[asterisk-dev] Dynamic Payloads

[Alexander Traud]

> Is there any advantage to falling back to [the currently used] numbers as defaults for a given format?

If the remote party cares about a dynamic RTP payload type number but ignores its "=rtpmap:", this is a software bug for sure [1][2][3]. Consequently, you are asking whether SIP/SDP implementations are known which are still in use and contain such a software bug.

Yes, I know such an implementation, released last year by a company in the SIP market since day one. If I use another number in the range 96-127, I face no audio on incoming calls at that phone (and then it must be restarted to continue to do any audio). I have to use exactly the type number expected by that phone. This bug was reported and is confirmed. Although this is a severe bug, a resolution is questionable. Therefore, yes, I see an advantage.

[1] <https://tools.ietf.org/html/rfc4566#section-5.14> sub-field <fmt>
[2] <https://tools.ietf.org/html/rfc4566#section-6> attribute =rtpmap:
[3] <https://tools.ietf.org/html/rfc4566#section-8.2.3>