Re: Erik Walthinsen's API proposal

ChristianHJW wrote:
> <alex@xxxxxxxxxx> schrieb im Newsbeitrag
> news:20021004235646.A25430@xxxxxxxxxxxxxxxxxxxxxx
>
> > Keep in mind that his "middle tier", being underneath UCI, is entirely
> > codec-specific.  In order to use that level of things in MCF you'd have to
> > write codec-specific code into MCF for every format you wanted to use with
> it
>
> > (this is exactly what UCI (which fills the role of the top tier of Erik's
> > model) is intended to avoid).
> So i misunderstood his intentions here. While UCI may be the perfect
> solution for codec handling on MCF, we still need another 'layer' to be able
> to handle other formats, like DV where audio and video are in one stream. I
> thought Erik's 'middle tier' could be used as a solution here.
That's right, so far UCI is OK except for the DV type...
One option would be to use the AV demuxing with an MCF specific API 
(UCI++ ? ;) and rely on UCI for the "simple" formats like "audio" and 
"video".
UCI++ (or whatever) would simply be a demuxer where you could select the 
streams you want to read, and then these streams could use simple UCI 
codecs (or not).



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf