[PD] strategy to apply shaders/motion blur to pmpd in a 1900x800 rectangular gemwin

Cyrille Henry ch at chnry.net
Tue Nov 5 12:35:52 CET 2013


hello Marco,

better than snap2tex, you can use rendering in framebuffer. The framebuffer can directly be rendered as a texture, no snapping is needed.

i did not understand the way you make your motion blur.
I  usually have my model to run at a frequency 5 or 10 time faster than the rendering. then, you can just render 10 primitives for every image...
or make 10 rendering per frame, and average them in the final render.

cheers
c




Le 05/11/2013 12:08, Marco Donnarumma a écrit :
> Hi all,
>
> I have a 1900 x 800 gemwin. I render a few geos on a big [sphere] and a pmpd system on top of it.
>
> I'd need some cpu-friendly motion blur and glow effects on everything rendered in the gemwin.
>
> My strategy so far has been to split the final rendering into two 800x800 squares. That is:
>
> I have two [snap2tex] snapping two separate [gemwin -1] onto two [square]. One [snap2tex] has 0 0 offset, and the other has 800 0 offset. Then, by modulating the opacity of [colorRGB] for the 2 squares I can get motion blur effect.
>
> This works fine, but the same splitting strategy doesn't work for shaders. If I apply a shader to the two [square] the coordinate of the texture are all mixed up.
>
> Also, when applying the shaders the cpu go up +30%, whereas [snap2 tex] takes way less resources. is that normal?
>
> What's the best way to go here?
>
> thanks!
> best,
>
> --
> Marco Donnarumma
> New Media + Sonic Arts Practitioner, Performer, Teacher, Director.
> Embodied Audio-Visual Interaction Research Team.
> Department of Computing, Goldsmiths University of London
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Portfolio: http://marcodonnarumma.com
> Research: http://res.marcodonnarumma.com
> Director: http://www.liveperformersmeeting.net
>
>
> _______________________________________________
> Pd-list at iem.at mailing list
> UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
>



More information about the Pd-list mailing list