The conclusion from the latest meetings and discussions on the list is that there are multiple ways of doing progressive coding in AVIF. The best ways involve multi-layer coding, but it's not completely obvious exactly how it should be done. The following additions to AVIF would be good:
Some pointers on when 'lsel' should be used to expose layers at the container level. Always? Only for larger images where you may want to decode at lower resolution?
Recommended structure of a file for optimal interoperability and progressive decoding. Do we want to recommend a specific layering structure?
Examples
A decoder model that has recommended behavior for how a decoder should handle progressive decoding of the given examples.
The conclusion from the latest meetings and discussions on the list is that there are multiple ways of doing progressive coding in AVIF. The best ways involve multi-layer coding, but it's not completely obvious exactly how it should be done. The following additions to AVIF would be good: