Re: [AVT] Application of RFC3497 in pro-MPEG work

[Stephen Casner <casner at acm.org>]

Chuck,

> FYI - RFC3497, RTP Payload Format for SMPTE 292M Video - has been
> cited in a commercial Code of Practice issued by the Pro-MPEG forum.

Thanks for this notice.  It is pleasing to see our work used, but I
am concerned about one aspect of the adaptation:

> The RTP format proposed is that from RFC3497 with the following
> extensions:
>  - For Standard Definition, the RTP timestamps shall be derived
>    from a 27MHz clock locked to the input, not from the 148.5MHz
>    clock required by RFC3497.

This is not really a problem since the clock rate can be specified in
SDP.  We have had other examples of payload formats where new
applications were identified that needed a clock rate not anticipated
in the payload format specification.

>  - For other video systems the RTP timestamp shall be the word
>    clock rate of the video system.

Ditto.

>  - To assist with picture concealment in the event of lost packets,
>    the data in a packet shall only come from one line1 of video.
>    As an implication of this the EAV field will always occur
>    at the start of a packet.

Application profiling, again not a problem.

>  - In addition to this, there shall be a custom RTP header
>    extension to assist with defining where in a line a particular
>    packet starts which equipment which complies with this code
>    of practice shall adopt.

This is unfortunate.  If the additional information is really needed,
then a new payload format should have been defined to add this
information to the payload header.

                                                        -- Steve

_______________________________________________
Audio/Video Transport Working Group
avt at ietf.org
https://www1.ietf.org/mailman/listinfo/avt