Wetterberg

I've just had the single most productive day working with Jitter. It's 

I've been quite upset over the last weeks that I had no clue how to work 

with jit.gl.mesh and matrices together... well; today the dime dropped. 

I've just had the single most productive day working with Jitter. It's 
really really growing on me now!

I've been quite upset over the last weeks that I had no clue how to work 
with jit.gl.mesh and matrices together... well; today the dime dropped. 
Wicked tings.
Will post patch tomorrow.

thank-you-2

Here's yesterdays work - I'm really pleased with the way this came together.

A modular version of this patch, inspired in a way by flying carpets (?)

is going to be part of a bigger system that we are building at

Funny enough it's one of the first patches I've done in a long time that

doesn't use abstractions, but if it still doesn't work for you, please

A modular version of this patch, inspired in a way by flying carpets (?)
is going to be part of a bigger system that we are building at
www.thecovertoperators.org

Funny enough it's one of the first patches I've done in a long time that
doesn't use abstractions, but if it still doesn't work for you, please
let me know. Thanks.

Damn. Sorry about that, attachments to the email list don't seem to make it to the forum.

Bas van der Graaff skrev:
> I'm interested, but where's here?
Damn. Sorry about that, attachments to the email list don't seem to make it to the forum.

Here it is again as a forum attachment.
Hope this works,

> Damn. Sorry about that, attachments to the email list don't seem to make

> Here it is again as a forum attachment.

On 8/14/07, Andreas Wetterberg  wrote:
>
> Bas van der Graaff skrev:
> > I'm interested, but where's here?
> Damn. Sorry about that, attachments to the email list don't seem to make
> it to the forum.
>
> Here it is again as a forum attachment.
> Hope this works,
>
> Andreas.
>
>
>
>

I just realized why I was getting jaggies in the patch though - I had 

forgotten @antialias 1 on the meshes. Doing that will make things look 

Alright, I also realized that my RGB matrix coming in only controls the 

Z position of the mesh dims... Can anyone help me decipher what I've 

done wrong? I have a feeling that the problem may stem from the 

[jit.matrix 3 char 8 8 @planemap 1 2 3] - unfortunately controlmatrices 

are very much Black Box tech to me at this point. Cracking this will 

I'd really appreciate any help in getting this patch (and my next ones) 

> On 8/14/07, *Andreas Wetterberg* < awetterberg@post.cybercity.dk 

> Damn. Sorry about that, attachments to the email list don't seem

> ------------------------------------------------------------------------

I just realized why I was getting jaggies in the patch though - I had 
forgotten @antialias 1 on the meshes. Doing that will make things look 
1000% sweeter.

Alright, I also realized that my RGB matrix coming in only controls the 
Z position of the mesh dims... Can anyone help me decipher what I've 
done wrong? I have a feeling that the problem may stem from the 
[jit.matrix 3 char 8 8 @planemap 1 2 3] - unfortunately controlmatrices 
are very much Black Box tech to me at this point. Cracking this will 
make the patch really kick-ass, though.

I'd really appreciate any help in getting this patch (and my next ones) 
to the next level,

yair reshef skrev:
> thank you Andreas, great patching
>
> On 8/14/07, *Andreas Wetterberg* < awetterberg@post.cybercity.dk 
> > wrote:
>
>     Bas van der Graaff skrev:
>     > I'm interested, but where's here?
>     Damn. Sorry about that, attachments to the email list don't seem
>     to make it to the forum.
>
>     Here it is again as a forum attachment.
>     Hope this works,
>
>     Andreas.
>
>     jitter@cycling74.com 
>
>
>
> ------------------------------------------------------------------------
>
>


Quote: Wetterberg wrote on Wed, 15 August 2007 10:20

----------------------------------------------------

> Alright, I also realized that my RGB matrix coming in only controls the 

> Z position of the mesh dims... Can anyone help me decipher what I've 

> done wrong? I have a feeling that the problem may stem from the 

> [jit.matrix 3 char 8 8 @planemap 1 2 3] - unfortunately controlmatrices 

> are very much Black Box tech to me at this point. Cracking this will 

> make the patch really kick-ass, though.

