On Monday 04 Nov 2002 10:51 am, Rolf Dubitzky wrote: > Hi, > > Here a short idea of how I do things in piave. > > I store the render setup in a single tree. This tree is made up of > "MediaElements". There are two Kinds of MediaElements: Nodes and Leaves. > All kinds of audio and video data is a Leave. E.g. a DV file , or a JPEG > image, or an audiofile.
<snip> > Clear? Questions? ;-) You have essentially described what I consider to be a "scene" :-) The only difference that I can see is that in my definition of a scene, I have stated that any videos, pictures, transitions, effects, etc. start and end on scene boundaries. Essentially, a scene never "plays one video for 2 seconds, and then transitions into a second video for 2 seconds". You would replace this with : First scene plays video for 2 seconds. Second scene transitions from first video to second video for 2 seconds. Hopefully, you see that this is essentially just a "break down" of what your idea is into smaller chunks, which can be just as easily merged together again as they were to break up. I've attatched a small picture of the timeline with several clips on it. For arguments sake, assume that there is an operator on each making them slightly transparent. i.e. underlying videos need to interact with the overlying videos. The green lines in the picture represent where scenes would be defined. > Now PropertyPath: Every Operator needs a set of parameters which define its > behaviour. There are static operators which have a single set of parameters > and DynamicOperators. A dynamic operator owns a ProprtyPath, which is a > vector of Properties. Example: binOp1 from above. In it's parents localtime And now you've just described what I call "interpolations". Isn't it nice to know that we agree with each other? :-) In this case, the only problem that I really have is figuring out the best way to describe a PropertyPath (or interpolation) withing > In the timeline these property-nodes must be visible. You must be able to > insert new nodes, delete old nodes and popup a window to edit the values > for a node. Or in other words, I need to write a KeyFrame widget :-) Yes, I intend to do something like this (and the timeline is pretty much designed to accomodate them). It will start off having only linear tranistions, but I would like to add bezier, "step" (where the value simply jumps from one value to another), and then a few more exotic ones, such as "sine wave". The idea is though, that again, the cutter would say in cutter abilities the types of interpolation it supports, and then Kdenlive would make sure that only those ended up in the cutting list. > The usual separation in "transitions" and "effects" makes sense with analog > cutting mashines, it is just nonsense if you have a computer. Hmmm... It depends on what editing analogy you choose for the editor. I haven't made that choice yet for Kdenlive. For instance, Adobe Premier has a seperate "transition" line on the timeline (though I never enjoyed that way of editing because it is awfully implemented). Whilst I can envisage an interface which models a multicamera shoot such as you would do whilst mixing a live show. In that case, having effects and transitions seperate does make sense. Cheers, Jason -- Jason Wood Homepage : www.uchian.pwp.blueyonder.co.uk -------------- next part -------------- A non-text attachment was scrubbed... Name: scene_example.png Type: image/png Size: 2658 bytes Desc: not available URL: <http://mail.kde.org/pipermail/kdenlive/attachments/20021104/d403ac36/attachment.png>
