keithmanlove

I have avoided jit.expr like the plague, which I know is a bad idea.

I've done things with it, and I've read the tutorials about it. It

still evades me. I feel like I get it, but then I look at Andrew B's

shatter example and I'm confused. I understand what purpose it has in

the patch, and I get the form. Would someone mind stepping me through

this equation? I don't mind if you treat me like a moron... I

probably deserve it. Any help would be greatly appreciated.

jit.expr @expr "(cell[0]%4.>0.)*(cell[0]%4.

"(cell[0]%4>1)*0.052" "0." "(cell[0]%4.>0.)*(cell[0]%4.

Maybe this will help me over whatever block I've been having.

I have avoided jit.expr like the plague, which I know is a bad idea.
I've done things with it, and I've read the tutorials about it.  It
still evades me.  I feel like I get it, but then I look at Andrew B's
shatter example and I'm confused.  I understand what purpose it has in
the patch, and I get the form.  Would someone mind stepping me through
this equation?  I don't mind if you treat me like a moron... I
probably deserve it.  Any help would be greatly appreciated.

jit.expr @expr "(cell[0]%4.>0.)*(cell[0]%4.
"(cell[0]%4>1)*0.052" "0." "(cell[0]%4.>0.)*(cell[0]%4.
"(cell[0]%4>1)*0.025"

jit-expr-and-shatter

This looks to be for quad generation, thus the % 4. The % 4 basically

says count from 0 to 3 in loops producing a patter 0, 1, 2, 3, 0, 1,

for cells with loop pattern with values 1 or 2, put the value 0.052 in the cell

for cells with loop pattern 2 and 3 put value 0.052

this results in quads with vertices [0, 0, 0] [0.052 , 0, 0], [0.052,

for cells with loop pattern with values 1 or 2, put the value 0.025 in the cell

for cells with loop pattern 2 and 3 put value 0.025

This results in texcure coordinates [0, 0] [0.025 , 0], [0.025,

This looks to be for quad generation, thus the % 4.  The % 4 basically
says count from 0 to 3 in loops producing a patter 0, 1, 2, 3, 0, 1,
2, 3, ...

Vertices
x:
"(cell[0]%4.>0.)*(cell[0]%4.
for cells with loop pattern with values 1 or 2, put the value 0.052 in the cell

y:
"(cell[0]%4>1)*0.052"
for cells with loop pattern 2 and 3 put value 0.052

this results in quads with vertices [0, 0, 0] [0.052 , 0, 0], [0.052,
0.052, 0], [0, 0.052, 0]

TexCoords:
s:
"(cell[0]%4.>0.)*(cell[0]%4.
for cells with loop pattern with values 1 or 2, put the value 0.025 in the cell

t:
"(cell[0]%4>1)*0.025"
for cells with loop pattern 2 and 3 put value 0.025

This results in texcure coordinates  [0, 0] [0.025 , 0], [0.025,
0.025], [0, 0.025]

> This looks to be for quad generation, thus the % 4. The % 4 basically

> says count from 0 to 3 in loops producing a patter 0, 1, 2, 3, 0, 1,

> for cells with loop pattern with values 1 or 2, put the value 0.052 in the cell

> for cells with loop pattern 2 and 3 put value 0.052

> this results in quads with vertices [0, 0, 0] [0.052 , 0, 0], [0.052,

> for cells with loop pattern with values 1 or 2, put the value 0.025 in the cell

> for cells with loop pattern 2 and 3 put value 0.025

> This results in texcure coordinates [0, 0] [0.025 , 0], [0.025,

On 8/16/07, Wesley Smith  wrote:
> This looks to be for quad generation, thus the % 4.  The % 4 basically
> says count from 0 to 3 in loops producing a patter 0, 1, 2, 3, 0, 1,
> 2, 3, ...
>
> Vertices
> x:
> "(cell[0]%4.>0.)*(cell[0]%4.
> for cells with loop pattern with values 1 or 2, put the value 0.052 in the cell
>
> y:
> "(cell[0]%4>1)*0.052"
> for cells with loop pattern 2 and 3 put value 0.052
>
> z:
> "0."
> always 0
>
> this results in quads with vertices [0, 0, 0] [0.052 , 0, 0], [0.052,
> 0.052, 0], [0, 0.052, 0]
>
>
> TexCoords:
> s:
> "(cell[0]%4.>0.)*(cell[0]%4.
> for cells with loop pattern with values 1 or 2, put the value 0.025 in the cell
>
> t:
> "(cell[0]%4>1)*0.025"
> for cells with loop pattern 2 and 3 put value 0.025
>
> This results in texcure coordinates  [0, 0] [0.025 , 0], [0.025,
> 0.025], [0, 0.025]
>
> wes
>


