Joint Collaborative Team on Video Coding (jct-vc) Contribution

6.3HL syntax for range extensions and single-layer HEVC coding (5)

JCTVC-M0042 No display SEI message [J. Boyce, D. Hong, W. Jang (Vidyo)]

Chaired by Y-.K. Wang.

A “no display” SEI message is proposed, which indicates that a coded picture be decoded but not displayed. JCTVC-L0179 proposed the same SEI message as one of several options to enable a middle box to make a change to a coded bitstream to indicate that a particular picture not be displayed. During the Geneva meeting, discussion on several SEI messages was postponed until after the technical freeze of HEVC version 1. The meeting notes for the review of JCTVC-L0179 note at the end, “Tentative plan is to consider creating an SEI message in a future version.” The proposed SEI message is suggested for adoption for the next HEVC release.

It is noted that the functionality provided by the proposal is the same as in the AVC full-frame freeze SEI message, but with simplified persistence characteristics.

Decision: Adopt (into RExt draft), with the persistence scope to be changed to apply to the associated picture.

JCTVC-M0146 VUI extension and SEI for chroma sampling filter [T. Chujoh (Tosiba), K. Kazui (Fujitsu Lab.), P. Topiwala, W. Dai, M. Krishnan (FastVDO LCC.)]

Since chroma formats that are called 4:2:0, 4:2:2 and 4:4:4 are used at the video coding technology, some kinds of chroma format conversions are needed when output video is translated or displayed. Although the sample location of chroma format can be specified at VUI in current draft specification, there is no information regarding chroma sampling filter. This contribution provides a framework to send recommended up-sampling and down-sampling filter coefficients for chroma format conversion by using VUI and SEI message. Especially, it is possible to solve a problem of error accumulation when chroma conversions are repeated.

Proposed method includes a new VUI syntax elements in an SPS extension and an SEI message. If the VUI specifies a user-defined filter, the SEI message is used to carry the user-defined filter coefficients.

It was questioned why a specific entry in the VUI table was proposed for a particular filter that is not specified in any standard, and not just using a user defined entry.

An alternative would be to use only an SEI message instead of a hybrid solution between a VUI and an SEI message.

The contribution does not contain information about the normalization stage.

For further study.

JCTVC-M0281 Additional experiment results for frame packing arrangement SEI message for 4:4:4 content in 4:2:0 bitstreams [Y. Zhang, Y. Wu, S. Kanumuri, S. Sadhwani, G. J. Sullivan, H. S. Malvar (Microsoft)]

M0229 is somewhat related to this.

This contribution proposes a method to extend the use of the frame packing arrangement SEI message to represent 4:4:4 content in nominally 4:2:0 bitstreams. The contribution is an update of the prior contributions JCTVC-K0240 and JCTVC-L0316 that provides additional experiment results. With the proposed method, it is reported that one constituent frame (e.g. in a top-bottom packing or alternating-frame coding scheme) can be decoded compatibly as an ordinary 4:2:0 image, or can be supplemented with the data from another constituent frame to form a complete 4:4:4 image representation. It is proposed to include support for the additional scheme into the frame packing arrangement SEI message in both AVC and HEVC, to facilitate deployment of systems using this method. Since 4:2:0 is the most widely supported format in products, it is asserted that having an effective way of conveying 4:4:4 content through such decoders can provide the substantial benefit of enabling widespread near-term deployment of 4:4:4 capabilities (especially for screen content coding). The proposed method operates by packing the samples of a 4:4:4 frame into two 4:2:0 frames and encoding the two 4:2:0 frames as the constituent frames of a frame packing arrangement. The semantics of 'content_interpretation_type' are extended to signal this packing arrangement. The proposed scheme is asserted to be of high practical value for applications involving screen content. Relative to native 4:4:4 encoding, the proposed scheme can provide the advantage of compatibility with the ordinary 4:2:0 decoding process that is expected to be more widely supported in decoding products.

Proposes to use the frame packing SEI message for a different purpose than stereo support. Proposed for both HEVC and AVC.

Experimental results provided using the screen coding coding content, and also a comparison with the RExt 4:4:4 encoder, although it is noted that this contribution is specifically aimed at providing 4:4:4 when a 4:4:4 codec is not supported. Provides a limited form of chroma format scalability using a main profile decoder.

Images are provided which illustrate some visual quality improvements vs. 4:2:0 coding. A question raised about at what bitrate the proposed method outperforms 4:2:0 coding.

Proposes a signaling method to provide chroma upsampling filter information.

3 new content interpretation types proposed. Downsampling filtering operation for each type is provided, using simple filters.

Contribution does not address interlaced source.

Text provided, but without figures in the text, which would be useful.

For further study. Request to have software and/or processed video sequences provided.

JCTVC-M0126 Spatial substream definition [M. Arena (RAI), P. Sunna (RAI), G. Ballocca (Sisvel Technology)]

Withdrawn.

JCTVC-M0181 SEI message: independently decodable regions based on tiles [Y. Ye, Y. He, Y. He (InterDigital), X. Yang, P. Yue, Y. Zhang (Huawei), M. Horowitz (eBrisk Video)]

This is a follow-up proposal to JCTVC-L0049, JCTVC-K0116 and JCTVC-K0248. At the 11th and 12th JCT-VC meetings, it was proposed to add an SEI message to support independently decodable regions using tiles. Simulation results of the proposed SEI message showed that, for typical use case scenarios, the restrictions on motion compensated prediction as required by the proposed SEI message incur RD performance penalty of 1.4% and 2.5% when loop filter across tile boundary is disabled and enabled, respectively. Subjective viewing of the reconstructed video did not reveal noticeable coding artefacts.

Address tiles only, not slices.

Identifies tile regions, and explicitly lists which tiles are contained within the region.

Some experimental results provided at QP = 22 only.

Refer to discussion in JCTVC-M0235.

JCTVC-M0235 Motion-constrained tile sets SEI message [Y. Wu, G. J. Sullivan, Y. Zhang (Microsoft)]

This contribution proposes a "motion constrained tile sets" SEI message to indicate that inter prediction processes within one or more specified sets of tiles are constrained to reference only regions within each corresponding set of tiles in other pictures. Using the proposed SEI message, it would be possible for a decoder to correctly decode the specified set(s) of tiles within the pictures of a coded video sequence without needing to decode the entire content of each picture. It is asserted that the proposed SEI message can enable a form of complexity scalability for region-of-interest decoding and display, provide improved loss robustness, and enable enhanced decoder parallelism.

After submission of a first version of this contribution, the authors became aware of JCTVC-M0181 and the prior contributions that it references (JCTVC-K0116, JCTVC-K0248, and JCTVC-L0049). While the concepts are similar, the authors suggest that the syntax found in this proposal is more appropriate than that in JCTVC-M0181, as this contribution is asserted to enable rectangular regions of tiles to be grouped in a more logical and efficient fashion. Additionally, this contribution proposes a syntax element establishing a relationship with the pan-scan rectangle SEI message (as in the previous motion-constrained slice group SEI message) and proposes a tile set identifier syntax element.

Proposes a rectangular region which contains an integer number of tiles.

Includes a pan scan rect.

Includes a motion constrained tile set id to identify the particular rectangular region.

Decision: Adopt (into RExt draft) with revised text to remove pan_scan_rect and change 4 parameters to 2 parameters. A revised version with agreed syntax and semantics to be uploaded in a revision.