-
1. Re: Sample Designs
greenman Jan 20, 2003 7:25 PM (in response to spsa)> I am proposing the initial vertical services for the
> media engine so there you go. They are meant to be
> implemented as ServiceMBeans:
<snip>
There will need to be some sort of mechanism for handling transcodings as well. This would be some sort of registry again, I suspect.
> JMF is going to be used for most of the above.
> Capturing requires a native version of JMF , which
> comes with its native libraries. We need to find a
> proper way to incorporate native libraries into jboss
> ( ???? !?!? ) or decide that a proper installation of
> jmf is needed.
Given that the JMF 'power packs' are only available for a limited number of platforms, AFAIK, we probably shouldn't rely upon them architecturally, but as a quick and easy of making certain media types / conversions available.
Limited linear audio support, for example could be achieved with javax.sound.
I hadn't, to this point, considered on-the-fly capture in this scenario, imagining instead that most people would have been using persisted media chunks. How do you see its major application? For serving up live broadcasts etc.? -
2. Re: Sample Designs
spsa Jan 21, 2003 8:53 AM (in response to spsa)Well ,
For capturing what I was thinking was both live broadcasting and mobile devices. Let's say you have a small device somewhere with limited storage. You can then capture stuff and send it back to your main server for storing.
One other plausible scenario is the following: You have an array of computer doing real time face detection. The face images you get are then send back to a main server for recognition.
SpSa -
3. Re: Sample Designs
spsa Jan 21, 2003 8:54 AM (in response to spsa)Well ,
For capturing what I was thinking was both live broadcasting and mobile devices. Let's say you have a small device somewhere with limited storage. You can then capture stuff and send it back to your main server for storing.
One other plausible scenario is the following: You have an array of computer doing real time face detection. The face images you get are then send back to a main server for recognition.
SpSa