[Cellar] Re: Introductions and a new I-D that may be appropriate for CELLAR

[Sam Whited <[e-mail]>]

Hi Spencer,

Thanks for the reply!

To answer your main question, I believe this is just a profile and no 
changes need to be made to Matroska to support stem files. Whether 
there's any benefit to the cellar working group taking the draft on as 
opposed to me just submitting it to the editor myself I don't know, but 
I do believe there is some value in having these be standardized for use 
in multiple DJ applications since right now there isn't any open 
standard and buying stem files is a bit of a mess depending and depends 
on where you get them.

My current main sticking point is whether Matroska or Ogg is a better 
wrapper for it. I'm currently leaning towards Ogg because of it's 
simplicity and how much easier it is to read and write the files, but 
Matroska has much better support for metadata which is nice, and if I'm 
not mistaken you can seek faster in Matroska (though DJ applications 
would likely pre-compute seek points anyways to avoid any possibility of 
latency).

Also, in either format, I'm struggling to figure out where best to 
indicate that this is a specific profile of the format (webm uses the 
doctype, but that's a violation of the matroska spec if I'm reading this 
correctly). I'll likely just make up some metadata for this, but it 
seems like something for this would already exist that I should imitate.

—Sam


On 2026-03-02 12:17, Spencer Dawkins at IETF wrote:
> Hi, Sam,
>
> You're definitely in the right place to be asking questions.
>
> We talked about this draft last week at our monthly meeting, mostly 
> from the perspective of what CELLAR might need to do.
>
> The meeting minutes are here: 
> https://datatracker.ietf.org/meeting/interim-2026-cellar-02/materials/minutes-interim-2026-cellar-02-202602242000-00, 
> but just to save you a click, I can summarize ...
>
>     Mailing list question about draft-swhited-ogg-stems/
>     Is this draft/technology in scope for CELLAR? It's really a request
>     for a Matroska profile
>     What standardization is actually needed?
>     We have already started mailing list discussion with Sam Whited (the
>     draft-swhited-ogg-stems/ author). We'll have more discussion,
>     and will figure out whether anything needs to be updated in our
>     working group specifications, and whether anything must be written
>     down in order to achieve interoperability
>
>
> Our biggest question is whether there are changes to existing working 
> group documents needed, in order to support STEMS.
>
> Beyond that - if STEMS is a profile, and does not require any changes 
> to Matroska, we can talk about the advantages of producing that 
> profile in CELLAR. One obvious question is whether the profile would 
> help interoperability if it's published by CELLAR, as opposed to just 
> appearing as a web page (there is nothing wrong with web pages!).
>
> But you're engaged with the right people for now, and focused on 
> technical discussion. Please continue!
>
> Best,
>
> Spencer

-- 
Sam Whited
[e-mail]