by

Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • 00:22

    KelSolaar on develop

    Ensure that CMFS are properly p… Merge branch 'feature/v0.3.16' … (compare)

  • May 27 16:45
    nick-shaw commented #591
  • May 26 05:02

    KelSolaar on develop

    Add "Cedric Lejeune" to sponsor… Merge branch 'feature/v0.3.16' … (compare)

  • May 26 05:01

    KelSolaar on v0.3.16

    Add "Cedric Lejeune" to sponsor… (compare)

  • May 26 04:58

    KelSolaar on v0.3.16

    Add "Cedric Lejeune" to sponsors (compare)

  • May 26 04:48

    KelSolaar on v0.3.16

    (compare)

  • May 25 11:19

    KelSolaar on idt

    (compare)

  • May 25 11:19

    KelSolaar on incorrect_deprecation

    (compare)

  • May 25 11:18

    KelSolaar on venice

    (compare)

  • May 25 11:18

    KelSolaar on hexadecimal

    (compare)

  • May 25 11:18

    KelSolaar on precision

    (compare)

  • May 24 03:13

    KelSolaar on develop

    Ensure that automatic colour co… Ensure that most "Colour" honou… Document new environment variab… and 2 more (compare)

  • May 24 03:13
    KelSolaar closed #593
  • May 24 01:06
    coveralls commented #593
  • May 24 00:56
    coveralls commented #593
  • May 24 00:54
    coveralls commented #593
  • May 24 00:32
    coveralls commented #593
  • May 24 00:31
    KelSolaar commented #593
  • May 24 00:31
    KelSolaar commented #593
  • May 24 00:30
    coveralls commented #593
MirageInc
@MirageInc
If you'd be interested in including a thin-film interference piece of code just a bit like your blackbody stuff, I'd be happy.
You'd be required to give an angle of incidence for your light, the thickness of your film, the refractive index of your liquid OR say that it's water and instead give a temperature and wavelength (because I can then calculate the respective index directly given a separate empirical algorithm I know of) and then we can generate a similar plot to your blackbody stuff, such as a spectrum and the XYZ overall look.
Equally I could give you the kind of code I'm actually using in my project, where it can generate a spectrum with different thicknesses.
Either way, depending on what you're interested in adding, I can try to work that extra bit harder and give you something workable that is compatible with the rest of your library.
Thomas Mansencal
@KelSolaar
Both seems equally great, no rush anyway but it sounds interesting enough that I thought I would ask!
gustain
@gustain
Dear Thomas,
Dear Thomas, I would like to obtain the values that are used for the D65 illuminants. By reading the docs I found how to plot it, but do I have the possibility to get the actual value, as a array or a dictionary ? Thank you
gustain
@gustain
Sorry to bother I just found it !
Thomas Mansencal
@KelSolaar
Hi! Glad you found them, for other people reading they are available like that:
  • Chromaticity coordinates: colour.ILLUMINANTS['CIE 1931 2 Degree Standard Observer']['D65']
  • Chromaticity coordinates alias: colour.ILLUMINANTS['cie_2_1931']['D65']
  • Spectral distribution: colour.ILLUMINANTS_SDS['D65']
MirageInc
@MirageInc
Hey Thomas
I have finished with what I was working on with my project
and I have written a "sanitised" file containing the functions I thought you might like to add to the library
and it contains an explanation of the things.
How can I send it to you?
Once you see it, you probably won't be that impressed.
Thomas Mansencal
@KelSolaar
If you want to be listed as contributor to the repository, the easiest is probably to fork Colour and drop the file maybe in the colour/phenomena directory or anywhere you think most of the code should be and push back to your fork
I can then work from your commit so that you would be shown in the history! :)
MirageInc
@MirageInc
Just because I'm not the smartest guy and also because I don't really know how to use Git, I'll explain what I did:
I forked Colour
then I added my file to colour/phenomena
then I made a pull-request for this fork...
I assume that works?
And again, I've stripped down all the stuff to just 3 functions that a general user would want to use and I hope they're cool enough to be included..
MirageInc
@MirageInc
Yeah... I expected the pull-request to wait for some kind of approval but I think it was outright blocked...
Thomas Mansencal
@KelSolaar
Thanks I can see the PR here: colour-science/colour#525
Thomas Mansencal
@KelSolaar
I will ask a few questions over there!
Lypheo
@Lypheo
Hi there! I’ve got a question about the ST2084 EOTF that’s been bugging me for a while, and this is the only public chatroom I could with people that likely know their shit about, well, colour stuff, so here goes… The SMPTE 2084 standard specifies that PQ-encoded signals are to be displayed at EOTF(signal) cd/m^2 if all RGB components are equal. Now of course this begs the question what physical luminance signals like, for example, RGB = (0.1, 0.2, 0.1) should correspond to. As far as I can tell, the standard doesn’t address this. Does anyone have an idea?
Thomas Mansencal
@KelSolaar
Hi Lypheo, you can ask that on Github too! Each channel of the signal is encoded the same way with the inverse EOTF, irrespective of the per-channel value, they are decoded likewise.
Lypheo
@Lypheo
Yes, I realize that. But I am unsure how to interpret PQ-encoded RGB channels, because ST2084 is meant to be an absolute transfer function, so a specifc color value should map unambiguously to some physical output luminance of the respective subpixel. However, the standard only specifies what luminance it maps to in the case that all channels are equal (R=G=B).
How, then, do I determine the intended subpixel luminances of a PQ-encoded RGB signal which is not greyscale?
Lypheo
@Lypheo
grafik.png
here’s the relevant paragraph
If I’m getting this right, it implies that a linear value of, say, R = 0.01 should correspond to an output of 100 nits for the red subpixel
Lypheo
@Lypheo
which is another point of confusion for me: since nits are a measure of luminous intensity and thus account for the differences in perception, wouldn’t an output of 100 nits on all 3 subpixels lead to non-white output? I was under the impression that equal amounts (not luminance) of R, G, B would produce white
Thomas Mansencal
@KelSolaar
A ST 2084/PQ encoded value of 1 is equivalent to 10000cd/m2
nits are a measure of luminance (not luminous intensity), it is weighted per surface area in a given direction
you are already in the photometric realm here, so equal channel values should yield white
Alexey Kuleshevich
@lehins
In case anyone is doing Haskell here. I wrote a library for color space conversion, chromatic adaptation etc. https://github.com/lehins/Color
I used python colour package quite a bit to teach myself the concepts or to validate my understanding, so I thought I'd share with you.
Thomas Mansencal
@KelSolaar
Hi @lehins, that is great!
I'll add it to awesome colour!
Thomas Mansencal
@KelSolaar
Btw, if Colour has been useful, a simple mention or comment about it somewhere would be sweet!
Alexey Kuleshevich
@lehins
@KelSolaar Thank you for addition of Color to awesome-colour.org and for your feedback. I added some links in the readme, to colour and awesome-colour.org
Thomas Mansencal
@KelSolaar
Cheers @lehins ! :)
And you are welcome
OmarWagih1
@OmarWagih1
Hello all, i was interested in joining Colour-Science for GSoC, I'm mostly interested in the benchmarking and performance improvement project but I had a question, would you like to setup a new repo for benchmarking so as not to add to Colour's dependancies (and make it easier to run the benchmarking) or would you like it to be on the main repo? Thanks!
Thomas Mansencal
@KelSolaar

Hello!

Nice to meet you @OmarWagih1!

Would you mind sending a mail to colour-developers@colour-science.org please? We will follow-up there!

Cheers,

OmarWagih1
@OmarWagih1
Hello, nice to meet you too! sure will do.
marcelo claro laranjeira
@MarceloClaro
Good night ... I would like some help, I need to plot hex to munsull ... how can I convert?