JVET-D0008 JVET AHG report: 360 video test conditions (AHG8) [J. Boyce, E. Alshina, G. v. d. Auwera, P. Philippe, Y. Ye]

The AHG used the main JVET reflector, jvet@lists.rwth-aachen.de, with an [AHG8] indication on message headers. There were approximately 35 email messages.

The availability of two software packages for conversions between projection formats was announced on the reflector, and the packages were made available to the participants. Features were added to the software packages based upon requests from AHG participants.

Several different projection formats were discussed, including equirectangular projection (ERP), several variants of cube map projection (CMP), octahedron projection (OHP), icosahedron projection (ISP), Castor panoramic projection (CPP), equal area projection (EAP) and truncated square projection (TSP). Additionally, alternatives were discussed regarding how to properly compare objective quality metrics which used different projection methods, and how the projection conversion process itself impacted quality. Some HM encodings were done and preliminary results made available on the reflector. Many input contributions to this meeting contain additional experimental results.

Newly available 360 video content was announced.

There were 30 contributions identified as related to 360 video. Some of these contributions go beyond the scope of the AHG mandates, but were listed in the report for completeness.

The AHG recommended to review the relevant input contributions. The AHG also recommended that the following items be worked on during the meeting:

• Arrange viewing sessions of 360 video test sequences
• Define initial 360 video test conditions in an output document, including
  • Identify initial 360 video test sequence set
  • Identify software and configuration to be used for projection conversions
  • Select one or more projection formats to be used as input to the encoder
  • Select one or more objective quality metrics to be reported

A BoG (coordinated by J. Boyce) was asked to work on the aspects from the recommendations above, plus

• Assess inputs on performance of JEM on 360 content
• Investigate inputs on specific coding tools for this type of content, for getting an initial idea about additional compression performance