Author Topic: A proposed sequence builder for rgb megatrees. meighan.net/seqbuilder/index.php  (Read 14530 times)

Offline smeighan

  • Patron Member
  • Sr. Member
  • ****
  • Posts: 2285
    • Nutcracker RGB Sequence Builder
Re: A proposed sequence builder for rgb megatrees
« Reply #15 on: January 16, 2012, »
i have rewritten the web page to show animated jpegs instead of animated gif's

I have a java script that loads the jpegs and animates them. This also seemed to solve my issue with sometime not seeing the animations. IF's can only display 256 colors. I wanted the animations to reflect the actual RGB color that will be seen in your show.

It has only been a few days so this is still Alpha software.

You are not allowed to view links. Register or Login

I am going to change the math model to rotate 3 spirals at once and that. If this works as expected, i hope to get more people contributing the models of different animations.

Sean
Littleton, CO
Latest releases You are not allowed to view links. Register or Login
xLights/Nutcracker Forum You are not allowed to view links. Register or Login
Fbook You are not allowed to view links. Register or Login

Offline smeighan

  • Patron Member
  • Sr. Member
  • ****
  • Posts: 2285
    • Nutcracker RGB Sequence Builder
Re: A proposed sequence builder for rgb megatrees
« Reply #16 on: January 16, 2012, »
I am posting to two places, one here and the other on the lsp forums.

Here is a thread from there:

The targets can be any shape. Currently i have modeled megatrees and a flat plane. I will also make a snowman of three spheres just to show how it would look. a 10' tall snowman, hmmm.

anyways the math models are independent of the targets. in fact you could have many target objects, say 2 mega-trees next to each other, and the projection from the math model could be projeted onto both at the same time.

Just need the math model and the target models to be close to the same size. Currently i am using a 10' high cone with a 5' diameter. I then have modeled the physical megatree 10. tall and 5' diametere. i move the model on top of the target and do a match to the closest pixel.

I find the closest pixel by doing
dist = sqrt((xm-xt)**2 + (ym-yt)**2 ++ (zm-zt)**2).
In any direction, dist is now how far apart the pixel is from the source model.
xm,ym,zm = coordnates from the math model
xt,yt,zt = coordinates of the models.

I am thinking of reaching out to jr colleges or high school math departments and see if their students might want to think up some of the math models.

I have an idea of how to do snowflakes, garlands. I want to get as many different models as we can.

The source model could come from something other than math models.

For example: maybe use a tablet and store  the data as people write. You would then see whatever was drawn projected onto the target models.

other things maybe would be to read avi files and project onto tthe targets

right now, i just wanted to get the math models. For example the single spiral you see has a single parameter to it to add more spirals. 3, 6, 12? Model is built projection is the same. It is these more complicated models where i think it is a valuable tool.



Quote from: Fasteddy;13836
So this is basically an effects generator which is an exciting prospect for LSP, I hope this feature will be able to be used on more than just a mega tree. I would like to see the same features as with transitions where you can map those generated effects to either an element or the whole display.
This then would really make it a madrix competitor and take LSP to the next level

Thanks for the great work and input, im  sure the community will be much better off with your input.
« Last Edit: January 19, 2012, by smeighan »
Sean
Littleton, CO
Latest releases You are not allowed to view links. Register or Login
xLights/Nutcracker Forum You are not allowed to view links. Register or Login
Fbook You are not allowed to view links. Register or Login

Offline RJ

  • Administrator
  • Sr. Member
  • *****
  • Posts: 8519
Re: A proposed sequence builder for rgb megatrees
« Reply #17 on: January 16, 2012, »
Looks like some neat stuff going on here.

RJ
Innovation beats imitation - and it's more satisfying

Offline smeighan

  • Patron Member
  • Sr. Member
  • ****
  • Posts: 2285
    • Nutcracker RGB Sequence Builder
Re: A proposed sequence builder for rgb megatrees
« Reply #18 on: January 16, 2012, »
I have added one more model to web site.
The math model for the spiral has a parameter for setting the number of spirals. I created a 1,2 and 3 spiral source model.
3 spirals would probably be very complicated if you had to program the sequence by hand.

You are not allowed to view links. Register or Login

This starts to show where i think this will be valuable. I dont have the mapping to the target models done yet.
« Last Edit: January 19, 2012, by smeighan »
Sean
Littleton, CO
Latest releases You are not allowed to view links. Register or Login
xLights/Nutcracker Forum You are not allowed to view links. Register or Login
Fbook You are not allowed to view links. Register or Login

Offline devoidelk

  • Sr. Member
  • ****
  • Posts: 208
  • North western Indiana, 47960
    • Facebook
Re: A proposed sequence builder for rgb megatrees
« Reply #19 on: January 17, 2012, »
Looks good so far i dont have any rgbs so when I get some to make a mega tree your programs the one ill be looking up...

