Test with this one. You should get the motion blur on the torus (here from your webcam). Ifnot, there is a broblem somewhere else...
Le 06/10/2013 22:13, Py Fave a écrit : > Hello , i just opened your last patch and am looking at it . > weanwhile i reply to your points > > 2013/10/6 Jack <[email protected]>: >> Hello, >> >> Le 05/10/2013 22:51, Py Fave a écrit : >> >> in fact my problem is not solved . >> it seems like the feedback effect happens everywhere but on the >> surface drawn by the shader . >> >> According to the patch, the feedback is applied on a [square 4] (where there >> is also the [torus 2]). >> Then, [gemframebuffer] 'snap' the whole 'inside' [square 4]. And your torus >> is inside this snapshot (as texels and not vertices). >> > yes ,and then there is some feedback on the texels of the square . > but in my patch if i rotate the torus i get some feedback but not on the > shaded part (only the silhouette of the torus . i don't know why . > i would like everything to go in the feedback loop . It should be OK if you put the shader for the torus on the Gem chain with [gemhead 5] and then let the feedback chains safe like in the patch attached. Tell me if there is a problem now. ++ Jack > > in fact i try to have trails that show "geodesics" on a surface > later i plan to replace my torus with other models . > > > >> i use a torus , in perspective wich has a shader applied . >> >> i tried the following workaround(archive is joined ) , but still no luck . >> the file to open has a OPEN in name . >> >> It seems there is no circle4.geom, [receive13] and [nshader0] >> objects/abstractions in your .zip. > sorry i forgot . > > receive13 only is missing . > it's part of ext13 library (and is in pd-extended iirc) > http://puredata.info/downloads/ext13 > > [receive13 nothing ]could be replaced by receive nShader1 > or all the shader abstraction can be replaced by the classical shader network > with fragment shader = circle4.frag and vertex shader = circle4.vert. > no .geom was used > > >> i would like the dots to have the feedback , >> or dots + geometry alltogether. >> >> I can't see dot in your patch (or gem window), maybe they come from geometry >> shader (which is missing) ? >> > i didn't use geometry shader . just fragment . > glsl program complains in the abstraction but no harm here i think > >> i tried pix_snap2tex ( using gems.feedback from pdmtl abstractions) too >> but no luck because we can't choose the mixing function . >> and i would like only white to make trails . >> and i don't know how to keep the supporting geometry from appearing . >> the torus should be full black with a white trail only . >> >> A simple possibility is to use 2 torus, a black 'on' a white. And use the >> white for the trail ? >> >> An other way : >> >> white torus as texture >> |__> in the feedback __> output feedback as texture >> |__> convert white in black texel __> output black torus as texture >> >> Then mix (multiply) black torus with white feedback to get a black torus >> with white trails. >> ++ >> >> Jack >> > i'll send a modified patch a bit later . > > Thank you so much i see some good things from the last patch you sent . > > it is not what i want to do but opens other directions > and perhaps i was focusing on one method but there seems to be alternatives . > > i'll update this thread tomorrow . > i have to keep some life away from computer :-) > > Thanks for your help > > Py > > _______________________________________________ > [email protected] mailing list > UNSUBSCRIBE and account-management -> > http://lists.puredata.info/listinfo/pd-list
test.pd
Description: application/puredata
// Jack/RYBN 2013
#extension GL_ARB_texture_rectangle : enable
uniform sampler2DRect MyTex1, MyTex2;
void main() {
vec2 tex_coord = gl_FragCoord.st;
vec4 color = texture2DRect(MyTex1, tex_coord);
vec4 color0 = texture2DRect(MyTex2, tex_coord);
color *= 0.01;
color0 *= 0.99;
gl_FragColor = color+color0;
}
_______________________________________________ [email protected] mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
