On Friday 22 November 2002 05:19 pm, Jason Wood wrote: > The generation of scenes is not really for the benefit of the GUI - it is > for the benefit of the pattern matching scheduler (still to be written :-) > ), which will look at the render list, search for areas of it that have > already been rendered, and replace them with the pre-rendered versions.
Ah, ok, that makes sense. > It sounds like a lot of work, and if you tried to implement this using the > xml structure that Piave is currently using, as we would have to implement > a full pattern checking algorithm ourselves. > By using scenes, you can simply compare the XML tree generated by one scene > to the XML tree generated by another scene. If they are identical, then we > know we don't have to re-render :-) > This is going to be very useful if you have a piece of footage with effects > that you use multiple times (but don't want to render down because you > might change those effects), it is useful for implementing undo > functionality, as we don't need to keep track of what is and isn't in the > scene render cache (yet more functionality in the future) Oh! I start to understand what you mean. I agree, it makes sense to start splitting to scenes when we have such a scheduler. Recognizing such parts in a tree is much more difficult even if you would implement some kind of 'reference' which points to the same subtree or so. > Fine by me, I am going to work on the client interface within Kdenlive > first, before I start looking at rewriting nullcutter. Ok. I'll start then from the other side, implementing the server interface in piave. It'll be a little different from the nullcutter anyway. Cheers, Rolf *************************************************************** Rolf Dubitzky e-mail: Rolf.Dubitzky at Physik.TU-Dresden.de s-mail see http://hep.phy.tu-dresden.de/~dubitzky/ ***************************************************************
