2 different camera angles of the same composition

There isn't any way to do multi-pass rendering in QC (you're essentially doing MP here, since you're rendering the same data twice).

Iterators won't work here because particle systems don't work in iterators (lame).

You could edit the composition in Property List Editor, and give both particle system patches the same random seed -- that should make them identical.

Would it be possible to move the camera every other frame to sort of simulate this?

Thanks for the info guys. That's really a shame about the multipass thing but I'm glad I know now and don't have to learn this the long and hard way.

I'll try the plist trick for the particles. That would be great if that makes the particles emit and behave the same.

I'm still really learning my way around Quartz and haven't found all of it's boundries and understand all the execution order stuff yet.

I do a lot of work in Max/MSP/Jitter and am much more fluent in that environment. For instance, I've already made a stereo out thing in Jitter, it's really easy actually, but those who know Jitter, it's not near as fast and efficient as Quartz.

I thought about the idea of modulating a 3d transform object to ping pong between left and right camera positions and have it switching which half it's feeding very fast, but haven't tried it yet. Does that seem feasible? Perhaps it would only get 30fps max because the nature of the operation.

If that method could work, it would eliminate the duplicate patches and simplify that part.

Right after writing my first post, I did some experimenting with Spooky Send and Receive and the "knight vs. goblin" composition. I tried to send Object data from one subpatch "goblin" to another goblin and found although the second subpatch would work, it was delayed as well as had a lower framerate. Instead duplicating the subpatch all together yielded much better results.

I guess it all has to do with QC internal scheduling and although the "Spooky" pair are a godsend, they are at a deferred status compared to the Patchtime driving both subpatches. Is that right?

I can see now I will want to really implement Patch Time into my patches to lock the clocks between the duplicates that are running. The whole gag relies on the two perspectives (left eye, right eye) to draw at the same time.

thanks again

Stuart

This really is an interesting concept.

I've been studying the "splode-timefreeze" demo composition that comes with Kineme Particle Tools. It really illustrates an interesting use of patch time and camera positioning to create that whole frozen time effect from the matrix.

A lot to be learned from this example.

big props to whoever made this composition.

S.

Would it work for your needs if you put one particle system in a render in image, and then patched the output to two sprites and maybe angle them an extremely small amount, and offset z a little?

Also, on another note, if you want to pursue the multiple particle system route... you can just publish all of the inputs on each particle system, and put all of the controls (randoms and such) on the outside of that level, so that both particle systems are fed by the same value data.

Sorry if neither of those things help you any... but you never know.

Thanks gtoledo3

Good suggestions. I considered the render in Image method, but don't think that will work quite right since we are rotating a 2d video plane with the same particles on it. I guess I don't know if the particles on the rotated plane are going to be in the right virtual space over the Z axis to work in the 3d presentation environment. I'll definitely try it.

The hardest part is right now is truly testing it without the 3d screen. I'm considering buying one of those Nvidia Stereo graphics cards and a set of glasses, but even then, that might be in a different output format all together than what I'm after. Those work by shuttering the glasses left and right really fast to get the two different images and the way this 3d LED screen is going to work is by the use of polarizing film and glasses due to cost restrictions. Buying glasses for 20 or even 100 isn't so bad, but for 10,000 + a night is another thing.

I'm beginning to experiment with running this on one machine vs. two each following a third which issues the command to both. Mostly because of the output resolution I'm after and complexity. Quartz is fast, but even it has it's limits.

As always thanks for all the help from the great members on this forum.

Stuart

Wow, editing the random seed in a Plist editor worked like a charm. I just tested it with 2 Particle generators in a single patch and they appear to be exactly the same.

Cheers Chris

S.

If you are using Kineme particles, cant you attach the particle scene output to 2 separate particle renderers, which are within 2 different 3d transforms?

Attached is a very simple example of what I mean.

Thanks Steve

Kineme's "scene" really does put it all in there to allow it to be split into multiple rendering objects.

Thanks so much for the example, this illustrates how simple it could be if working within a Kineme 3d scene.

cheers

Stuart

http://code.google.com/p/dlp-stereo-osx/source/list

This page may be of semi-related interest.