RGBSB target model builder

[smeighan]

This thread is to talk about the portion of RGBSB that will model your megatree so RGBSB can put effects on it.

This is VERY important. If you cannot model your tree , you will NOT be able to use RGBSB.

This initial release is everything except for the editing of an existing tree. Since defining your tree should take less than 60 seconds, just build multiple models.

This release now allows you to create an account on rgbsb. One you create an account you will come into the target modeler screen.

This is what i need from you. Is there any reason you cannot model your existing or future tree?

I am aware of some people who have built tree in what i call a snake method. take a string, go up and then down and then go up again. I will be adding this soon.

Your models are being saved to the database, but it is not like you would lose alot of work. Like i said, you should be able to model your tree in a few minutes.

I expect the ability to edit an existing model will occur tonight.

Todays release is alpha, things may work and stop working. I am releasing code. I expect the stable version ot be released on Monday. I am doing the early release to get feedback.

So have fun You are not allowed to view links. Register or Login

thanks

[smeighan]

Many of you have created a model, Great job!

I have been asked, "What Now?"


What happens now is I verify that everyone can model their tree. I will choose some of the models entered to be come part of my test suite. This was the goal of this weekend.


the next part is you need to create effects in your user space. Think of the effects you have seen, say the barber pole, the most simple spiral model i have.
That effect has these variables.

Number of stripes. The Barber pole had this set to 2
How many rotations a stripe makes going from the top to the bottom. The barber pole on my web page had this set to 3
Color of each stripe. RED and WHITE were used for Barber Pole.
Thickness of stripe, in Barber Pole case you say to fill it in till the next stripe
Rotation cw or ccw. I think i had it cw
Do you want the whole tree used or only half a mega tree. I set this to be half a megatree


Now you will save this as say BARBER1 in your effects library.

Then you go to your personal library and select a target (Example: MTREE1)and a effect, (Example: BARBER1)
I will ask you for the time quantum, say 50ms,
and how long you want a sequence, say 5 seconds.

I now create one frame with every string#, pixel#, rgb value. In this case I will produce 100 objects

At this point I can make animated gif's. Notice so far the software is sequencer agnostic.
This point is where i have been but i was cheating by setting the variables inside the software and not pulling them from tables.

At this point I expect you will look at the animations and maybe go back and modify your effects and try again.
If you are only going to have one megatree, you will never go back to the target modeler. I do expect you will start to makde 10's or 100's of effects that you personally like.

When satisfied, you select a sequencer and I produce the Lms,vix or lsp file. You then import it into your sequencer and should see what I previewed for you,
Cut and paste it ( here I get fuzzy, since I have used any sequencer software. I expect all three sequencer groups will give me suggestions on how to make this easiest for you.)

In summary
1) you must build your rgb using the target modeler . You just finished this step.
2) u need to select an effect from your personal library of effects u like. I have not released this tool yet. It should be released within the next 1-2 weeks.
3) optional at this point is to get an animated gif to preview
4) finally, you get an output file for your sequencer

STATUS as of feb 25th:

(1) is functional now. I know I will want to change the layout, the usability, the prompts and expand it to model matrix or grids.

(2) is maybe 70% done. Actually building the models, discovering what variables need to be changed, is complete.
What is left is another form where you select from this list
Spirals, bands, meteors, pictures,text
You then get prompted for the variables that belong to that effect. U name this effect and it goes into your library

(3) is done, it needs some clean related to scheduling Linux tasks, cleaning files and such. But the guts of it is done

(4) XML output for sequencers. I have LOR maybe 40% done, vixen 10%, LSP 5%. I need some people's models to work here, what people have done this weekend will help.

I have stated all 4 of these will be done by summer.

Thanks for the feedback

[smeighan]

over 40 people registered today. 30 models were created. I got some feedback and i fixed some bugs. A very good start.

Thanks

You are not allowed to view links. Register or Login

[rimist]

I know that I'm one of the few, but I decorate for many more holidays than Christmas. That being said, I wonder if you would consider as part of your long term plan some models using different shapes: a heart 2d/3d, a clover 2d/3d, a rainbow with a pot of gold, a flag, a tombstone, a pumpkin 2d/3d, a skull 2d/3d.

