[AVT] replying to a question on timed text in the AVT minutes

Hi, this question:

    Another issue is the ability to repeat modifier fragments and sample
    descriptions: it was asked if this feature should be retained, or
    disallowed in favour of external methods like FEC or retransmission?
    No suggestions were made.

I think that many text streams will be very low-bandwidth, and that 
therefore though the cost of duplication seems expensive in terms of 
the multiplier on the stream (i.e. 100% overhead if everything is 
repeated once) and that FEC could achieve better resilience with a 
lower multiplier, nonetheless the overall effect on the session may 
be negligible as we are doubling a very small number.  It may well be 
the most cost-effective solution from a system point of view, rather 
than implementing a more complex FEC algorithm to save very small 
amounts of bandwidth.  So, I think it is fine to say it is allowed; 
it's really easy for a terminal to detect and discard duplicates. 
And the sender/systems analysis can work out the trade-off (and use 
FEC instead if in fact it is available and appropriate).  Requiring 
the decoder to detect and handle duplicates drops out almost free, so 
seems worth having.
-- 
David Singer
Apple Computer/QuickTime

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