Thanks Wes for explaining this a little better. Here is a simplified 

version of the patch that might make things a little clearer.

Thanks Wes for explaining this a little better.  Here is a simplified 
version of the patch that might make things a little clearer.

Oh, Andrew... Wes... this is soooo great. I'm doing stuff that's been

stumping me for a while. I'll post something soon. This is a very

> Thanks Wes for explaining this a little better. Here is a simplified

> version of the patch that might make things a little clearer.

> #P comment 15 379 381 196617 The jit.expr expressions are designed to

> generate a simple billboard quad complete with texture coordinates for

> every 4 points in a matrix (horizontally). The resulting matrix can be

> added to another matrix that gives position offsets. The position matrix

> should have 4 redundant value for each quad so that all of the vertices

> #P user jit.cellblock 267 325 628 345 3 9 8 1 45 17 0 1 1 0 0 0 0 0 1 0

> 0 0 255 255 255 0 0 0 0 0 0 191 191 191 0 0 0 215 215 240 1 1 1 0 4 0 0 0;

> #P user jit.cellblock 267 304 628 324 3 9 8 1 45 17 0 1 1 0 0 0 0 0 1 0

> #P user jit.cellblock 267 283 628 303 3 9 8 1 45 17 0 1 1 0 0 0 0 0 1 0

> #P user jit.cellblock 267 261 628 281 3 9 8 1 45 17 0 1 1 0 0 0 0 0 1 0

> #P newex 267 218 79 196617 jit.unpack 5;

> #P user jit.cellblock 267 241 628 261 3 9 8 1 45 17 0 1 1 0 0 0 0 0 1 0

> #P newex 361 154 75 196617 jit.window foo;

> #P newex 361 135 82 196617 jit.gl.render foo;

> #P newex 17 241 248 196617 jit.gl.mesh foo @draw_mode quads @color 1. 1.

> #P newex 17 110 109 196617 jit.matrix 5 float32 8;

> #P newex 17 140 299 196617 jit.expr @expr

> #P comment 127 110 226 196617 simple matrix with 8 vertex points ,

> which will generate 2 quads (1 for every 4 vertices);

> #P comment 15 352 381 196617 drawmode quads expects a matrix with

> horizontal dimension that is a multiple of 4. The quads are drawn in

> order top-left , top-right , bottom-right , bottom-left.;

Oh, Andrew... Wes... this is soooo great.  I'm doing stuff that's been
stumping me for a while.  I'll post something soon.  This is a very
great object.  Thanks so much.

On 8/17/07, andrew benson  wrote:
> Thanks Wes for explaining this a little better.  Here is a simplified
> version of the patch that might make things a little clearer.
>
> Best,
> Andrew B.
>
> #P window setfont "Sans Serif" 9.;
> #P window linecount 5;
> #P comment 15 379 381 196617 The jit.expr expressions are designed to
> generate a simple billboard quad complete with texture coordinates for
> every 4 points in a matrix (horizontally). The resulting matrix can be
> added to another matrix that gives position offsets. The position matrix
> should have 4 redundant value for each quad so that all of the vertices
> will be added to the same value.;
> #P user jit.cellblock 267 325 628 345 3 9 8 1 45 17 0 1 1 0 0 0 0 0 1 0
> 0 0 255 255 255 0 0 0 0 0 0 191 191 191 0 0 0 215 215 240 1 1 1 0 4 0 0 0;
> #P user jit.cellblock 267 304 628 324 3 9 8 1 45 17 0 1 1 0 0 0 0 0 1 0
> 0 0 255 255 255 0 0 0 0 0 0 191 191 191 0 0 0 215 215 240 1 1 1 0 4 0 0 0;
> #P user jit.cellblock 267 283 628 303 3 9 8 1 45 17 0 1 1 0 0 0 0 0 1 0
> 0 0 255 255 255 0 0 0 0 0 0 191 191 191 0 0 0 215 215 240 1 1 1 0 4 0 0 0;
> #P user jit.cellblock 267 261 628 281 3 9 8 1 45 17 0 1 1 0 0 0 0 0 1 0
> 0 0 255 255 255 0 0 0 0 0 0 191 191 191 0 0 0 215 215 240 1 1 1 0 4 0 0 0;
> #P window linecount 1;
> #P newex 267 218 79 196617 jit.unpack 5;
> #P user jit.cellblock 267 241 628 261 3 9 8 1 45 17 0 1 1 0 0 0 0 0 1 0
> 0 0 255 255 255 0 0 0 0 0 0 191 191 191 0 0 0 215 215 240 1 1 1 0 4 0 0 0;
> #P newex 361 154 75 196617 jit.window foo;
> #P toggle 361 69 15 0;
> #P newex 361 88 51 196617 qmetro 5;
> #P newex 361 135 82 196617 jit.gl.render foo;
> #P newex 361 107 50 196617 t b erase;
> #P button 17 75 15 0;
> #P newex 17 241 248 196617 jit.gl.mesh foo @draw_mode quads @color 1. 1.
> 1. 1.;
> #P newex 17 110 109 196617 jit.matrix 5 float32 8;
> #P window linecount 2;
> #P newex 17 140 299 196617 jit.expr @expr
> "(cell[0]%4.>0.)*(cell[0]%4.1)" "0."
> "(cell[0]%4.>0.)*(cell[0]%4.1)";
> #P comment 127 110 226 196617 simple matrix with 8 vertex points ,
> which will generate 2 quads (1 for every 4 vertices);
> #P comment 15 352 381 196617 drawmode quads expects a matrix with
> horizontal dimension that is a multiple of 4. The quads are drawn in
> order top-left , top-right , bottom-right , bottom-left.;
> #P connect 2 0 4 0;
> #P fasten 2 0 12 0 22 175 272 175;
> #P connect 3 0 2 0;
> #P connect 12 4 16 0;
> #P connect 12 3 15 0;
> #P connect 12 2 14 0;
> #P connect 12 1 13 0;
> #P connect 12 0 11 0;
> #P connect 9 0 8 0;
> #P connect 8 0 6 0;
> #P fasten 6 1 7 0 406 128 366 128;
> #P connect 6 0 7 0;
> #P connect 5 0 3 0;
> #P window clipboard copycount 18;
>
>
>


> Thanks Wes for explaining this a little better. Here is a simplified 

Oh Andrew, this has so much potential for my work... I really really 

andrew benson skrev:
> Thanks Wes for explaining this a little better.  Here is a simplified 
> version of the patch that might make things a little clearer.
>
> Best,
> Andrew B.
>
> #P window setfont "Sans Serif" 9.; 
Oh Andrew, this has so much potential for my work... I really really 
wish it made sense to me :-/

I think it's the overall syntax of jit.expr - I'll give it a try:

- so basically every dimension is in " " .

- and this is two logical operations multiplied by each other, then multiplied by 0.052 - meaning both TRUE then 0.052, otherwise 0?

I guess upon closer investigation it does *read* alright, I just 

wouldn't know where to begin when using it on my own.

I remember the jitter tuts having a few lessons on jit.gl.sketch with 

the messages getting increasingly advanced - were there jit.expr lessons 

andrew benson skrev:
> Hi Andreas,
> What isn't making sense?
>
I think it's the overall syntax of jit.expr - I'll give it a try:

 - so basically every dimension is in " " .

"(cell[0]%4.>0.)*(cell[0]%4.
- and this is two logical operations multiplied by each other, then multiplied by 0.052 - meaning both TRUE then 0.052, otherwise 0?

I guess upon closer investigation it does *read* alright, I just 
wouldn't know where to begin when using it on my own.

I remember the jitter tuts having a few lessons on jit.gl.sketch with 
the messages getting increasingly advanced - were there jit.expr lessons 
like this too?

> I think it's the overall syntax of jit.expr - I'll give it a try:

> - so basically every dimension is in " " .