Offline smeighan

  • Patron Member
  • Sr. Member
  • ****
  • Posts: 2285
    • Nutcracker RGB Sequence Builder
Re: A proposed sequence builder for rgb megatrees
« Reply #20 on: January 17, 2012, »
I have a new web link for the sequence builder

You are not allowed to view links. Register or Login

This code now is actually building the models.  Please update any bookmarks you may have

thanks
sean
Sean
Littleton, CO
Latest releases You are not allowed to view links. Register or Login
xLights/Nutcracker Forum You are not allowed to view links. Register or Login
Fbook You are not allowed to view links. Register or Login

Offline kgustafson

  • Coop Manager
  • Sr. Member
  • *
  • Posts: 1120
    • Lost Weekend Productions
I can see my Linear Algebra classes in college might actually be put to use.  LOL.
------
Visit at: You are not allowed to view links. Register or Login

Offline Warrant

  • Jr. Member
  • **
  • Posts: 7
  • Ho! Ho! How does that work again?
The awesomeness of this project is...awesome!

One idea comes to mind, I've seen some folks make what I'll call, "half-tree's".  Mega-trees but only the front 180 degrees of it.  Those types would seem to be more used for animations, figures, shapes, pictures, etc.  A full 360 degree tree would have "light pollution" from the back row of lights.  Do-able?

Offline Rod R

  • Coop Manager
  • Sr. Member
  • *
  • Posts: 902
Nice work Sean,  Since I am an LSP user as well I see this being a very handy tool for us and other software programs.

I can see the math model animations in all three examples but only the target for the 8/16/32.  In the first example (green pixels).

The 2nd and 3rd examples do not show any animations on the targets. Using Win XP, Firefox.
« Last Edit: January 18, 2012, by Rod R »

Offline smeighan

  • Patron Member
  • Sr. Member
  • ****
  • Posts: 2285
    • Nutcracker RGB Sequence Builder
I have released a new version of the seq builder. There were some major changes in the matching algorithym.
You are not allowed to view links. Register or Login

I also wrote a description of how the logic works.

Now i am modeling the targets for all three source math models. I also changed the targets so the data looks more like LED's.


One issue i am working on is how to turn off animations and turn them on demand.
Currently when the web page runs and creates the png files, it is making 2600 of them, 11 megabytes. It is caching all of these in the browser and then turning them on using a java script.

Each png file is 1-2kbytes.

still a work in progress. maybe next week ill work on exporting the data to vixen or lsp. then maybe someone can try these and see what they look like.

With this release i think you can get an idea what your megatrees would look like.

thanks
sean

BTW, there is another thread here for this software: You are not allowed to view links. Register or Login
Sean
Littleton, CO
Latest releases You are not allowed to view links. Register or Login
xLights/Nutcracker Forum You are not allowed to view links. Register or Login
Fbook You are not allowed to view links. Register or Login

Offline mkozik1

  • Sr. Member
  • ****
  • Posts: 326
  • Monroe, GA 30655
I too am using Firefox but can see the third sequence but not the second.  The third took a little while to get going but once everything finished downloading it moves pretty quickly.  The second only shows "fast animation example".
- Mark

Offline Steve Gase

  • Sr. Member
  • ****
  • Posts: 2915
    • WinterLightShow in Georgetown, TX
I like your progress!  You are doing well with the idea to widen the sprirals...  I think they do require more thickness to come off well.
You are not allowed to view links. Register or Login  |  110K channels, 50K lights  |  Nutcracker, Falcon, DLA, HolidayCoro

Offline smeighan

  • Patron Member
  • Sr. Member
  • ****
  • Posts: 2285
    • Nutcracker RGB Sequence Builder
You are not allowed to view links. Register or Login
The awesomeness of this project is...awesome!

One idea comes to mind, I've seen some folks make what I'll call, "half-tree's".  Mega-trees but only the front 180 degrees of it.  Those types would seem to be more used for animations, figures, shapes, pictures, etc.  A full 360 degree tree would have "light pollution" from the back row of lights.  Do-able?

Yep, this is very easy. I just will build a target tree with half the segments. Currently i have a variable match algorithm. If i am in the first 10% of the top of the tree, i match when the pixels are within 1" of the model. at the very top a 32 segment tree is only a few tenths of an inch apart. As i go down the tree the radius increases and the strings separate. I am using 4" for the bottom of a 10' tree. If i tighten this matching routine to a few inches then the spirals to animate multiple pixels. In other words i will set up a set of matching routines and allow the user to pick how fussy the match will be.

i will add half trees later this week and see what it looks like.

thanks
sean
Sean
Littleton, CO
Latest releases You are not allowed to view links. Register or Login
xLights/Nutcracker Forum You are not allowed to view links. Register or Login
Fbook You are not allowed to view links. Register or Login

Offline Ozmin

  • Patron Member
  • Sr. Member
  • ****
  • Posts: 136
WOW!!   This will be very helpful in the future.  Since channel counts are going up exponentially with the new RGB stuff and this will cut the sequencing time down a bunch.  Thank you for all the hard work!!

Bill A.

Offline Rod R

  • Coop Manager
  • Sr. Member
  • *
  • Posts: 902
I can see everything now WinXP, firefox.   Nice job in the short time you've been working on it.