Of itu-t sg16 wp3 and iso/iec jtc1/SC29/WG11

JCTVC-M0124 SCE5.1.5: Inter-layer motion information prediction [T.-D. Chuang, Y.-W. Huang, S. Lei (MediaTek)]
JCTVC-M0160 SCE5 – Results of 5.1.6 on inter-layer motion vector prediction [D. Bugdayci, K. Ugur (Nokia)]

`

JCTVC-M0059 SCE5.1.7: Inter-layer motion data inheritance [Z. Deng, W. Zhang, L. Xu, Y. Han, X. Cai, Y. Chiu (Intel)]

JCTVC-M0047 SCE5: Results of test 5.1.10 on inter-layer motion information prediction [J. Lee, H. Lee, J. W. Kang, J. S. Choi (ETRI)]
JCTVC-M0283 SCE5 5.2.1 on Effectiveness of temporal motion vector prediction (TMVP) [J. Xu, A. Tabatabai, K. Sato, S. Lu (Sony), Y. H. Tan, C. Yeo (I2R)]
JCTVC-M0141 SCE5: Result of SCE5.2.2 [K. Sato, J. Xu (Sony), J. Park, B. Jeon(LG)]
JCTVC-M0392 SCE5: Crosscheck of JCTVC-M0141 on result of SCE5.2.2 [T.-D. Chuang, Y.-W. Huang (MediaTek)] [late]
JCTVC-M0142 SCE5: Buffer Size Estimation of SCE5.2.x [K. Sato, J. Xu (Sony)]

On the usage of motion data buffer for SHVC is being studied under SCE5.2.x.

It is proposed by SCE 5.2.1 that temporal motion prediction be omitted at the enhancement layer to reduce the required buffer size, as the collocated base layer motion information takes part of TMVP in the enhancement layer.

It is proposed by SCE 5.2.2 to postpone motion data compression after encoding/decoding of the enhancement layer, or 2-stage motion data compression for improving coding efficiency.

At the 12^th JCT-VC meeting in Geneva, it was suggested to study on the buffer size requirements for SCE5.2.x proposals.

This contribution provides information on the buffer sizes required for each of SCE5.2.x proposals. Trade-off of buffer sizes and coding efficiency is also studied.

36 bits per MVx (16), MVy (16) and refidx (4), 40 byte aligned

In worst case (which is 1x scaling i.e. SNR scalability and RA, bi-pred) the memory increase relative to EL picture buffer would be approx. 40% in case of 4x4 (uncompressed), 10% in case of 8x8 compressed

It is mentioned by one expert that the memory for uncompressed could potentially be further reduced, as 4x8/8x4 only allow unipred; hypothetically it should not be larger than for 8x8 bi-pred, but additional signalling and logic would be necessary.

The uncompressed case has too large a memory increase.

Further information was requested about the actual additional memory for 8x8 case, as an example some level with 4K video.

Another potential option for further investigation could also be to make the compression ratio dependent on the upscaling factor (e.g. 8x8 only for 2X case).

JCTVC-M0289 SCE5 : Inter-layer motion vector prediction in AMVP [J. Park, B. Jeon (LG)]