Re: [Cellar] AV1 mapping Matroska

2018-07-02 21:07 GMT+02:00 Michael Niedermayer <michael@niedermayer.cc>:
> On Mon, Jul 02, 2018 at 01:14:02PM +0200, Jerome Martinez wrote:
>> On 02/07/2018 12:51, Steve Lhomme wrote:
>> >2018-07-02 12:34 GMT+02:00 Michael Niedermayer <michael@niedermayer.cc>:
>> >>Hi
>> >>
>> >>On Mon, Jul 02, 2018 at 09:47:45AM +0200, Moritz Bunkus wrote:
>> >>>Hey,
>> >>>
>> >>>>Looking at the term "coded video sequence" (CVS) in other codecs (H.264
>> >>>>and H.265) it seems it's a common term. And for those codec one Segment
>> >>>>correspond to one CVS, with the parameters of that CVS stored in the
>> >>>>CodecPrivate (SPS + PPS for H.264 for example). So we should probably go
>> >>>>in that simple way.
>> >>>I have quite a lot of h.264 samples here, mainly M2TS from DVB, where
>> >>>SPS/PPS change mid-stream, often multiple times. For such files the first
>> >>>occurrences of SPS/PPS make up the AvcC in CodecPrivate, but all key frames
>> >>>are still prefixed with the currently active SPS/PPS — because that's the
>> >>>only way to signal that stuff has actually changed.
>> >>IMHO All SPS/PPS should be in the "global header" (CodecPrivate) which the
>> >>global header applies to, not just the first.
>>
>> Not possible with e.g. real time streaming (you don't know in advance the
>> next SPS/PPS)
>
> In cases where the SPS/PPS can change and it is not known when the global
> header needs to be produced (that is for example real time streaming).
> it follows that there should be no global header or it should not contain
> any SPS/PPS. But only things which are known for the whole stream.

In Matroska this can be achieved by starting a new Segment when the
codec parameters change. That has an impact on all tracks at the same
time so it may not be good/friendly in all cases. We could describe
this scenario in the specs when a `PrevUID` should be used to specify
the new Segment is a continuation of the previous segment. Actually
even the NextUID could be generated for each Segment and written when
it starts. Technically it would be a correct segment hard-linking but
in the context of live streaming (unknown duration per segment).

> I would argue this is kind of the definition of a global header.
> It applies globally to the whole stream.
> Storing only one out of several PPS/SPS in the header
> (which was mentioned above) feels  rather incorrect to me. And any
> software which makes decissions based on SPS data like resolution and
> aspect ratio could benefit from knowing if it needs to scan the whole
> stream for SPS changes or if it can trust the first or global header
>
>
> [...]
> --
> Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB
>
> If you drop bombs on a foreign country and kill a hundred thousand
> innocent people, expect your government to call the consequence
> "unprovoked inhuman terrorist attacks" and use it to justify dropping
> more bombs and killing more people. The technology changed, the idea is old.
>
> _______________________________________________
> Cellar mailing list
> Cellar@ietf.org
> https://www.ietf.org/mailman/listinfo/cellar
>

-- 
Steve Lhomme
Matroska association Chairman