> - and this is two logical operations multiplied by each other, then multiplied by 0.052 - meaning both TRUE then 0.052, otherwise 0?

Exactly. Just remember that jit.expr doesn't impement order of

operations so use parentheses excessively. I approach using jit.expr

by sketching what I want on paper and then writing down the equations

in a general form like you would when doing a math problem and then

seeing how the data maps to a matrix. Once that's done, I adapt the

> I think it's the overall syntax of jit.expr - I'll give it a try:
>
>  - so basically every dimension is in " " .
>
> "(cell[0]%4.>0.)*(cell[0]%4.
> - and this is two logical operations multiplied by each other, then multiplied by 0.052 - meaning both TRUE then 0.052, otherwise 0?

Exactly.  Just remember that jit.expr doesn't impement order of
operations so use parentheses excessively.  I approach using jit.expr
by sketching what I want on paper and then writing down the equations
in a general form like you would when doing a math problem and then
seeing how the data maps to a matrix.  Once that's done, I adapt the
general form to a jit.expr expression.

i have been f=playing with shatter, i (now) understand the concept of

jit.expr, but fail when it comes to the texture coordinates. if i understand

correctly the s and t are NOT per vertex but a scaler. i fail to

understanding the paragraph mentioning them in the tutorial and failed

understanding that concept when digging the red book.

in the example i attach i took the shatter and tinyvideo patch's (Andrew,

both are to be adored), and combined some of their functions.

> > I think it's the overall syntax of jit.expr - I'll give it a try:

> > - so basically every dimension is in " " .

> > - and this is two logical operations multiplied by each other, then

> multiplied by 0.052 - meaning both TRUE then 0.052, otherwise 0?

> Exactly. Just remember that jit.expr doesn't impement order of

> operations so use parentheses excessively. I approach using jit.expr

> by sketching what I want on paper and then writing down the equations

> in a general form like you would when doing a math problem and then

> seeing how the data maps to a matrix. Once that's done, I adapt the

i have been f=playing with shatter, i (now) understand the concept of
jit.expr, but fail when it comes to the texture coordinates. if i understand
correctly the s and t are NOT per vertex but a scaler. i fail  to
understanding the paragraph mentioning them in the tutorial and failed
understanding that concept when digging the red book.

in the example i attach i took the shatter and tinyvideo patch's (Andrew,
both are to be adored), and combined some of their functions.
but the texture dont sit right, fail...

On 8/18/07, Wesley Smith  wrote:
>
> > I think it's the overall syntax of jit.expr - I'll give it a try:
> >
> >  - so basically every dimension is in " " .
> >
> > "(cell[0]%4.>0.)*(cell[0]%4.
> > - and this is two logical operations multiplied by each other, then
> multiplied by 0.052 - meaning both TRUE then 0.052, otherwise 0?
>
> Exactly.  Just remember that jit.expr doesn't impement order of
> operations so use parentheses excessively.  I approach using jit.expr
> by sketching what I want on paper and then writing down the equations
> in a general form like you would when doing a math problem and then
> seeing how the data maps to a matrix.  Once that's done, I adapt the
> general form to a jit.expr expression.
>
> wes
>

success. but thru trail and error, i still don't know why it works....

> > > I think it's the overall syntax of jit.expr - I'll give it a try:

> > > - so basically every dimension is in " " .

> > > - and this is two logical operations multiplied by each other, then

> > multiplied by 0.052 - meaning both TRUE then 0.052, otherwise 0?

> > Exactly. Just remember that jit.expr doesn't impement order of

> > operations so use parentheses excessively. I approach using jit.expr

> > by sketching what I want on paper and then writing down the equations

> > in a general form like you would when doing a math problem and then

> > seeing how the data maps to a matrix. Once that's done, I adapt the

> > general form to a jit.expr expression.

