On 3/19/2017 5:49 PM, clipka wrote:
> Am 18.03.2017 um 23:08 schrieb Mike Horvath:
>> On 3/18/2017 4:36 PM, Mike Horvath wrote:
>>> On 3/17/2017 4:49 AM, clipka wrote:
>>>> There. Now you know the common property of all reflection spectra that
>>>> map to the surface of the shape you're after. I hope this helps you
>>>> devise an algorithm to compute them in a systematic fashion.
>>>>
>>>
>>> I was able to get some coding help from Bruce Lindbloom to generate the
>>> XYZ coordinates. So, I have points and I have colors. But I do not have
>>> a mesh. Do you have any tips on generating the mesh? Did you use one
>>> color for each triangle, or did you create some sort of gradient?
>
> I created and linked the points according to the following scheme:
>
>     .###.... -- ..###... -- ...###.. -- ....###.
>          \        /  \        /  \        /
>           \      /    \      /    \      /
>            \    /      \    /      \    /
>             \  /        \  /        \  /
>           ..##.... -- ...##... -- ....##..
>             /  \        /  \        /  \
>            /    \      /    \      /    \
>           /      \    /      \    /      \
>          /        \  /        \  /        \
>     ..#..... -- ...#.... -- ....#... -- .....#..
>
> where each `.` represents an interval in the spectrum that's at 0% power
> and each `#` represents an interval in the spectrum that's at 100% power.
>
> - On each line from left to right, the "box" spectra have a constant
> width, but the starting point and end point shift from blue to red.
>
> - On each line from bottom left to top right, the "box" spectra have a
> constant starting point, but their width increases as the end point
> shifts from blue to red.
>
> - On each line from top left to bottom right, the "box" spectra have a
> constant end point, but their width decreases as the starting point
> shifts from blue to red.
>
> Remember to define a point where your spectrum wraps around from red to
> blue. The "mod" function helps a lot with that.
>
>> Also, is the shape stacked like a wedding cake? Are there the same
>> number of sections in each layer?
>
> The layers shown as horizontal lines in the diagram above will /not/ be
> perfectly horizontal in the colour space, due to the varying brightness
> associated with each wavelength interval; but each layer will have the
> same number of points.
>

Did you use a function to color the shape, or did you color each 
triangle in a solid color? I am doing the latter at the moment using the 
xyz2RGB() macro that comes with Lightsys. It works okay, but you can see 
each and every triangle as a result. I don't like that.

I also tried re-using the function I used to color the sRGB gamut, and 
the result is very wonky and random looking. Not good.


Mike

Post a reply to this message