Andres Quezada

Hi! I'm trying to solve this but I'm very newbie at Jitter so I think I should post it. I'm creating a Lorenz attractor (chaotic attractors, a very interesting topic) and I succeed with help from various YouTube videos. The thing is that I want to draw the curved trajectory that the points make over time. There are 36 points in this patch and I need them to draw their trajectory in the jit.world window. Is that even possible? Is there an object that creates lines/curves over time with XY coordinates? My goal is to have the 36 points in the jit.world window and having them draw the trajectories over time infinetely.

slider

flonum

number

comment

jit.gl.camera

loadmess

jit.*

jit.gen

param

swiz

toggle

message

jit.gl.mesh

jit.matrix

button

jit.gl.handle

jit.fpsgui

jit.world

jit.noise

panel

Hi! I'm trying to solve this but I'm very newbie at Jitter so I think I should post it. I'm creating a Lorenz attractor (chaotic attractors,

lorenz-attactor

 http://www.mikezed.com/portfolio/lorenz/

Here's my solution: http://www.mikezed.com/portfolio/lorenz/ 

jit.lcd or mgraphics lets you easily create dots, lines, points, whatever you wish.

however the looks of a "tail" of "moving" points will then still depend on the rate of the datastream.

with GL you might like that there is already a 3D coordinate system present for free and you dont have to calculate the third axis yourself. 

myself i am using the same oldschool method as mzed - i only display one or two axis per lcd. (one could try to use color for the third axis but i think that look silly.)


jit.lcd or mgraphics lets you easily create dots, lines, points, whatever you wish.

however the looks of a "tail" of "moving" points will then still depend on the rate of the datastream.

with GL you might like that there is already a 3D coordinate system present for free and you dont have to calculate the third axis yourself. 

myself i am using the same oldschool method as mzed - i only display one or two axis per lcd. (one could try to use color for the third axis but i think that look silly.)


There's a jit.gl.sketch example in my help patch, too. It's 3D and everything.

Hi, thank you very much for the possible solutions. The "lorenz" object is amazing! It solves many things I wanted to see this week, thank you! Last night I implemented jit.mgraphics and I succeeded with the attractor drawing. I tried for a couple of hours a way of implementing the jit.window points into the jit.mgraphics but I don't know if that is even possible. I want a "guide" for the jit.mgraphics in order to know "where" the attractor is at every single moment. I'm using this jit.mgraphics for 4 performers in order to interpret the attractor for a live improvisation scenario. It would be amazing if I can generate the drawing AND see where the trajectory is with a red point or something like that. Is that possible in Jitter? I'm loving this stuff, I never had the opportunity to learn about Jitter. Thanks in advance!

if you for example simply clear the gl space every 2 seconds, you have a short tail already.

depending on what you are going to control with it you might also like to have a display which shows each dimension separate.


if you for example simply clear the gl space every 2 seconds, you have a short tail already.

you could also do it every 25 new points

depending on what you are going to control with it you might also like to have a display which shows each dimension separate.


Just an example of drawing with lines, from the last point to the current, instead of just drawing points, and accumulating the result.

How do you do this? You mean to reset the jit.mgraphics?

Actually I want to graphic points as I need to see where I am on the X Y axis.

Is not possible to add a red dot that follows the trajectory?

if you for example simply clear the gl space every 2 seconds, you have a short tail already.
you could also do it every 25 new points

a dot is easy, because you are directly controlling its center. therefore it never will be perfectly round and it is also hard to control its exact size. . :)

you need to calculate the offset yourself. for example when the dot has a diameter of 5 pixels you would like to substract 2 from the parameter values for X and Y before you create the drawing command from those values.

lorenz attractor Y: "17" --> [- 2] --> "15"

lorenz attractor X: "29" --> [- 2] --> "27"

--> "ellipse $i1 $i2 5 5, fill" --> [jit.mgraphics]

there are other ways... but i would create a parallel instance (to what you have in your above patch) of the mgraphics object, draw the red dot on black background there, and then mix the 2 video frames with a certain copy mode to ge the red dot "above" the attractor history figure.

however for music applications 1D is also great and you can do that with a scolling multislider. ;)

grafik.png


in gl a dot is easy, because you  are directly controlling its center. therefore it never will be perfectly round and it is also hard to control its exact size. . :)

in mgraphics or lcd you need to calculate the offset yourself.  for example when the dot has a diameter of 5 pixels you would like to substract 2 from the parameter values for X and Y before you create the drawing command from those values.

lorenz attractor Y: "17" --> [- 2] --> "15"
lorenz attractor X: "29" --> [- 2] --> "27"

--> "ellipse $i1 $i2 5 5, fill" --> [jit.mgraphics]

there are other ways... but i would create a parallel instance (to what you have in your above patch) of the mgraphics object, draw the red dot on black background there, and then mix the 2 video frames with a certain copy mode to ge the red dot "above" the attractor history figure.

however for music applications 1D is also great and you can do that with a scolling multislider. ;)



That makes total sense! I recreated the patch and it works great, thank you :) How do you mix/blend two instances of jit.mgraphics? I tried jit.glue but it copies the mgraphics instances side by side. I also tried jit.alphablend but I can't get it to work right.

CleanShot 2021-08-25 at 10.39.40.png

EDIT: I ended using jit.op @op * and it works! Is there any other method that is better than this?

