Am Dienstag, den 29.01.2008, 21:27 +0100 schrieb Richard Spindler:
2008/1/29, [EMAIL PROTECTED] <[EMAIL PROTECTED]>:
> > You should note though that the individual render node *must not* hold
> > any time varying data because of thread safety. Any data will be passed
> > in as "frames" (video, audio, control data) with the individual "pull"
> > calls for rendering. The render network is shared, and the backend needs
> > complete freedom for pulling rendered frames as it sees fit.
>
> Problems to keep in Mind:
Richard,
obviously you didn't get my point...
> * Requiring Plugins to provide random access to frames is nice in theory, but
> in
> practice, many iterative Video Filters exist and produce nice Visuals.
I wasn't talking about "theory" but pointed out, that in our current design the
render nodes themselves are stateless, and get a context passed in. If we find
that we need to support sequential processing, then this has to be organized via
the context...
> * Complete Separation of UI and Plugins is nice in theory, but in practice,
> you might want to attach additional UI relevant data to an Effect.
Again, I wasn't talking about "theory" but pointed out we need to accomodate
for
pasing some opaque data owned by the upper layers...
Hermann V.
_______________________________________________
Cinelerra mailing list
[email protected]
https://init.linpro.no/mailman/skolelinux.no/listinfo/cinelerra
