RFC 4103, "RTP Payload for Text Conversation", June 2005
Note: This RFC has been updated by RFC 9071Source of RFC: avt (rai)
Errata ID: 5032
Publication Format(s) : TEXT
Reported By: Gunnar Hellstrom
Date Reported: 2017-06-07
Rejected by: Francesca Palombini
Date Rejected: 2023-11-07
Section 5.2 says:
After an idle period, the transmitter SHOULD set the M-bit to one in the first packet with new text.
It should say:
After an idle period, the transmitter SHOULD set the M-bit to one in the first packet with new text. A number of approaches can be taken for how to compose the initial packets in the session, and the packets sent at resumption after an idle period. In order to harmonize transmitter behavior, and fulfill requirements in RFC 2198 and RFC 4102, transmitters SHOULD apply the following mechanism: Initially in the session and at resumption of transmission after an idle period, when redundancy is used, the packets to send SHOULD contain the same level of redundancy as specified for the session. If redundant data for the specified number of generations is not available for transmission, empty T140blocks SHOULD be inserted in the packet for transmission to make it contain the specified level of redundancy.
RFC 4103 does not exactly specify how to compose the first packets in the session and the packets after an idle period, when redundancy is used in the session.
Even if receivers should be prepared to decode any valid packet composition, it eases interoperability when transmitters behave consistently.
RFC 2198 requires that the redundant format must carry at least the primary and one redundant level. RFC 4102 requires that if different compositions of the payloads in the packet is to be used, then each combination needs to be assigned its own payload type number. Assuming that that includes use of varying levels of redundancy with the same payload in the redundant data, these requirements lead to the recommendation to use the approach documented in the corrected text.
Your comment is not in scope for errata reports, which are meant to collect errors in the documents, things that were actual errors at publication and that would have been fixed at that time had the working group or document authors noticed them -- they were just missed. What you've reported goes beyond what can be done in errata. The change, therefore, if it is to be applied needs to be achieved through a consensus document.