CleanShot 2021-08-25 at 11.00.55.png

I will definitely make use of the multislider for the musicians <3

yes jit.op has the basics and is effective.

you might run into situations with it where it only works when the background of A is black or white - but where this is intended arithmetics are perfect. :)

i am an lcd guy and have no experience with mgraphics, it can be that i got the offset wrong. to check if the dot is exactly centered, just change the order for a moment so that the black pixel is on top.


yes jit.op has the basics and is effective.

you might run into situations with it where it only works when the background of A is black or white - but where this is intended arithmetics are perfect. :)

i am an lcd guy and have no experience with mgraphics, it can be that i got the offset wrong. to check if the dot is exactly centered, just change the order for a moment so that the black pixel is on top.


Yes, I subtracted -5 from the coordinate and it works :) Do you use jit.lcd or only lcd? Is there any major difference? I tend to think that lcd offers more options than mgraphics, I will read about it these days. Thank you very much for everything, I'm very happy with the results at the moment.

For reference purposes, here's an OpenGL version with the red dots.

they are almost identical. investing in one of them prepares you for the other. 

the major difference is that the jit buddy requires a [t b l] at its inlet to output stuff.

i usually start with lcd and then later use jit.lcd when i run into issues regarding idle mouse order or when i need 3 layers and copymodes. copymodes exist 

lcd, but in jit.lcd you can have both. :)

it can also be fun/useful to apply simple video effects on a jit.lcd, for example "greyscale" or "invert" - not to speak of geometric mirroring or scaling to 50% - will be much more work using lcd commands, while as matrix operation after a jit.lcd it is a nobrainer.


they are almost identical.  investing in one of them prepares you for the other. 

the major difference is that the jit buddy requires a [t b l] at its inlet to output stuff.

i usually start with lcd and then later use jit.lcd when i run into issues regarding idle mouse order or when i need 3 layers and copymodes. copymodes exist within lcd, but in jit.lcd you can have both. :)

it can also be fun/useful to apply simple video effects on a jit.lcd, for example "greyscale" or "invert" - not to speak of geometric mirroring or scaling to 50% - will be much more work using lcd commands, while as matrix operation after a jit.lcd it is a nobrainer.




Thank you Pedro, that patch was exactly what I was trying to achieve in gl.

I will try some stuff with lcd, I'm guessing it will bring other possibilities within this project <3

Hi Pedro, I was checking your patch and it gives me a lot of ideas for this project. I was wondering if there is a way to isolate the points of the attractor. In the patch below I'm generating 50 points in jit.window but it would be great if I can isolate one of the points in another jit.pwindow throughout the patch. Is there a way to isolate the points? I'm generating the coordinates with jit.spill and below I'm getting one of the points with jit.mgraphics but I love the feedback that you created in the jit.window. It would be amazing if I can isolate the 50 points one by one in different jit.pwindows but I don't know how to resolve this. Thank you very much in advance, I'm learning a lot doing this project :)

Here's an alternate approach that consists of generating the geometry instead of relying on pixel post-processing effects (accumulation). Here I accumulate the last 500 geometry points and draw the lines with it.

To draw a specific instance in a different window, just create another jit.world with a different name, and restrict the matrix that goes into another jit.gl.mesh to contain only the geometry coordinates of that line.

Here's an example, I hope it makes sense to you.

Hi! Sorry for the delay.
Here's an alternate approach that consists of generating the geometry instead of relying on pixel post-processing effects (accumulation). Here I accumulate the last 500 geometry points and draw the lines with it.
To draw a specific instance in a different window, just create another jit.world with a different name, and restrict the matrix that goes into another jit.gl.mesh to contain only the geometry coordinates of that line.
Here's an example, I hope it makes sense to you.


Hi Pedro, sorry for the late response too. This is great!!! It's exactly what I was looking for. I'm trying to understand the patch as there are a lot of objects I didn't knew about but I'm learning in the process. I'm trying a way to send the different jit.world windows to different monitors. Maybe a solution would be to use jit.displays? Here I'm drawing three instances.

Thanks! Looking forward to learning more about jitter.

Hi, Andres. You can drag each window to a particular screen and press the Escape key, making it fullscreen.

Alternatively, you can write the positioning of the window and the fullscreen activation directly in each jit.world object. Here's an example assuming 2 FullHD (1920*1080) monitors side by side and positioning the window in the right/secondary screen.

jit.world lorenz @windowposition 1920 0 @fullscreen 1

The 1920 pixels of the first screen start counting from 0 so 1919 is the rightmost pixel. So, the coordinate 1920 is the 1st pixel of the second screen and so on. You don't need to position the window on the left/upper corner anyway, you just need to position the window inside the screen (for instance, "2000 100" would work), because the window will be made fullscreen

Hi, Andres. You can drag each window to a particular screen and press the Escape key, making it fullscreen.
Alternatively, you can write the positioning of the window and the fullscreen activation directly in each jit.world object. Here's an example assuming 2 FullHD (1920*1080) monitors side by side and positioning the window in the right/secondary screen.

The 1920 pixels of the first screen start counting from 0 so 1919 is the rightmost pixel. So, the coordinate 1920 is the 1st pixel of the second screen and so on. You don't need to position the window on the left/upper corner anyway, you just need to position the window inside the screen (for instance, "2000 100" would work), because the window will be made fullscreen

I hope this explanation helps.

method #3 involved [screensize] and window messages to [thispatcher]

Thank you Pedro and Roman, I will try this on Tuesday as I will be in a room with 3 monitors.

Lorenz attactor