Wednesday, April 29, 2009

Call for Evidence on High-Performance Video Coding (HVC)

--excerpted from N10553 which can be found on

1 Introduction

A large quantity of video material is already distributed in digital over broadcast channels, digital networks and packaged media. More and more of this material will be distributed with increased resolution and quality demand.

Technology evolution will soon make possible the capture and display of video material with a quantum leap in quality (temporal and spatial resolution, color fidelity, amplitude resolution). Networks are already finding it difficult to carry HDTV resolution and data rates economically to the end user. Therefore, further data rate increase will put additional pressure on the networks. For example:
  • High-definition (HD) devices (displays and cameras) are affordable for consumer usage today, while the currently available internet and broadcast network capacity is not sufficient to transfer large amount of HD content economically. While this situation may change slowly over time, the next generation of ultra-HD (UHD) contents and devices, such as 4Kx2K displays for home cinema applications and digital cameras, are already appearing on the horizon.
  • For mobile terminals, video quality using resolutions such as QCIF at low frame rates and low bit rates today is largely unacceptable. While the overall data rate will increase with the evolution of 3G/LTE and 4G networks, also the number of users increases simultaneously with their quality demand. Anticipating that lightweight HD resolutions such as 720p or even beyond will be introduced in the mobile sector to provide similar perceptual quality as for the home applications, lack of sufficient data rates as well as the prices to be paid for transmission will remain a problem for the long term.
MPEG has concluded that video bitrate (when current compression technology is used) will go up faster than the network infrastructure will be able to carry economically, both for wireless and wired networks. Therefore a new generation of video compression technology that has sufficiently higher compression capability than the existing AVC standard in its best configuration (the High Profile), is needed. A study has been started on the feasibility of HVC, which is mainly intended for high quality applications, in particular expecting
  • Performance improvements in terms of coding efficiency at higher resolution,
  • Applicability to entertainment-quality services such as HD mobile, home cinema and Ultra High Definition (UHD) TV.
First results that were reported to MPEG indicate that compression technology giving higher compression performance than AVC might exist for such application cases. It is therefore planned to enter a more rigorous evaluation phase, with a Call for Evidence on new high-performance video compression technology potentially followed by a formal Call for Proposals.

Companies and organizations who have developed compression technology that they believe to be better than AVC High Profile are kindly invited to bring such information to MPEG in the context of this Call for Evidence. The main purpose of a possible subsequent Call for Proposals and standards development will be to increase compression performance.

  • Test sequences and AVC High Profile anchors available: 2009-04-30
  • Expression of interest to participate: 2009-06-01
  • Submission of contributions (descriptive document): 2009-06-22
  • Decoded sequences, bitstreams and binary decoders available at London by 2009-06-27
  • Evaluation of responses: July 2009 MPEG meeting (shall be attended by submitters)
  • Depending on the outcome of the Call for Evidence, MPEG intents to issue a Draft Call for Proposals by the end of the July meeting
2 Test cases

Submitters are encouraged but not required to submit results for all test cases. However, submitters are required to provide results for all sequences in a given class.

2.1 Sequence formats and frame rates

Test material is progressively scanned in 4:2:0 color sampling and 8 bit per component.

The classes of video sequences are:
A) Cropped areas of size 2560x1600 taken from the following sequences (frame rates unchanged): Traffic (4096x2048/30), People on Street (3840x2160/30), Park Joy (3840x2160/50).

B) 1920x1080@24 frames per second: ParkScene, Tennis, Kimono1

C) 832x480@30 frames per second (WVGA): Mobisode 2, Keiba, Flowervase

D) Downsampled versions to half spatial resolution (416x240@30fps, WQVGA) of sequences in C)

2.2 Coding Conditions of AVC anchors
Anchors have been generated by encoding the above sequences using an AVC High Profile encoder (JM15.1) with
  • QP settings: 25, 29, 33 and 37 for class A, 25, 28, 31 and 34 for class B, 25, 29, 33 and 37 for classes C, 25, 29, 33 and 37 for classes D
  • GOP with an Intra picture every 24, 28 and 48 pictures for 24 fps, for 30 fps and for 50 fps sequences, respectively
  • Hierarchical B pictures IbBbP coding structure
  • 4 reference pictures
  • RD Optimization enabled
  • RDOQ enabled (fast mode)
  • Weighted prediction enabled
  • Fast motion estimation. (range 128x128)
For more precise information, refer to config files that come with the anchors.

Submissions to the call shall
  • Be encoded to within +/-5 % of the target bit rates defined above
  • Allow for random access at intervals not more than 1 second apart
  • Not use preprocessing
  • Only use postfilter if it is part of the decoding process.

3 Evaluation methodology

The test method most suitable when assessing video at high visual quality levels is the DSCQS test method as described in ITU-R Recommendation BT 500-11. This method is well known and properly described in the text of Recommendation BT-500, therefore this document does not includes further details on this method and its usage.
To speed up the evaluation of results of the Call for Evidence, we are considering a new effective approach to formal subjective testing of video, which is described in Annex C.

In addition, proponents are required to submit an input contribution with documentation of PSNR values (at least average of frame PSNR for each sequence and encoding point, separate for luma and chroma components, as well as Bjntegaard Delta-Rate and Delta-PSNR [1] [2] compared to the anchors) and, if possible, documentation of the compression technology. Furthermore, in addition to the decoded sequences, a binary decoder as well as bitstreams for all submitted test cases shall be brought to the MPEG July meeting, such that the results can be verified.

4 Logistics

Prospective contributors to the Call for Evidence should contact:

Prof. Dr. Jens-Rainer Ohm, chair of MPEG Video Subgroup
RWTH Aachen University, Institute of Communications Engineering
Melatener Str. 23, 52074 Aachen, Germany
Tel. +49-241-8027671, Fax. +49-241-8022196, email

Dr. Vittorio Baroncini,
Senior Researcher, Audio Video Signal Processing Area
Fondazione Ugo Bordoni
Via B. Castiglione, 59 00142 – Rome - Italy
Tel. +39-06-54802134, Fax. +39-06-54804405, email

Expression of interest is requested by sending an email to Prof. Ohm by 2009-06-01 latest. Interested parties are kindly invited to express their intent of participation to the MPEG video chair as early as possible.

MPEG participants should upload their contributions as regular input documents. Prof. Ohm will collect contributions from outside MPEG and make them available via the MPEG document registry.

Test sequences and anchors (including configuration files used for encoding) will be made available by request by contacting one of the persons above.

[1] Gisle Bjontegaard, "Calculation of Average PSNR Differences between
RD curves", ITU-T SG16/Q6, 13th VCEG Meeting, Austin, Texas, USA, April 2001, Doc. VCEG-M33.

[2] Gisle Bjontegaard, "Improvements of the BD-PSNR model", ITU-T
SG16/Q6, 35th VCEG Meeting, Berlin, Germany, 16th - 18th July, 2008, Doc.VCEG-AI11.

No comments: