Number
|
Session
|
Title
|
Source
|
Dispositions
|
m32255
|
File Format
|
Sync Sample Signalling in Multi-Layered HEVC Extensions
|
Vinod Kumar Malamal Vadakital, Miska M. Hannuksela
|
Noted
|
m32256
|
File Format
|
Signalling of Region of Interest Enhancement layers in Layered HEVC
|
Vinod Kumar Malamal Vadakital, Miska M. Hannuksela
|
Rejected
|
m32257
|
File Format
|
Proposal for an unified 14496-15 annex for all multi-layered HEVC extensions
|
Vinod Kumar Malamal Vadakital, Miska M. Hannuksela
|
Accepted
N14123
|
m32260
|
File Format
|
File format extensions for multi-layered HEVC sequences
|
Vinod Kumar Malamal Vadakital, Miska M. Hannuksela
|
Noted
|
m32284
|
File Format
|
Clarifications on HEVC Tile Tracks
|
Jean Le Feuvre, Cyril Concolato, Franck Denoual, Frédéric Mazé, Eric Nassor
|
Accepted
N14123
|
m32285
|
File Format
|
Multi-Layer HEVC and Tile Tracks
|
Jean Le Feuvre, Cyril Concolato, Franck Denoual, Frédéric Mazé, Eric Nassor
|
Accepted
N14123
|
m32286
|
File Format
|
Multi-Layer HEVC Dependencies
|
Jean Le Feuvre, Cyril Concolato, Franck Denoual, Frédéric Mazé, Eric Nassor
|
Accepted
N14123
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Summary of discussions
m32255 Sync Sample Signalling in Multi-Layered HEVC Extensions
We note that the ‘ras ‘ sample group provides similar functionality, and perhaps we might never have had a sync sample table if we had had sample groups back then. Rather than re-designing the sync sample box, we would prefer to point people at the mechanism we prefer. Indeed, we have a historical accumulation of tools here, and we could deprecate some in a scalable HEVC context (e.g. pre-roll sample groups). We note that the sync sample table does enable some codec-unaware operations (e.g. deriving parts of timelines, fragmentation at RAP boundaries). We prefer not to re-design the sync sample table.
We should check that part 12 is clear that a sync sample is both a decoding and presentation clean start.
We have ‘rap’, ‘sap’ (proposed), ‘ras’, ‘roll’, ‘sync’ sample groups, and the sync sample table; this is too many; we should try to simplify, with perhaps a new part 12 definition and some deprecation.
We want to study both these problems further.
m32256 Signalling of Region of Interest Enhancement layers in Layered HEVC
It seems that the tracks all need to indicate precisely what they provide, within them. Providing a ‘fast hint’ in the track references seems like a dangerous door to open. If we want fast tagging, couldn’t it, or shouldn’t it, be a tag within the track rather than the nature of the reference.
Dostları ilə paylaş: |