tb definition is not clear on how to calculate top bitrate in case of different adaptation sets.
Example use cases:
(a) Two audio adaptation sets, Klingon stereo (single representation, 128Kbps) and Quenya 5.1 (single representation, 384Kbps. When Klingon audio is played, tb can be reported as either 128 or 384. It should be 128 in this case, but it is not clear in the spec.
(b) Same as above, when 128Kbps is AAC and 384Kbps is E-AC3 (which is not supported on a particular device).
(c) Multi-key encryption: the top bitrate of the whole MPD can be 24Mbps (2160p, 60fps), but a browser client will often be limited to a 2Mbps 540p due to DRM robustness. Therefore the top bitrate for that client should be 2Mbps and if you report it as 24Mbps you are reporting a non-existent network problem.
Can we clarify the text and say something on the lines of "the top bitrate shall only account for representations / rendition available to the client given constraints such as codec, language, and content protection"?
tb definition is not clear on how to calculate top bitrate in case of different adaptation sets.
Example use cases: (a) Two audio adaptation sets, Klingon stereo (single representation, 128Kbps) and Quenya 5.1 (single representation, 384Kbps. When Klingon audio is played, tb can be reported as either 128 or 384. It should be 128 in this case, but it is not clear in the spec. (b) Same as above, when 128Kbps is AAC and 384Kbps is E-AC3 (which is not supported on a particular device). (c) Multi-key encryption: the top bitrate of the whole MPD can be 24Mbps (2160p, 60fps), but a browser client will often be limited to a 2Mbps 540p due to DRM robustness. Therefore the top bitrate for that client should be 2Mbps and if you report it as 24Mbps you are reporting a non-existent network problem.
Can we clarify the text and say something on the lines of "the top bitrate shall only account for representations / rendition available to the client given constraints such as codec, language, and content protection"?