You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It is possible to use more than one ContentEncoding in a track. But there is nothing that describes how it should be applied. In particular in what order.
Also different ContentEncoding may have a different ContentEncodingScope. But that wouldn't work well if a different ContentEncoding is applied after or before. So there's probably some constraints needed there.
Noone probably ever used 2 ContentEncoding in a single track but the specs allow it. But it might even be useful (see zstd Magic Number header stripping). So we need to clarify this.
The text was updated successfully, but these errors were encountered:
There is actally ContentEncodingOrder which is supposed to tell in which order the list of ContentEncoding has to be applied.
Chrome supports multiple ContentEncoding.
It doesn't support content compression, which makes the multiple support useless, unless there's encryption on different scope or multiple layers of encryption.
The only place it uses the list is to pick the CENC key. And it's only using the first content, without even checking the ContentEncodingOrder values.
Multiple layers may be useful though. You could do compression > header stripping of compressed data > encryption. But there's currently no implementation that supports more than one ContentEncoding (also checked libmatroska 1 and 2 as well as libavformat).
It is possible to use more than one
ContentEncoding
in a track. But there is nothing that describes how it should be applied. In particular in what order.Also different
ContentEncoding
may have a differentContentEncodingScope
. But that wouldn't work well if a differentContentEncoding
is applied after or before. So there's probably some constraints needed there.Noone probably ever used 2
ContentEncoding
in a single track but the specs allow it. But it might even be useful (see zstd Magic Number header stripping). So we need to clarify this.The text was updated successfully, but these errors were encountered: