Implement aesthetic options for "colour.plotting.RGB_chromaticity_coordinates_chromaticity_diagram_plot_*" definitions. #391

tashdor · 2018-03-02T05:52:35Z

It would be great to have control over some parameters when plotting RGB samples on a CIE1931 diagram:
style (x, +, or circle)
fill color (black, white, rgb source color)
size (relative or width in pixels)

…gram_plot_CIE*" definitions. References #391.

KelSolaar · 2018-03-02T09:12:42Z

So I have added a scatter_parameters argument which get passed to pylab.scatter directly with defaults yielding the following plot:

Defaults

a = np.random.random((32, 32, 3))
RGB_chromaticity_coordinates_chromaticity_diagram_plot_CIE1931(a)

and with some customisations:

Custom Parameters

RGB_chromaticity_coordinates_chromaticity_diagram_plot_CIE1931(
    a,
    scatter_parameters={'c': 'black', 'marker':'+', 'alpha': 0.5}
)

This is available in latest develop branch.

nick-shaw · 2018-03-02T09:17:51Z

Nice. I've ended up temporarily editing the parameters in the module before, to make the points more visible. Probably not an advised approach!

KelSolaar · 2018-03-02T09:19:28Z

We might want to fine tune the defaults, the points were fine when plotting in Jupyter notebook but they are a tad big from command line right now.

WilliamCoulter · 2018-03-02T09:26:50Z

I've always thought that the CIE model in 2D could be enhanced in a few ways. Namely, the thickness of the line of the diagram is proportional the Y color matching function (brightness), as this says how much a color point of that wavelength is needed to pull other color points toward it. Also, X, Y, and Z appear on the diagram. X is at (1,0), Y is at (0,1) and Z is (0,0). Thank you for your time, William Coulter (559)-901-4788 "There is nothing noble in being superior to your fellow man. True nobility is being superior to your former self" -Ernest Hemingway

…

On Fri, Mar 2, 2018 at 1:17 AM, Nick-Shaw ***@***.***> wrote: Nice. I've ended up temporarily editing the parameters in the module before, to make the points more visible. Probably not an advised approach! — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#391 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AZGPY6eLsde4GIt887a3t6SEfYTAnQM9ks5taQ5CgaJpZM4SZb35> .

KelSolaar · 2018-03-02T09:38:03Z

@Willingo : Do you have an example image of what you are thinking about?

Cheers,

Thomas

WilliamCoulter · 2018-03-02T09:52:24Z

No but I can probably implement it myself in Matlab quickly. Imagine weighting the thickness of the locus to the intensity of the y color matching function for a given nanomenter of light. Like, 553nm would be thickest. I can draw the XYZ (1,0),(0,1),(0,0) it out and explain how I think it is useful. It allows one to know where a color point moves if you imagined increasing the intensity of an interval by a certain amount. I'm heading to sleep now though. Thank you for your time, William Coulter (559)-901-4788 "There is nothing noble in being superior to your fellow man. True nobility is being superior to your former self" -Ernest Hemingway

…

On Fri, Mar 2, 2018 at 1:38 AM, Thomas Mansencal ***@***.***> wrote: @Willingo <https://github.com/willingo> : Do you have an example image of what you are thinking about? Cheers, Thomas — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#391 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AZGPY5BeLv64DxGpOG18zLIipBONMUruks5taRMAgaJpZM4SZb35> .

WilliamCoulter · 2018-03-02T09:54:08Z

It effectively gives you lumen / radiant watt. 't help much except to realThis is useful in the lighting industry for energy reasons as well as interpreting reflective curves, but maybe it's not that useful in general. Thank you for your time, William Coulter (559)-901-4788 "There is nothing noble in being superior to your fellow man. True nobility is being superior to your former self" -Ernest Hemingway On Fri, Mar 2, 2018 at 1:52 AM, William Coulter <wwcoulter6@gmail.com> wrote:

…

No but I can probably implement it myself in Matlab quickly. Imagine weighting the thickness of the locus to the intensity of the y color matching function for a given nanomenter of light. Like, 553nm would be thickest. I can draw the XYZ (1,0),(0,1),(0,0) it out and explain how I think it is useful. It allows one to know where a color point moves if you imagined increasing the intensity of an interval by a certain amount. I'm heading to sleep now though. Thank you for your time, William Coulter (559)-901-4788 <(559)%20901-4788> "There is nothing noble in being superior to your fellow man. True nobility is being superior to your former self" -Ernest Hemingway On Fri, Mar 2, 2018 at 1:38 AM, Thomas Mansencal ***@***.*** > wrote: > @Willingo <https://github.com/willingo> : Do you have an example image > of what you are thinking about? > > Cheers, > > Thomas > > — > You are receiving this because you were mentioned. > Reply to this email directly, view it on GitHub > <#391 (comment)>, > or mute the thread > <https://github.com/notifications/unsubscribe-auth/AZGPY5BeLv64DxGpOG18zLIipBONMUruks5taRMAgaJpZM4SZb35> > . >

KelSolaar · 2018-03-02T10:27:22Z

@Willingo : Just to confirm, did you have something like that in mind?

tashdor · 2018-03-02T17:19:36Z

This is all working great! What is the best way to limit the number of samples created by read_image to speed up performance?

…

-- Tashi Trieu 720-206-4229 Tashi@TashiTrieu.com

On Fri, Mar 2, 2018 at 2:27 AM, Thomas Mansencal ***@***.***> wrote: @Willingo <https://github.com/willingo> : Just to confirm, did you have something like that in mind? [image: image] <https://user-images.githubusercontent.com/99779/36894696-1d6c6f92-1e71-11e8-815a-93d2f628f535.png> [image: image] <https://user-images.githubusercontent.com/99779/36894711-2bc09f1e-1e71-11e8-9981-acce6d2fe99f.png> — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#391 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AA1ex6u76Pgelm-iIbHXG0qmjQWdVoCrks5taR6UgaJpZM4SZb35> .

KelSolaar · 2018-03-02T19:20:31Z

You could try RGB[::2, ::2, 3], RGB[::4, ::4, 3] to take every second or fourth pixels in the image, akin to nearest neighbour interpolation in Nuke or Photoshop. The clean way would be to resample the array with scipy.misc.imresize although I noticed it is on its way to deprecation.

tashdor · 2018-03-02T19:24:30Z

I'm getting the error "IndexError: index 3 is out of bounds for axis 2 with size 3" when using "RGB[::4, ::4, 3]"

…

-- Tashi Trieu 720-206-4229 Tashi@TashiTrieu.com

On Fri, Mar 2, 2018 at 11:20 AM, Thomas Mansencal ***@***.***> wrote: You could try RGB[::2, ::2, 3], RGB[::4, ::4, 3] to take every second or fourth pixels in the image, akin to nearest neighbour interpolation in Nuke or Photoshop. The clean way would be to resample the array with scipy.misc.imresize <https://docs.scipy.org/doc/scipy/reference/generated/scipy.misc.imresize.html> although I noticed it is on its way to deprecation. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#391 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AA1exwcScRPeVmPjMoa5Ipj_wSNTtJVeks5taZuFgaJpZM4SZb35> .

KelSolaar · 2018-03-02T19:35:02Z

Sorry, my array slicing is incorrect, try something like that: RGB[::2, ::2]

KelSolaar · 2018-03-03T06:14:13Z

I have updated the overall plotting and diagrams generation code, notable things:

The colour.plotting.DEFAULT_PLOTTING_COLOURSPACE attribute can be overriden to define the colourspace used to render the relevant figures colours.
The Chromaticity Diagram generation has been refactored so that tuning the Spectral Locus is easier:

import numpy as np
from colour.plotting import *

colour_plotting_defaults()

RGB = np.random.random((32, 32, 3))

RGB_colourspaces_chromaticity_diagram_plot_CIE1931(
    ['ITU-R BT.709', 'ACEScg', 'S-Gamut', 'Pointer Gamut'])

RGB_chromaticity_coordinates_chromaticity_diagram_plot_CIE1960UCS(
    RGB,
    'ITU-R BT.709',
    colourspaces=['ACEScg'],
    show_diagram_colours=False,
    spectral_locus_colours='RGB')

RGB_chromaticity_coordinates_chromaticity_diagram_plot_CIE1976UCS(
    RGB,
    'ITU-R BT.709',
    colourspaces=['ACEScg'],
    show_diagram_colours=False,
    spectral_locus_colours='RGB',
    spectral_locus_labels=[])

tashdor · 2018-03-03T06:44:50Z

Thanks! This is fantastic!

KelSolaar · 2018-03-03T06:52:38Z

Thanks @tashdor! If you need anything else, let me know, I haven't implemented @Willingo suggestions yet as I'm not sure about what he wanted.

KelSolaar · 2018-03-03T07:02:12Z

Oh and I removed most (still a few to go through) of the unneeded hardcoded styling elements, thus you can override globally marker size, line width, etc...:

import matplotlib.pyplot as plt
import matplotlib as mpl
import numpy as np

from colour.plotting import *

plt.style.use(['dark_background'])
mpl.rcParams['lines.linewidth'] = 1
mpl.rcParams['lines.markersize'] = 2

RGB = np.random.random((32, 32, 3))

RGB_chromaticity_coordinates_chromaticity_diagram_plot_CIE1960UCS(
    RGB,
    'ITU-R BT.709',
    colourspaces=['ACEScg'],
    spectral_locus_colours='white',
    grid=False)

tashdor · 2018-03-03T07:03:37Z

Very cool. Is there a way to produce a diagram with alpha channel and with out the frame, header, or axes?

KelSolaar · 2018-03-03T07:16:30Z

Something like that should do:

import matplotlib.pyplot as plt
import matplotlib as mpl
import numpy as np

from colour.plotting import *

plt.style.use(['dark_background'])
mpl.rcParams['lines.linewidth'] = 1
mpl.rcParams['lines.markersize'] = 2

RGB = np.random.random((32, 32, 3))

figure = RGB_chromaticity_coordinates_chromaticity_diagram_plot_CIE1960UCS(
    RGB,
    'ITU-R BT.709',
    colourspaces=['ACEScg'],
    spectral_locus_colours='white',
    grid=False,
    title='',
    standalone=False,
    legend=False)

figure.patch.set_visible(False)
figure.gca().axis('off')
figure.tight_layout()

with open('/Users/kelsolaar/Downloads/diagram.png', 'w') as png_file:
    figure.canvas.print_png(png_file)

tashdor · 2018-03-03T07:20:34Z

Perfect!

KelSolaar · 2018-03-06T08:52:02Z

Closing this one for now!

KelSolaar changed the title ~~Implement aesthetic options for RGB_chromaticity_coordinates_chromaticity_diagram_plot_CIE1931~~ Implement aesthetic options for "colour.plotting.RGB_chromaticity_coordinates_chromaticity_diagram_plot_*" definitions. Mar 2, 2018

KelSolaar added Enhancement API Normal labels Mar 2, 2018

KelSolaar added this to the v0.3.12 milestone Mar 2, 2018

KelSolaar added a commit that referenced this issue Mar 2, 2018

Update "colour.plotting.RGB_chromaticity_coordinates_chromaticity_dia…

9dbda83

…gram_plot_CIE*" definitions. References #391.

KelSolaar added P2 and removed Normal labels Mar 6, 2018

KelSolaar closed this as completed Mar 6, 2018

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Implement aesthetic options for "colour.plotting.RGB_chromaticity_coordinates_chromaticity_diagram_plot_*" definitions. #391

Implement aesthetic options for "colour.plotting.RGB_chromaticity_coordinates_chromaticity_diagram_plot_*" definitions. #391

tashdor commented Mar 2, 2018

KelSolaar commented Mar 2, 2018 •

edited

Loading

nick-shaw commented Mar 2, 2018

KelSolaar commented Mar 2, 2018

WilliamCoulter commented Mar 2, 2018 via email

KelSolaar commented Mar 2, 2018

WilliamCoulter commented Mar 2, 2018 via email

WilliamCoulter commented Mar 2, 2018 via email

KelSolaar commented Mar 2, 2018

tashdor commented Mar 2, 2018 via email

KelSolaar commented Mar 2, 2018

tashdor commented Mar 2, 2018 via email

KelSolaar commented Mar 2, 2018

KelSolaar commented Mar 3, 2018

tashdor commented Mar 3, 2018

KelSolaar commented Mar 3, 2018

KelSolaar commented Mar 3, 2018

tashdor commented Mar 3, 2018

KelSolaar commented Mar 3, 2018

tashdor commented Mar 3, 2018

KelSolaar commented Mar 6, 2018

Implement aesthetic options for "colour.plotting.RGB_chromaticity_coordinates_chromaticity_diagram_plot_*" definitions. #391

Implement aesthetic options for "colour.plotting.RGB_chromaticity_coordinates_chromaticity_diagram_plot_*" definitions. #391

Comments

tashdor commented Mar 2, 2018

KelSolaar commented Mar 2, 2018 • edited Loading

nick-shaw commented Mar 2, 2018

KelSolaar commented Mar 2, 2018

WilliamCoulter commented Mar 2, 2018 via email

KelSolaar commented Mar 2, 2018

WilliamCoulter commented Mar 2, 2018 via email

WilliamCoulter commented Mar 2, 2018 via email

KelSolaar commented Mar 2, 2018

tashdor commented Mar 2, 2018 via email

KelSolaar commented Mar 2, 2018

tashdor commented Mar 2, 2018 via email

KelSolaar commented Mar 2, 2018

KelSolaar commented Mar 3, 2018

tashdor commented Mar 3, 2018

KelSolaar commented Mar 3, 2018

KelSolaar commented Mar 3, 2018

tashdor commented Mar 3, 2018

KelSolaar commented Mar 3, 2018

tashdor commented Mar 3, 2018

KelSolaar commented Mar 6, 2018

KelSolaar commented Mar 2, 2018 •

edited

Loading