>
>
>
> On 8/18/07, Wesley Smith  wrote:
> >
> > > I think it's the overall syntax of jit.expr - I'll give it a try:
> > >
> > >  - so basically every dimension is in " " .
> > >
> > > "(cell[0]%4.>0.)*(cell[0]%4.
> > > - and this is two logical operations multiplied by each other, then
> > multiplied by 0.052 - meaning both TRUE then 0.052, otherwise 0?
> >
> > Exactly.  Just remember that jit.expr doesn't impement order of
> > operations so use parentheses excessively.  I approach using jit.expr
> > by sketching what I want on paper and then writing down the equations
> > in a general form like you would when doing a math problem and then
> > seeing how the data maps to a matrix.  Once that's done, I adapt the
> > general form to a jit.expr expression.
> >
> > wes
> >
>
>

some more demos, this one draws a morph target using jit.lcd. 

i broke a higher res matrix into 6 different pieces and got to 1600x1200*6 total texture size, but this is using only one shatter "module", peace and patch to come.

http://www.zshare.net/download/3896458e9db738/

the video is h.624 (4mb), hope it plays well for all.

some more demos, this one draws a morph target using jit.lcd. 
i broke a higher res matrix into 6 different pieces and got to 1600x1200*6 total texture size, but this is using only one shatter "module", peace and patch to come.
http://www.zshare.net/download/3896458e9db738/
the video is h.624 (4mb), hope it plays well for all.


Very cool Yair! I really like the use of jit.lcd for controlling motion 

Very cool Yair!  I really like the use of jit.lcd for controlling motion 
and shape.  Thanks for sharing.

yeah, i like that a lot, thanks for sharing. 

Quote: yair r. wrote on Fri, 28 September 2007 14:21

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

> some more demos, this one draws a morph target using jit.lcd. 

> i broke a higher res matrix into 6 different pieces and got to 1600x1200*6 total texture size, but this is using only one shatter "module", peace and patch to come.

> the video is h.624 (4mb), hope it plays well for all.

Quote: yair r. wrote on Fri, 28 September 2007 14:21
----------------------------------------------------
> some more demos, this one draws a morph target using jit.lcd. 
> i broke a higher res matrix into 6 different pieces and got to 1600x1200*6 total texture size, but this is using only one shatter "module", peace and patch to come.
> http://www.zshare.net/download/3896458e9db738/
> the video is h.624 (4mb), hope it plays well for all.
>
----------------------------------------------------


thanks everybody for the kind words. i'll continue to update on what i

accomplish with shatter and the xray objects.

> Quote: yair r. wrote on Fri, 28 September 2007 14:21

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

> > some more demos, this one draws a morph target using jit.lcd.

> > i broke a higher res matrix into 6 different pieces and got to

> 1600x1200*6 total texture size, but this is using only one shatter "module",

> > the video is h.624 (4mb), hope it plays well for all.

thanks everybody for the kind words. i'll continue to update on what i
accomplish with shatter and the xray objects.

On 9/29/07, karl-otto von oertzen  wrote:
>
>
>
> very nice !
>
> Quote: yair r. wrote on Fri, 28 September 2007 14:21
> ----------------------------------------------------
> > some more demos, this one draws a morph target using jit.lcd.
> > i broke a higher res matrix into 6 different pieces and got to
> 1600x1200*6 total texture size, but this is using only one shatter "module",
> peace and patch to come.
> > http://www.zshare.net/download/3896458e9db738/
> > the video is h.624 (4mb), hope it plays well for all.
> >
> ----------------------------------------------------
>
>
> --
> karrrlo
>
>

Could you confirm this comment is false, please.

For quads the vertices position seems to be bottom-left, bottom-right, top-right, top-left .

Could you confirm this comment is false, please.
For quads the vertices position seems to be bottom-left, bottom-right, top-right, top-left .

Could you be a little more specific as to what you are seeing?

I'm drawing 1 quad and I'm having 2 matrix - for vertex & texcoords position.

If I change the value for cell 0,0 then it appears to be the bottom-left point of the quad.

the points location/number seems not to be the same for quads and simple mesh points-made ...?

What I understand is even with flipped texture, vertex #1 is bottom-left [for quads], since values for it appear to be in cell 0,0 in a [jit.matrix]. 

But my logic could be inappropriate since I'm a dummy in openGL.

What I understand is even with flipped texture, vertex #1 is bottom-left [for quads], since values for it appear to be in cell 0,0 in a [jit.matrix]. 
But my logic could be inappropriate since I'm a dummy in openGL.


jit.expr and shatter