[xiph-rtp] Section 6, Congestion (Re: xiph-rtp Digest, Vol 3, Issue 9)

[Ramón García]

Michael,

The sending of scaling the sending of RR already takes into account
the issue of scaling, as Phil Kerr has pointed out. In fact, this was
the only important change in the RTP protocol from RFC 1889 to RFC
3550.

My only disagreement with that paragraph is that it is almost
redundant. The RFC 3550 already says that clients SHOULD send RR (in
section 6). Perhaps it is useful to put a sentence to emphasize the
importance of these records for bitpeeling.

I would write.

"Although clients SHOULD send RR acording to RFC 3550, in the case of
Vorbis packets these records are specially useful, because the Vorbis
codec is designed to allow to change the bitrate easily essentially by
removing some bytes at the end of a packet (bitpeeling). Therefore,
this SHOULD should be stronger for Vorbis packets."

Ramon