-Rimist (via Tapatalk)

[smeighan]

On another thread, i said sometime this year i plan on supporting Google Sketchup as an input for the target models.

This resulted in this thread


Hi;

"So if I send you an .ipt or .dwg, that's cool?"

it will be when i write the parser. i use google docs and have thought of expanding the places i can get a source file. i was thinking these three sources for target models.

1) rgbsb builds the target model
2) User provides and ascii file , they do all the work
3) take google docs, dwg files .etc.

Today, (1) is the only choice. I said i would look into google docs after i get some major milestones done

1) Build targets using rgnsb, Mostly Done. Need to allow user to delete models they dont want and to edit an existing model
2) Allow users to create effects into their library, In Progress
3) Allow user to select from a pull down of effects from their library and also select a target, this week
4) Produce an animated gof so user can see what that effect will look like on their virtual tree by next week
5) produce xml files for sequencers, 2-8 weeks
6) Create models for other rgb targets , maybe by june
7) New effects, movies, snowstorms, fireworks i will do this as a break from (5)
Write a user manual, before june
9) Look to using a google docs file as a target summer, fall

I am working on 2 & 3 now. I hope within the week people will be trying effect son their trees and seeingg the result on the web page.

I expect #5 to take 2-8 weeks.,

I want to be done with 1-5 by june.

thanks

[smeighan]

Released new target builder code Feb 27

This release now has

1) editing (yeah!). Click on a model name and it will load the form, edit and save. Still dont have the ability for you to delete models.
2) a unit of measure (inches or cm)
3) There is a summary at the bottom for the form showing what sequencers the users want to support.
4) Changed fields to light blue in form. I had a hard time seeing where i should type.
5) Fixed bug where "How will you build your tree" = UP/DOWN 180 was not working.
6) modified wording of this prompt.
old: **Total number of Pixels on this string
new: **Total number of Pixels on this string (Note, this is NOT the total pixels on your tree, just the count on one string):
Some people had inputted the total pixel count of their tree. Sorry, if it was not clear.

The hyperlink for editing was not complete and caused people to have a blank user name. I believe this is now fixed since editing should work.

thanks

[smeighan]

Hi all;

All of the original mega tree targets on the web page were 32 strands of lights, each strand had 35 pixels.
You are not allowed to view links. Register or Login

i am getting ready to release the next phase of rgbsb.

Phase 1) user can create a target modeling their tree. DONE
Phase 2) User will select from the effects classes (Spirals, Bands, text, Pictures, Meteors, .etc.) , customize the effect and save it in their private library. IN PROGRESS. Probably will be finished within next 7 days.
Phase 3) User select a target model and and effect and produces an animated GIF. IN PROGRESS, This is 75% complete, it will come one oine within next 14 days. This will be a big deal for the users. With this release you can try hundreds of effects on your tree and get a visual idea of what different effects look  like. I think this ability to simulate your tree will be popular.
Phase 4) User now chooses to output their data into an xml file for the 3 different sequencers. 2-12 weeks.

I will also start writing a user manual when Phase 3 comes online.

So to give you an idea of what this might look like, i built a different target model. Instead of 32 strands x 35 pixels, here is effects projected onto a 48 strand by 60 pixel tree.

6 spirals, 100ms delay, only show front half of tree



The pictures, as expected, project with better clarity.
 







And finally, text looks better



thanks

[smeighan]

Question/suggestion:

Should "pixel size" be introduced into your model?

With the current visualizations I see these nicely done pictures projected onto a tree and I think "that's nice!"  but I suspect when this is applied to my tree that the images may not be as recognizable.  My reason for suggesting this?  When I look at the nutcrackers I see one tree with a pixel size that is just as big as the gap between that pixel and the next.  in another nutcracker, the is very little gap.  i reality the smart pixels are about .39in... and the gap between pixels is 2.61in.

