Numbering:
My pixel numbering is based on my original attempt to reduce the number of cuts and splices as I add pixels to the base of coroflake. I start from the top branch going clockwise, then down the trunk -- then repeat for each branch in a clockwise fashion. No doubt my wiring was a personal decision, but once the first was implemented, each of my coroflakes were done in the exact way.
In trying to sequence this, I often come back to logical groupings where I can apply the same/similar colors to a common group:
1) a panel/matrix with a coroflake-shaped mask
2) Line-drawing -- the snake
3) concentric rings
4) group by branch... group 1 has pixels 1-10, group 2 has pixels 11-21, etc.
5) unrolled to create a panel
The mask:Pretty straight-forward... pixels are available on a panel -- but you can only light those pixels which are physically present -- in effect, a coroflake-shaped mask is applied to the existing effects. Its nice, but nothing great.
The snake:This is not doing anything special, but some effects like "Garland" would work nicely. Lights can chase around the flake and accumulate at the end of the string.
Concentric Rings:In this case, I want the effects to radiate outwards as rings -- there are several places where the pixel density is higher, and it leads to defining circles.. circles that rotate like wheels, sometimes in opposing directions.
Flattened-Trees:If you "unroll" the coroflake in your mind the branches become a grove of trees, then some of the effects like "Fire" work nicely -- radiating from the center outward.
Inverted Trees:Sometimes you might want to radiate inward... so a model which is flipped around the center axis would be nice.
Starburst: (naming apologies to combusionmark)The starburst, or porcupine ball is similar to a coroflake... strings (even irregular strings in the case of a coroflake) radiate from a center point. In the starburst it is more interesting because it happens in 3 dimensions. Rings, trees, branches/twigs, snakes, maybe matrix all can be applied to this element.
Some conclusions/observations:A) If
ANY effect could be flipped along an axis, then the need for inverted trees is not there. Maybe other effects could benefit... reverse garland, upward spirals, reverse meteor showers, etc.
B) if you unroll the model (the flattened trees) then the concentric circles become easier.
C) if you attempted to
flip between effects by using multiple models -- a "COROFLAKE-TREES", a "COROFLAKE-RINGS", a "COROFLAKE-BRANCHES", then you'd want the ability to reuse the same channel numbers between the models... and, if you were using the RINGS model for one phrase and switched to the TREES model in the next phrase, the RINGS model should not interfere with anything else that shared the same channels.
Is there a generic solution to these problems?
D)
multiple models sharing channel numbers which allow the pixels to take on new roles seems useful.
E) an
intermediate mapping layer that translates between the "virtual" matrix, and the channel assignments seems to be necessary... remapping to create a new model over the pixels would help a lot!
Thoughts?