Reply to thread

Message

<blockquote data-quote="mm1352000" data-source="post: 746985" data-attributes="member: 82144">The issue is with *encoding* changes in general. My understanding is that format changes (eg. 5.1 &lt;--&gt; 2 channel audio) sometimes happen in the broadcast world, and I think MP could handle that as long as your codecs can handle it (ie. the codecs would be the critical factor). Encoding changes are a completely different kettle of fish because MP would have to reconfigure the client side graph... like you said: support &quot;hostile&quot; changes.Technically speaking, the issue with doing that is that I don't think it would be sufficient for the Colossus. From what I saw in Wile E's server logs, the Colossus sometimes changes the audio encoding flags on the audio ES in the PMT *even though Wile E didn't change input or channel*. We don't want glitches because of random PMT changes - we only want to pay attention to the meaningful ones. I'm not yet sure what bearing the PMT chosen on the server side has on the client side, which is why I wanted to look through Wile E's TsReader logs. If this issue has to be handled in TsReader (or somewhere on the client side) then the any fix is almost certain to have to wait for 1.3.0 - there would just be too much risk in those kinds of changes...</blockquote>

[QUOTE="mm1352000, post: 746985, member: 82144"] The issue is with *encoding* changes in general. My understanding is that format changes (eg. 5.1 <--> 2 channel audio) sometimes happen in the broadcast world, and I think MP could handle that as long as your codecs can handle it (ie. the codecs would be the critical factor). Encoding changes are a completely different kettle of fish because MP would have to reconfigure the client side graph... like you said: support "hostile" changes. Technically speaking, the issue with doing that is that I don't think it would be sufficient for the Colossus. From what I saw in Wile E's server logs, the Colossus sometimes changes the audio encoding flags on the audio ES in the PMT *even though Wile E didn't change input or channel*. We don't want glitches because of random PMT changes - we only want to pay attention to the meaningful ones. I'm not yet sure what bearing the PMT chosen on the server side has on the client side, which is why I wanted to look through Wile E's TsReader logs. If this issue has to be handled in TsReader (or somewhere on the client side) then the any fix is almost certain to have to wait for 1.3.0 - there would just be too much risk in those kinds of changes... [/QUOTE]

Verification