Hello,
I'm in over my head with glsl shaders right now. The attached patch is a test I'm using to try and figure out how to properly use [texunit( messages, but it only partially works. Instead of a clear image of me in front of my webcam, I get a changing series of solid colors. It responds to some rapid movements, so I know my video is making it to the shader.
The shaders do nothing to alter the image, they are just there for testing purposes. I've tested by running the video straight through the shaders and it displays clear and normal, as expected.
I've read some things about the image size needing to be a power of 2, not sure if that's related.
Thanks, this stuff is nuts!
-Ben
after a very quick look, i noticied that it work if you connect the right output of the framebuffer to the right input of the 2nd texture, in order to set the coordinate.
i thought that Iohannes had recently made some change in order to improve this, but this patch is not working better than before. maybe because the texture is use not the texunit 0???
Cyrille
Ben Baker-Smith a écrit :
Hello,
I'm in over my head with glsl shaders right now. The attached patch is a test I'm using to try and figure out how to properly use [texunit( messages, but it only partially works. Instead of a clear image of me in front of my webcam, I get a changing series of solid colors. It responds to some rapid movements, so I know my video is making it to the shader.
The shaders do nothing to alter the image, they are just there for testing purposes. I've tested by running the video straight through the shaders and it displays clear and normal, as expected.
I've read some things about the image size needing to be a power of 2, not sure if that's related.
Thanks, this stuff is nuts!
-Ben
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
cyrille henry wrote:
after a very quick look, i noticied that it work if you connect the right output of the framebuffer to the right input of the 2nd texture, in order to set the coordinate.
i thought that Iohannes had recently made some change in order to improve this, but this patch is not working better than before.
yea, still everything is a bit in flux here, so i reverted the changes that introduced the automatic handling here. it seems like they need some more thinking about how and when to reset the texture-matrices. so this basically means, it hasn't become much better than it was some months ago. you should still premultiply the texturematrices in the shaders though, because i'm pretty sure they will come at some point in the near future.
mdgasr IOhannes
maybe because the texture is use not the texunit 0???
Thanks for the help Cyrille and IOhannes, that's gotten me up the next step.
-Ben
On Thu, Mar 4, 2010 at 9:39 AM, cyrille henry ch@chnry.net wrote:
after a very quick look, i noticied that it work if you connect the right output of the framebuffer to the right input of the 2nd texture, in order to set the coordinate.
i thought that Iohannes had recently made some change in order to improve this, but this patch is not working better than before. maybe because the texture is use not the texunit 0???
Cyrille
Ben Baker-Smith a écrit :
Hello,
I'm in over my head with glsl shaders right now. The attached patch is a test I'm using to try and figure out how to properly use [texunit( messages, but it only partially works. Instead of a clear image of me in front of my webcam, I get a changing series of solid colors. It responds to some rapid movements, so I know my video is making it to the shader.
The shaders do nothing to alter the image, they are just there for testing purposes. I've tested by running the video straight through the shaders and it displays clear and normal, as expected.
I've read some things about the image size needing to be a power of 2, not sure if that's related.
Thanks, this stuff is nuts!
-Ben
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
hello,
i am having a problem with inserting gemframebuffer in the right way in the attached patch.
as you can see i try to buffer the texture on the big rectangle and put it on a smaller one. the only way i could get a buffer image was to put it in parallel to the a-branch of |t b a| which seems odd to me. furthermore the produced texture is just showing a segment of the whole 1600 600. what am i doing wrong ?
thank you for your help.
best
erich
hi,
found the problem. i had to give a suitable |perspec( message to gemframebuffer so the whole texture would appear on the square.
best
erich
On Mon, 2010-03-08 at 16:17 +0200, erich wrote:
hello,
i am having a problem with inserting gemframebuffer in the right way in the attached patch.
as you can see i try to buffer the texture on the big rectangle and put it on a smaller one. the only way i could get a buffer image was to put it in parallel to the a-branch of |t b a| which seems odd to me. furthermore the produced texture is just showing a segment of the whole 1600 600. what am i doing wrong ?
thank you for your help.
best
erich
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list