Snowflake X/Y vs R/Theta geometry

[Dave L]

Nutcracker wizards.
I'm a newbie so forgive my vernacular if I get a few terms wrong.   I have a coro snowflake that I would like to apply models / effects to.  The geometry of the snowflake is 5 long radial arms with a smaller 5-spiked radial snowflake on the end of each long arm.  (This is much easier to look at and get the idea than to describe.)  So it seems to me that the nutcracker effects are mainly designed for rectangular X-Y type geometry (and they are great).  The snowflake is easier described by radial (R theta) geometry.  So, for example how do I model and get an effect where each leg might sequence around and then each smaller (end-flake) also circle around? Sort of like a ferris whell.  I know I can sort of reorder this visually into a vertical X-Y matrix - but visually the nutcracker effect looks much different in a X-Y geometry as compared to radial geometry.   If you have had luck with snowflakes or have alternative ways that I should think about this - I be glad to hear your thoughts.   Does nutcracker use different geometries or are all effects basically applied to an X-Y matrix?

Thanks for helping a NOOB.

[Steve Gase]

I have 9 coroflakes, 5 arms with 10 pixels each...  I do several things with them...

(1) custom model where the x/y coordinates of each pixel is mapped to a flat matrix

this is great for bars, butterfly, plasma, concentric rings, and other effects

(2) arches of 10 pixels each... so each flake is 5 arches

this is great with single strand, bars, and other effects.... it also gets a moving-concentric set of rings.

(3) arch of 50 pixels.

this is good with single strand, and and other effects...  it creates more of a moving wagon wheel effect.


In the preview  and nutcracker visualizers I show the custom model so that I can view how it will be seen, even though the effects are applied to one of the 3 models.

the point is that you can do a lot with nutcracker and a coroflake.

[Dave L]

Steve,

Thank you for the insights.  So what I am starting to understand is that since the coro flake is a unique geometry you have to in a sense transform (in your head) simpler geometries and apply them to the flake.

For case 1, your example of bars,  I am imagining you just start numbering you pixels from the center out on any given arm and then keep counting up arm after arm.  So first arm is 1 - 10.  2nd arm is 11 - 20. etc.  The nutcracker does its magic in scanning the square matrix in a bar pattern.  This can apply linear wipes, (up or down) and the like.

For case 2,  I can visualize the concentric expanding effect and then sort of a little spin around the outer sub- flake (or flaklet).

For case 3,  I can visualize the wagon wheel if you had a sort of chase with a chase group of 10 pixels.

How have you numbered you pixels as you get to the outer flaklet, do you just run around the points sequentially and always in the same "clock-sense"?

I know there is a long list of future feature enhancements, possibly there could be some effects specific to radial type geometries.

Thanks for your comments and helping me grasp these concepts.

Dave

[Steve Gase]

You should number the pixels in the order that they are found in the string.  The first pixel after your controller is '1', the second is '2' and so on.  So your pixel placement may differ from mine based on how you laid down your pixels.

I found it easy to take a picture of my coroflake -- ideally after you placed the pixels and before you glue down the cover.  Then use this picture as a background into excel and start numbering cells on top of the pixels.

I did mine in fairly high resolution... a 100x100 grid so that I'd have my custom model fairly precise.  I'd show you with a picture, but it is hard to resize the grid to where it shows my point.

as far as my placement, i started with the top arm, and the left-most 'bud' and numbered around that arm in counter-clockwise fashion, continuing down the the center of the flake (this was 1-10)... then continued with my string of pixels to the next arm in counter-clockwise fashion.  i chose this order and starting point in an arbitrary manner.  with a customer model, the order is not important as long as the numbers match up to the order on the string.

my order did lend itself to models (2) and (3).

btw... with 9 coroflakes I used the manual string channel assignment... and with the model (2) I combined the 9 flakes with 5 arms into a single model... 45 arches. 

similarly in the 50-pixel arch, i has a single model that had 9x50pixel arches.  this allowed the 9 flakes to work in unison.

...just found the original picture I used in the discussions: