Re: [Cellar] TrackTypes for non-standard data formats

[Dave Rice <[e-mail]>]

> On Apr 27, 2017, at 10:53 PM, Matt Gruenke <[e-mail]> wrote:
> 
> On Wed, Apr 26, 2017 at 4:47 AM, Steve Lhomme <[e-mail] <mailto:[e-mail]>> wrote:
> 2017-02-27 3:21 GMT+01:00 Matt Gruenke <[e-mail] <mailto:[e-mail]>>:
>  
> > In all of the examples I mentioned, I can foresee cases where these would
> > need to be time-stamped frames that might have a different sampling rate
> > than the video or audio streams, if present.  In fact, this is the case for
> > one camera I've used, where the depth sensor operates at a different rate
> > than the primary image sensor.
> 
> There is no 'sampling rate' in Matroska. Only the timestamp of a Block
> matters. So audio/video/other don't need to share anything in that
> regard.
> 
> What I meant by 'sampling rate' wasn't that there would be some exact frame rate.  I simply meant that there might not be a 1-to-1 relationship between frames of different types.  Again, to use the example of a certain camera with a depth sensor, it can capture RGB frames at a different rate than depth frames.
> 
> I mention this simply to point out that the depth frames seem like they belong in their own track.

I regards to this discussion, I ended up reviewing https://tools.ietf.org/html/rfc6648 <https://tools.ietf.org/html/rfc6648> which covers the deprecation of the "X-" prefix as a method to distinguish standardized and unstandardized descriptions. In particular I think the recommendations here https://tools.ietf.org/html/rfc6648#section-3 <https://tools.ietf.org/html/rfc6648#section-3> make sense to follow in the use of Codec IDs, in that a developer of an unstandardized codec id should do so with an assumption that that Codec ID may become standardized.

I sent a pull request for review at https://github.com/Matroska-Org/matroska-specification/pull/117 <https://github.com/Matroska-Org/matroska-specification/pull/117>, which adds this to the Codec Mapping page.

## Recommendations for the Creation of New Codec Mappings

Creators of new Codec Mappins to be used in the context of Matroska:

- SHOULD assume that all Codec Mappings they create might become standardized, public, commonly deployed, or usable across multiple implementations.

- SHOULD employ meaningful values for Codec ID and Codec Name that they have reason to believe are currently unused.

- SHOULD NOT prefix their Codec ID with "X-" or similar constructs.

These recommendations are based upon Section 3 of [@!RFC6648].

Best Regards,
Dave Rice