I agree with Colin to a large extend. We are talking about the case of interleaving where we have a certain interleaving pattern. In case of variable duration, any AU-index value (including zero) only occurs again after the applied interleaving pattern "passed sufficiently". If there are two subsequent RTP packets with the same AU-index value, then this is clearly one of the roll-over problems that the referenced rule states "MUST NOT" occur.
But note that the value zero (as any other value) may occur once in a while, but that the AU-index value always differs in subsequent RTP packets in case of variable duration. Which means that the answer to your original question is that you don't need to look further than two subsequent RTP packets.
Best regards,
Jan
Colin Perkins <csp@isi.edu>
2003-02-26 04:49 PM
To: Franceschini Guido <Guido.Franceschini@TILAB.COM>
cc: Jan vanderMeer/EHV/CE/PHILIPS@EMEA3
avt@ietf.org
Subject: Re: [AVT] New draft-ietf-avt-mpeg4-simple-07
Classification:
Guido,
The paragraph you quoted continues with "When transmitting Access Units
of variable duration ... the transmitter MUST use the AU-Index to encode
the index information required for reordering..." which implies non-zero
values to my reading.
Colin
--> Franceschini Guido writes:
>Colin,
>
>your reply makes me very happy. I'd love to see this draft becoming an RFC.
>So apparently it is already specified that AU-Index will never be 0 when
>"useful". That is just perfect -I was assuming that 0 was an acceptable
>coded value for the variable duration case-. I failed however to find such
>a statement. Could you or Jan point me to it ?
>
>Thanks again
>Guido
>
>-----Original Message-----
>From: Colin Perkins [mailto:csp@isi.edu]
>Sent: Wednesday, February 26, 2003 4:19 PM
>To: Franceschini Guido
>Cc: jan.vandermeer@philips.com; avt@ietf.org
>Subject: Re: [AVT] New draft-ietf-avt-mpeg4-simple-07
>
>
>Guido,
>
>--> Franceschini Guido writes:
>>Jan,
>>
>>unfortunately I have to disagree with the text you modified with respect
>>to Colin's text.
>>I am fine with supporting early implementations, but the wording used
>>is, IMHO, not suitable. More specifically you wrote:
>>
>> If the "constantDuration" parameter is not present, then Access
>> Units are assumed to have a variable duration, unless the AU-Index
>> is present and coded with the value 0 in each RTP packet.
>>
>>How can a receiver implement the rule above? Should it observe 1, 10,
>>1000 or infinite packets before determining that the "AU-Index is
>>present and coded with the value 0 in each RTP packet" ?
>
>One packet is sufficient. If the Access Units have variable duration, the
>AU-Index will never be zero, if they have fixed duration it will always be
>zero. The rest of the paragraph you quoted makes this explicit.
>
>>What is the exact legacy that you want to support ? Maybe there is a
>>different mechanism to support it. For example, a specific mode
>>definition might IMPLY a specific "constantDuration" value, without
>>writing this explicitely in the SDP.
>
>Not if it wishes to comply with this specification.
>
>Colin
>_______________________________________________
>Audio/Video Transport Working Group
>avt@ietf.org
>https://www1.ietf.org/mailman/listinfo/avt
>
>
>====================================================================
>CONFIDENTIALITY NOTICE
>This message and its attachments are addressed solely to the persons
>above and may contain confidential information. If you have received
>the message in error, be informed that any use of the content hereof
>is prohibited. Please return it immediately to the sender and delete
>the message. Should you have any questions, please contact us by
>replying to MailAdmin@tilab.com. Thank you
>====================================================================
>_______________________________________________
>Audio/Video Transport Working Group
>avt@ietf.org
>https://www1.ietf.org/mailman/listinfo/avt