Disclaimer: I know very little numpy
But if you don't have access to a lookup table (LUT) that actually defines
your non-linear color mappings, then numpy/scipy has functions for
interpolating values along a curve. Someone else can probably make this a
lot better, but here is something that is linearly interpolating from your
samples:
In [151]: import numpy as np
In [152]: rgb_samples = np.array([(0.0, 0.0, 0.0),
(0.044156629592180252, 0.044154495000839233,
0.044151946902275085),
(0.18418028950691223, 0.18418741226196289,
0.18419596552848816),
(0.48274806141853333, 0.48278585076332092,
0.48283094167709351),
(0.99991166591644287, 1.0000128746032715, 1.0001327991485596)])
In [153]: y_sample = np.array([0, 255*.25, 255*.5, 255*.75, 255])
In [154]: np.interp(.33, rgb_samples[:,0], y_sample)
Out[154]: 158.63533146734287
In [155]: np.interp(.33, rgb_samples[:,1], y_sample)
Out[155]: 158.63061312361012
In [156]: np.interp(.33, rgb_samples[:,2], y_sample)
Out[156]: 158.62497845173215
In [157]: float_reds = np.linspace(0., 1., 10)
In [158]: np.interp(float_reds, rgb_samples[:,0], y_sample)
Out[158]:
array([ 0. , 94.23304979, 135.62268918, 159.34706267,
183.07143616, 200.22487316, 213.92137707, 227.61788098,
241.3143849 , 255. ])
It is not going to be as accurate as having the entire LUT, but it
will be close.
scipy has a more advanced version called interp1d, which supports
different interpolation types, and will return you an interp function
you can reuse:
In [163]: import scipy as sp
In [164]: fit_r = sp.interpolate.interp1d(rgb_samples[:,0], y_sample,
copy=False)
In [165]: fit_g = sp.interpolate.interp1d(rgb_samples[:,1], y_sample,
copy=False)
In [166]: fit_b = sp.interpolate.interp1d(rgb_samples[:,2], y_sample,
copy=False)
In [167]: fit_r(.33)
Out[167]: array(158.63533146734287)
In [168]: fit_g(.33)
Out[168]: array(158.63061312361012)
In [169]: fit_b([.33, .45, .785736])
Out[169]: array([ 158.62497845, 184.24153582, 228.5786838 ])
In [170]: fit_b([.33, .45, .785736]).round().astype(int)
Out[170]: array([159, 184, 229])
On Sat, Jan 25, 2014 at 9:16 AM, Justin Israel <[email protected]>wrote:
> Yea at first I was going to suggest a fit function, because it seemed you
> wanted to remap your float 0-1 values to a new range between 0-255 ( or
> 20-235, etc). But towards the end of your later reply it does seem like you
> need to take a LUT into consideration.
> If your float values are already mapped through a LUT, then I would think
> you can just map them to 0-255 directly.
>
>
> On Sat, Jan 25, 2014 at 6:03 AM, Daz <[email protected]> wrote:
>
>> Heya
>>
>> Yes I know the basic conversion but that dont match the program.
>>
>> Basically if I set the color to be 0.25/0.25/0.25 of range value 0-1
>>
>> And then read the value that get processed to program it actually
>> is (0.044156629592180252, 0.044154495000839233, 0.044151946902275085) Which
>> makes me think that its being either gamma correct or something else in
>> mean time.
>>
>> I thought that I was facing simply vector conversion but it looks like i
>> have to find and reproduce the math for the curved emmm values in order to
>> match the app. Bit annoying for the moment I found work around but I'd
>> still want to know how to reproduce the match - or reverse engineer I
>> guess?
>>
>> Thanks for help.
>>
>> On Friday, 24 January 2014 16:48:25 UTC, damonshelton wrote:
>>
>>> 255.0 * each element for each color will give you what those 0-255 rgb
>>> values are, but not sure where you got your initial percentages versus
>>> values came from.
>>> 25% of 1.0 would be (.25, .25, .25)
>>>
>>>
>>> On Fri, Jan 24, 2014 at 6:41 AM, Daz <[email protected]> wrote:
>>>
>>>> Heya
>>>>
>>>> I'm trying to remap 3 dimensional vector to values of 0-255
>>>>
>>>> Can some1 suggest what algorithm I need to use to convert those vector
>>>> values to more human rgb values?
>>>>
>>>> I know that
>>>>
>>>> 0% = (0.0, 0.0, 0.0) - black
>>>> 25% = (0.044156629592180252, 0.044154495000839233, 0.044151946902275085)
>>>> 50% = (0.18418028950691223, 0.18418741226196289, 0.18419596552848816)
>>>> 75% = (0.48274806141853333, 0.48278585076332092, 0.48283094167709351)
>>>> 100% = (0.99991166591644287, 1.0000128746032715, 1.0001327991485596) -
>>>> white
>>>>
>>>> Thanks, bye.
>>>>
>>>>
>>> --
>> You received this message because you are subscribed to the Google Groups
>> "Python Programming for Autodesk Maya" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to [email protected].
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/python_inside_maya/0ff4eb6a-d890-4de2-966d-428758aa8d46%40googlegroups.com
>> .
>>
>> For more options, visit https://groups.google.com/groups/opt_out.
>>
>
>
--
You received this message because you are subscribed to the Google Groups
"Python Programming for Autodesk Maya" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/python_inside_maya/CAPGFgA0E3-DjORM2pvM53F4heLJZ%2B1GH%3D4%2BHgs22t%3D1UhoVgWg%40mail.gmail.com.
For more options, visit https://groups.google.com/groups/opt_out.
