Organisation internationale de normalisation

HEVC coding performance, implementation demonstrations and design analysis (5)

1. HM performance (0)

See the notes for JCTVC-U0051 and section 5.4.

2. RExt performance/verification testing (1)

1.1.1.1.1.1.1.1.19JCTVC-U0183 Visualization of BD-rate interpolation curves for MOS and PSNR results [A. S. Dias, M. Mrak (BBC)]

(Consideration of this topic was chaired by GJS on Thursday 06-24, 17:15-17:20.)

This contribution provided a spreadsheet to assist analysis of subjective quality scores with special treatment of cases where there is limited overlap between the bit rate and MOS curves. This was used for the verification testing of RExt and interlaced video.

It was remarked that some previous contributions have considered that problem for PSNR-based analysis. Further study was encouraged to check the way this was done before in relation to what is provided in this contribution.

1.1.1.1.1.1.1.1.20JCTVC-U0184 Interlaced video and format range extensions verification test results [G. Barroux (Fujitsu), C. Rosewarne (Canon), V. Baroncini (FUB)]

(Consideration of this topic was chaired by GJS on Wednesday 06-24, 12:00-13:00.)

This contribution provides the results of the HEVC interlace verification and the HEVC Range extensions verification tests.

The selected test sequences are first presented. The test conditions are then introduced, followed by a description of the test methods and logistic. Finally, the results are presented and analyzed.

The interlace verification tests showed that overall, HEVC offers a better encoding quality than AVC even when AVC uses twice the bit rate.

The Range extensions tests showed that HEVC consistently provides a better encoding quality than AVC for range extensions content even when AVC is used with double the bit-rate used with HEVC, and especially for 4:4:4 content.

The tests were operated under the conditions specified in JCTVC-T1003.

Performance curves were provided in the report based on the scheme described in JCTVC-U0183.

This contribution was used toward preparation of the output verification test report JCTVC-U1003.

3.