[codec] Discussion around ITU LS

[Cullen Jennings <[e-mail]>]

We received the LS at 

https://datatracker.ietf.org/documents/LIAISON/file1251.pdf

This raises some inter sting points that I think are worth some list discussion.


Section 1. 

Some issues are pointed out with the example code. Can these be fixed? 

WIll the implementation include switching between SILK and CELT modes. 

My understanding is the reference code is not meant to be an optimized version or work on a wide variety of platforms. It is simply supposed to be one implementation that defines the algorithm on a given architecture. Is this correct?

What would people like to see with respect to test vectors? 

Does the WG consider  time shortening / stretching functionality they would want to include in opus as a technique to use for PLC?


Section 2.

The goals of the WG are to develop a RF codec. We have several IPR filings reported. You can find them at 
https://datatracker.ietf.org/ipr/search/?option=document_search&id_document_tag=20667. Not all of them offer RF terms. The IETF does not want to be involved in deterring if the claimed IPR is valid or not - we simple report the claims we have. Many participants of the WG have looked at IPR disclosures and feel this codec meets the goals of the WG. The question we need to ask in LC is do people think this is ready to publish. 


Section 3.

Given the current state of who was willing to step up to do listening tests before the IESG has finalized a codec. I suspect that it will be difficult to get significantly better testing than we have today before moving to a Proposed Standard. Once we have the frozen specification in the form of a Proposed Standard, I expect to see continued work on characterization and testing. Christian has volunteer to be editor on the working group draft to capture these results.