[Kdenlive-devel] XML representation for effects.

[Rolf Dubitzky]

On Monday 08 March 2004 16:30, Dan Dennedy wrote:
> IMO, the SMIL Transition Effects module is perhaps the weakest in the
> specification. It is hard for me to comment on your implementation from
> this vantage point; however, it looks like you are treating an effect as
> a time container, when it is not.

Our treatment is only in syntax different from SMIL. In SMIL you define the 
tarnsition in the head and give it an id. Later you use the transitions named 
by id. We define the transitions in place. This is at the moment easier to 
generate/parse.

> It is more like a filter applied to a
> time region. I believe SMIL only allows them to be applied to media
> objects rather than any time container, 

If that is the case, then we can't live with the limitation. Anyway, it 
doesn't matter much. I for one don't plan to implement anything like SMIL 
player. We have enough of them. Just having the syntax in a similar notation 
helps people (me) to understand more quickly.

> but I would have to dig though
> the schema to know for sure. I was planning to not use SMIL for effects
> anyway. One thing you need to figure out is how to keep track of the the
> clips that have already been rendered for an effect.

I don't think that is a problem.
Step 1 :  don't care
Step 2 :  have each effect in a fast/lowquality and slow/highquality version
Step 3 :  have a caching and locking mechanism for prerendered/unchanges 
          segments

Cheers, Rolf

-- 
contacts: www.physi.uni-heidelberg.de/~dubitzky