Re: [MMUSIC] Simulcast: Short timelines, draft impending

[Byron Campen <[e-mail]>]

On 10/1/15 7:05 AM, Harald Alvestrand wrote:
> While waiting for the draft, one thing that puzzled me on third reading
> through this proposal:
>
> On 09/25/2015 04:34 AM, Adam Roach wrote:
>>   *
>>
>>     RIDs can be defined to be unidirectional, so as to allow
>>     implementations to signal the ability to send a different number of
>>     encodings than they can receive.
>>
>>   *
>>
>>     The values for each constraint can be specified to be different in
>>     each direction.
>>
>>
>>   *
>>
>>     The values (that is, identifiers) used for RID are proposed by the
>>     offerer in a session, and are symmetric. The RID values in an answer
>>     must be a subset of those present in the offer.
> I'm a bit confused. How do we have unidirectional but symmetric RIDs?
>
> If the answerer thinks that it is going to need more RIDs than the
> offerer proposes, is he required to start a renegotiation to get what he
> wants?
    Yes. If the offerer has offered to send 3 RIDs, there is no reason 
for the answerer to expect that 4 would be a possible value. Similarly, 
if the offerer has offered to receive 3 RIDs, there is no reason for the 
answerer to expect that 4 would be something that the offerer could make 
use of. The general pattern here is offer what you can send, and what 
you would like to receive, and the answerer can downnegotiate as it sees 
fit.
>
> Spinoff from that: Is it, in general, possible to add more RIDs at
> renegotiation time?
    Yes.

Best regards,
Byron Campen