i'm no expert in control matrices but... to me it looks like [jit.matrix 3 char 8 8 @planemap 1 2 3] is used to generate the texture itself and not the xyz (or just z) position of the dims.

that should be the part below [jit.matrix 3 char 64 64 @interp 1], which goes on to generate 3 gl.mesh... but i'm not convinced there is anything wrong with that. can you explain a bit more what you are trying to do at this point.

here is a patch illustrating a technique i am familiar with, which enables you to convert an RGB char matrix into an XYZ float32 matrix. not quite sure that's what you are after...

Quote: Wetterberg wrote on Wed, 15 August 2007 10:20
----------------------------------------------------
> Alright, I also realized that my RGB matrix coming in only controls the 
> Z position of the mesh dims... Can anyone help me decipher what I've 
> done wrong? I have a feeling that the problem may stem from the 
> [jit.matrix 3 char 8 8 @planemap 1 2 3] - unfortunately controlmatrices 
> are very much Black Box tech to me at this point. Cracking this will 
> make the patch really kick-ass, though.
----------------------------------------------------

that should be the part below [jit.matrix 3 char 64 64 @interp 1], which goes on to generate 3 gl.mesh... but i'm not convinced there is anything wrong with that. can you explain a bit more  what you are trying to do at this point.

> i'm no expert in control matrices but... to me it looks like 

> [jit.matrix 3 char 8 8 @planemap 1 2 3] is used to generate the 

> texture itself and not the xyz (or just z) position of the dims.

> that should be the part below [jit.matrix 3 char 64 64 @interp 1], which goes on to generate 3 gl.mesh...

> but i'm not convinced there is anything wrong with that. can you explain a bit more what you are trying to do at this point.

I am looking to get the RGB values to modulate the coords of the mesh. 

With some fiddling I have managed to get Y and Z movement, but not the 

ability to reach arbitrary positions, like when you send it jit.noise, 

for instance. I am very much flying blind here, but haven't got the info 

on how to correctly translate between 3 char 8 8 and my meshes.

> here is a patch illustrating a technique i am familiar with, which enables you to convert an RGB char matrix into an XYZ float32 matrix. not quite sure that's what you are after...

See, this is where I started, actually. Random values? No problem. But 

as soon as I start using a "real" matrix, I can only get so far.

Thanks for your help, Justin. Really appreciate it.

justin skrev:
> i'm no expert in control matrices but... to me it looks like 
> [jit.matrix 3 char 8 8 @planemap 1 2 3] is used to generate the 
> texture itself and not the xyz (or just z) position of the dims.
> that should be the part below [jit.matrix 3 char 64 64 @interp 1], which goes on to generate 3 gl.mesh...
of course, that's what I meant, too.
>  but i'm not convinced there is anything wrong with that. can you explain a bit more  what you are trying to do at this point.
>
I am looking to get the RGB values to modulate the coords of the mesh. 
With some fiddling I have managed to get Y and Z movement, but not the 
ability to reach arbitrary positions, like when you send it jit.noise, 
for instance. I am very much flying blind here, but haven't got the info 
on how to correctly translate between 3 char 8 8 and my meshes.
> here is a patch illustrating a technique i am familiar with, which enables you to convert an RGB char matrix into an XYZ float32 matrix. not quite sure that's what you are after...
>
See, this is where I started, actually. Random values? No problem. But 
as soon as I start using a "real" matrix, I can only get so far.
Thanks for your help, Justin. Really appreciate it.
Andreas.


Quote: Wetterberg wrote on Wed, 15 August 2007 20:20

> I am looking to get the RGB values to modulate the coords of the mesh. 

> With some fiddling I have managed to get Y and Z movement, but not the 

> ability to reach arbitrary positions, like when you send it jit.noise, 

> for instance. I am very much flying blind here, but haven't got the info 

> on how to correctly translate between 3 char 8 8 and my meshes.

sorry if parts of this are obvious to you, but i will try to break down the task:

part of the problem is the dim of the mesh compared to the dim of the gradient (8x8) - the smaller the dim the less complex the shape.

also part of the problem is that gradients may be too uniform a colour to create noticeably different geometry.

i would suggest you try using the gradient map as a way to modulate a noise source (jit.noise / jit.bfg) which in turn affects the geometry - perhaps this will give you more varied geometry.

