(20:33:05) Chris: I don't know if you're interested or not, but http://www.youtube.com/watch?v=YJV3xMP24fc is quite a nice demo of some recently added synfig features
(20:42:05) Darco: wow that is cool
(20:42:28) Darco: is it seed based?
(20:42:40) Darco: are the particles individual canvases or something?
(20:42:49) Chris: there's only one particle layer
(20:42:52) Chris: it's just a circle
(20:43:02) Chris: there's a new layer called "duplicate" on top of it
(20:43:11) Darco: of course, but you could make a paramater a canvas
(20:43:21) Darco: and then display the canvas instead of a circle
(20:43:22) Chris: which has parameters "from", "to" and "step" and makes copies of the stuff under it
(20:43:37) Chris: yes, it will duplicate whatever's under it
(20:43:58) Darco: I'm a little confused as to what duplicate does
(20:44:20) Chris: there's a new valuenode type called "random" which lets you vary parameters randomly over time
(20:44:37) Darco: based on a seed, right?
(20:44:47) Chris: http://synfig.org/Duplicate_Layer
(20:45:05) Chris: yes.  based on the same code as the mod_noise code - the noise gradient, etc.
(20:45:22) Chris: but as a valuenode, so it can be applied to any parameter of any layer
(20:46:33) Darco: nice
(20:47:56) Darco: I'm very impressed
(20:48:07) Chris: it can be nested too, to make a 2d array of circles from a single circle: http://synfig.org/forums/viewtopic.php?t=21&highlight=
(20:48:09) Darco: I haven't even looked at all the stuff on youtube
(20:50:24) Chris: this http://youtube.com/watch?v=V9EDol07WLE smoke simulation is another of genete's, using his idea of separate 'particle' layers
(20:50:47) Chris: I think that's from before I added the duplicate layer, so it's a huge 1000 layer composition :)
(21:01:54) Darco: heh, I see you found the xor layer. :P
(21:02:58) Chris: ha, yes.  didn't really get it though.
(21:03:25) Chris: something that has been noticed recently...
(21:03:58) Darco: It was an old, old layer that I used as a pattern for some of my fractal stuff.
(21:04:04) Chris: if we use studio to render an animation, and continue editing the animation while it renders, the changes made during rendering show up in the rendered file.  Is that intentional?
(21:04:29) Darco: No. That's a bug. 
(21:04:36) Chris: ok
(21:04:50) Chris: http://i85.photobucket.com/albums/k74/Genete/synfig/snail-1.png is pretty - another result of using the 'duplicate' layer
(21:04:52) Darco: I worked around it by forking before rendering
(21:05:48) Darco: but making a duplicate synfig composition in memory is expensive and slow. Forks are fast. So that's what we used at Voria.
(21:06:16) Darco: how fast/slow is the duplicate layer when it comes to rendering something like that?
(21:06:28) Chris: but the fork() call didn't get into the source tree in svn?
(21:06:46) Darco: I cannot remember.
(21:06:50) Chris: I think it's much the same speed as if the layers really were duplicated physically using copy/paste
(21:07:07) Chris: it just makes for a much smaller .sif file, and easier maintenance
(21:07:20) Darco: so, very slow. :)
(21:07:25) Darco: but that's ok
(21:08:38) Darco: You could make a render target which uses fork. The trick is that you need to communicate the progress back to the parent process. I used pipes for this.
(21:08:48) Darco: fork won't work on windows though
(21:08:53) Darco: something to keep in mind.
(21:09:42) Darco: I wanted to eventually have some sort of inexpensive composition copy mechanism which would somehow do a copy-on-change thing... but that turned out to be very difficult given the way that I designed things.
(21:09:58) Darco: but I need to get back to work
(21:10:00) Darco: :P
(21:10:02) Chris: ok
(21:10:03) Chris: :)
(21:10:05) Chris: enjoy
(21:10:37) Chris: if you have any of the code lying around that you used to do the fork & talk via pipes, I'd be interested to see it