In a different element -- like a coroflake, the pixels sit behind corrugated plastic, and there is no "gap"... instead light from one pixel blends into the area of the pixels around it. 

there are other coro examples -- like david moore's candy canes -- where pixels are segregated into separate compartments so that there is little blending/bleeding. 

...in the coro examples we might define a pixel as a much larger visible area... maybe 3" circles or 3" squares.

yep, never any lack of tasks. We get one idea, it opens 3 more.

The answer is yes to all of your questions. They can be modeled more and more accurately. as Phase 3 of rgbsb (The ability to have your tree display your effects on an animated gif) comes online, I think people will want to use the tool to do some initial sizing.


So in gnuplot you can specify how big the dots are drawn (in world coordinates).

So i removed the background skeleton that i draw in gray. This was so people could visualize the tree.

Here are 4 images drawn in different dot sizes

0.3" dots


0.6"


1.0"


let gnuplot pick the default dot size


I dont think i can draw them exactly to size, because a gif images is not glowing like a light. maybe 0.6" is closest to real life?

[smeighan]

New release of target modeler. Many people were getting some of the prompts wrong.

I have added logic to check ranges on a number of fields. I build a calculated tree given that you told me how long one string is. Your string is the hypotenuse of the triangle. If you are doing "UP/DOWN NEXT" or "UP/DOWN 180", i cut the hypotenuse in half.

Still, Alpha. Tool may break, data may be lost.
 I did need to reset at least half of the models. I am backing up the user and model table every few days.

so ...

Please try to create 1 or 2 models.

I have also removed these 3 prompts
1)orientation (Where is the front of the tree . I will add this back in later this summer. really not needed for now)
2) direction of rotation of strings , CW or CCW. Again this really does not affect us getting a model out. It will only affect if you want text scrolling right to left or left to right.
3) Type of string, CCR or smart string. again , i may need to add it back this summer, but it does not affect generating sequences

The secret to getting the model tool to work  is to get the number of pixels per string and the distance between the first and last string correct. You can now let the tool calculate some of the other dimensions for you.

Being RED on the tool is like the tool saying "I don't believe you but i will build the model that way". Your dimensions will always be used, not my suggestions.

If people could give some feedback on my ranges

1) Number of strings you will build a megatree from (1-100)
2) Number of pixels on a string (40-130)
3) Spacing between pixels (2-6" or 5-30cm)

thanks

[Mimir]

The strings I am using for my tree are different.  I only have a total of 36 nodes per string, and they are about 10 inches between nodes.  What I was planning on doing is to make my tree around 7.5 (90in) tall and each string would actually start at the bottom go up, down, up, and down again.  This would give me 4 virtual string for each physical string for a total of 24 virtual strings using 6 physical strings.

I hope that made sense reading it.   

I think what you are doing is fantastic.

John

[smeighan]

The strings I am using for my tree are different.  I only have a total of 36 nodes per string, and they are about 10 inches between nodes.  What I was planning on doing is to make my tree around 7.5 (90in) tall and each string would actually start at the bottom go up, down, up, and down again.  This would give me 4 virtual string for each physical string for a total of 24 virtual strings using 6 physical strings.

I hope that made sense reading it.   

I think what you are doing is fantastic.

John

I t amkes sense and i was looking at your target model earlier today. I will add these other options

Up/DOWN NEXT. This is what i currently have. Go Down to up and then Up to down.

UP/DOWN NEXT3. This will mean your string goes down to up, then up to down and finally backup. I will assume the next string will start Up to down, down to up, and up to down. In other words i will wrap a string 3 times.


UP/DOWN NEXT4. Same as above , but i will wrap a string 4 times. This means, in your example, that each strand is 22.5" (one 90" string wrapped 4 times.

Does this compute?

[Mimir]

It does compute.    The only confusion I see is that each of the virtual strings will be 90".  Each physical string is 360".

Of course, when I think about it the Up/Down Next model will describe the addressing of the nodes. The part I was struggling with when trying to create my model was how to answer the prompts to get the model correct.

Thanks.