Quote: Wetterberg wrote on Wed, 15 August 2007 20:20
----------------------------------------------------
> I am looking to get the RGB values to modulate the coords of the mesh. 
> With some fiddling I have managed to get Y and Z movement, but not the 
> ability to reach arbitrary positions, like when you send it jit.noise, 
> for instance. I am very much flying blind here, but haven't got the info 
> on how to correctly translate between 3 char 8 8 and my meshes.
----------------------------------------------------

> Quote: Wetterberg wrote on Wed, 15 August 2007 20:20

> ----------------------------------------------------

>> I am looking to get the RGB values to modulate the coords of the mesh. 

>> With some fiddling I have managed to get Y and Z movement, but not the 

>> ability to reach arbitrary positions, like when you send it jit.noise, 

>> for instance. I am very much flying blind here, but haven't got the info 

>> on how to correctly translate between 3 char 8 8 and my meshes.

> sorry if parts of this are obvious to you, but i will try to break down the task:

> part of the problem is the dim of the mesh compared to the dim of the gradient (8x8) - the smaller the dim the less complex the shape.

Yes, but I feel that's part of the beauty of it. But even if it was just 

a 3x3 matrix I should still be able to see offsets in all three axes... 

> also part of the problem is that gradients may be too uniform a colour to create noticeably different geometry.

Yeah, you're right. I just sat down and went crazy with the colour 

> i would suggest you try using the gradient map as a way to modulate a noise source (jit.noise / jit.bfg) which in turn affects the geometry - perhaps this will give you more varied geometry.

Good idea, that... my only concern is that I would like to see a close 

relationship between the simple matrix and the mesh shapes. I am hoping 

to reach a point where the function "makes sense" to the viewer.

justin skrev:
> Quote: Wetterberg wrote on Wed, 15 August 2007 20:20
> ----------------------------------------------------
>
>> I am looking to get the RGB values to modulate the coords of the mesh. 
>> With some fiddling I have managed to get Y and Z movement, but not the 
>> ability to reach arbitrary positions, like when you send it jit.noise, 
>> for instance. I am very much flying blind here, but haven't got the info 
>> on how to correctly translate between 3 char 8 8 and my meshes.
>>
> ----------------------------------------------------
>
> sorry if parts of this are obvious to you, but i will try to break down the task:
>
> part of the problem is the dim of the mesh compared to the dim of the gradient (8x8) - the smaller the dim the less complex the shape.
>
Yes, but I feel that's part of the beauty of it. But even if it was just 
a 3x3 matrix I should still be able to see offsets in all three axes... 
which I still don't.
> also part of the problem is that gradients may be too uniform a colour to create noticeably different geometry.
>
Yeah, you're right. I just sat down and went crazy with the colour 
selectors, but still nothing...
> i would suggest you try using the gradient map as a way to modulate a noise source (jit.noise / jit.bfg) which in turn affects the geometry - perhaps this will give you more varied geometry.
>
Good idea, that... my only concern is that I would like to see a close 
relationship between the simple matrix and the mesh shapes. I am hoping 
to reach a point where the function "makes sense" to the viewer.
> hth,
>
"t" certainly does "h",
Andreas.


Just a quick note to say that I have updated my flickr account to hold 

several very different hi-res screenshots from this patch in its newest 

http://www.flickr.com/photos/andreaswetterberg/sets/72157601487150672/

Feel free to have a look-see. CreativeCommons stuff, so grab what ya like.

- OH, and there's also a game in it - find the pic with the impaled 

Just a quick note to say that I have updated my flickr account to hold 
several very different hi-res screenshots from this patch in its newest 
state.
http://www.flickr.com/photos/andreaswetterberg/sets/72157601487150672/

Feel free to have a look-see. CreativeCommons stuff, so grab what ya like.
- OH, and there's also a game in it - find the pic with the impaled 
torus :-)

nice pictures, they are really good quality.

