Olivier Crête wrote: > On Fri, 2008-04-11 at 08:37 -0600, Peter Saint-Andre wrote: >> Olivier Crête wrote: >>> On Fri, 2008-04-11 at 06:50 -0600, Peter Saint-Andre wrote: >>>> Paul Witty wrote: >>>>> Peter Saint-Andre wrote: >>>>>> Lauri Kaila wrote: >>>>>> >>>>>>> 2008/4/10, Peter Saint-Andre <[EMAIL PROTECTED]>: >>>>>>> >>>>>>>> Offlist, someone mentioned to me that it might be appropriate for >>>>>>>> XEP-0167 to recommend or even require support for the ITU-T's G.711 >>>>>>>> technology, which is represented by payload-types 0 (PCMU) and 8 >>>>>>>> (PCMA) >>>>>>>> as registered with the IANA. >>>>>>>> http://www.itu.int/rec/T-REC-G.711/e >>>>>>>> >>>>>>>> Thoughts? >>>>>>>> >>>>>>>> >>>>>>> I think 64 kbits/s is too much to be mandatory. >>>>>>> >>>>>> I think the idea was that G.711 is supported everywhere, so recommending >>>>>> support for it in Jingle would help interoperability with existing SIP >>>>>> networks and the PSTN. >>>>>> >>>>> Indeed. We have a Jingle->SIP/H.323 gateway. I don't know of any SIP >>>>> or H.323 endpoints which support Speex, so if that is the only >>>>> recommended codec for Jingle audio, then we have to be prepared to >>>>> transcode the audio in all cases. By guaranteeing a minimum of G.711, >>>>> we should be always able to gateway between protocols without >>>>> transcoding if so desired, even if we may need to transcode for higher >>>>> quality audio e.g. Speex to AAC. >>>> Right, that's the idea. >>>> >>>> Is basic G.711 still patent-encumbered in any way? I'll have to do some >>>> research on that... >>> The G.711 standard was published in 1972, I don't know of any >>> jurisdiction where patents can last that long. >> I don't either. :) A quick search didn't yield any patents outstanding, >> so I think it's probably safe (IANAL etc.). Now the question is whether >> it should be mandatory to implement or just recomended. > > Making it mandatory will be a problem in any low-bandwidth environment > (its still 64kpbs.., compared to 6.4kbps for G.729).
Right. So making it recommended-to-implement might be best. But note that just because it is mandatory-to-implement does not mean it is mandatory-to-deploy or mandatory-to-use. Peter -- Peter Saint-Andre https://stpeter.im/
smime.p7s
Description: S/MIME Cryptographic Signature
