under "-qtype 0 -vhqmode 4". Only when under "-vhqmode 1" and within a certain "mid" bitrate range (40 to 70 under Teegdeck's comp-test), "-qpel" produces a slightly higher SSIM (but probably not high enough to justify the speed impact). Further, under a low bitrate situation (15 to 35), we should use the "EQM v3 LR" custom matrix (without -qpel).