Could you tell me how you get such a hi-resolution?

Main thing to get the high-quality image would be " @fsaa 1" in 

jit.window. I then fullscreen to one of my 1680*1050 screens and grab a 

Also, I've been asked how smooth the movement is... I've tried to 

capture that movement using the above software. Have a quick look here:

http://covops.dreamhosters.com/dropbox/Random%20Fez%20Grab%201.wmv

- horrible codec, but you get the idea. It does 50-60 fps @ 3360*1050 res.

> nice pictures, they are really good quality.

> Could you tell me how you get such a hi-resolution?

Main thing to get the high-quality image would be " @fsaa 1" in 
jit.window. I then fullscreen to one of my 1680*1050 screens and grab a 
.png screenshot using FRAPS;
http://www.fraps.com/ - pc only.

Also, I've been asked how smooth the movement is... I've tried to 
capture that movement using the above software. Have a quick look here:
http://covops.dreamhosters.com/dropbox/Random%20Fez%20Grab%201.wmv
- horrible codec, but you get the idea. It does 50-60 fps @ 3360*1050 res.

Chris Hipgrave skrev:
> nice pictures, they are really good quality.
>
> Could you tell me how you get such a hi-resolution?
>
> Thanks
> Chris
>
> .
>
>


Quote: Wetterberg wrote on Fri, 17 August 2007 15:21

> Main thing to get the high-quality image would be " @fsaa 1" in 

> jit.window. I then fullscreen to one of my 1680*1050 screens and grab a 

for osx users, press shift+apple+3 for the whole screen, or shift+apple+4 and you can select a portion of the screen. or there is the grab application in the utilities folder.

Quote: Wetterberg wrote on Fri, 17 August 2007 15:21
----------------------------------------------------
> Hi Chris,
>
> Main thing to get the high-quality image would be " @fsaa 1" in 
> jit.window. I then fullscreen to one of my 1680*1050 screens and grab a 
> .png screenshot using FRAPS;
> http://www.fraps.com/ - pc only.
----------------------------------------------------

> Quote: Wetterberg wrote on Fri, 17 August 2007 15:21

>> Main thing to get the high-quality image would be " @fsaa 1" in 

>> jit.window. I then fullscreen to one of my 1680*1050 screens and grab a 

> for osx users, press shift+apple+3 for the whole screen, or shift+apple+4 and you can select a portion of the screen. or there is the grab application in the utilities folder.

Sure, but that doesn't auto-name the file and puts it in the folder you 

specify in the file format you specify, nor does it record video, etc. 

justin skrev:
> Quote: Wetterberg wrote on Fri, 17 August 2007 15:21
> ----------------------------------------------------
>
>> Hi Chris,
>>
>> Main thing to get the high-quality image would be " @fsaa 1" in 
>> jit.window. I then fullscreen to one of my 1680*1050 screens and grab a 
>> .png screenshot using FRAPS;
>> http://www.fraps.com/ - pc only.
>>
> ----------------------------------------------------
>
> for osx users, press shift+apple+3 for the whole screen, or shift+apple+4 and you can select a portion of the screen. or there is the grab application in the utilities folder.
>
>
Sure, but that doesn't auto-name the file and puts it in the folder you 
specify in the file format you specify, nor does it record video, etc. 
etc. double-wink. ;-)

for those that don't want to use third party software or patch up something in max I have included a set-up for grabbing the rendered window image for both screenshots and recorded movies in the example files that come with p.jit.gl.tools. It uses the gl.sketch 'glreadpixels' render-to-matrix and 'exportimage' from a matrix methods.

you still have to choose a folder and file name yourself though...

for those that don't want to use third party software or patch up something in max I have included a set-up for grabbing the rendered window image for both screenshots and recorded movies in the example files that come with p.jit.gl.tools.  It uses the gl.sketch 'glreadpixels' render-to-matrix and 'exportimage' from a matrix methods. 

THANK YOU!