search

colour-science / colour

Colour Science for Python
https://www.colour-science.org
BSD 3-Clause "New" or "Revised" License
2.14k stars 263 forks source link

Add MacAdam ellipses plotting example. #18

Closed KelSolaar closed 6 years ago

KelSolaar commented 10 years ago

References

iinnovations commented 10 years ago

What is the plan for this? I have not found a well-reasoned MacAdam ellipse interpolation example from the literature. Or is the plan here to simply allow plotting of the data as recorded? There are a number of adhocs available, e.g. the fluorescent lighting standard.

KelSolaar commented 10 years ago

I created the issue more to remember to take a look at it but I have no particular plan for it right now.

iinnovations commented 10 years ago

Thanks Thomas. Do ellipse and/or ANSI bin generators currently exist in colour? Tests for inclusion? I have this code elsewhere and would consider porting them.

On Nov 16, 2014, at 10:26 AM, Thomas Mansencal notifications@github.com wrote:

I created the issue more to remember to take a look at it but I have no particular plan for it right now.

— Reply to this email directly or view it on GitHub.

KelSolaar commented 10 years ago

Neither of those exists right now! That would be very nice if you were to implement support for them actually!

nschloe commented 6 years ago

I'm currently trying to get MacAdams ellipses plotted in colorio, but I'm somewhat at loss on how to interpret the data in the original paper. Table 3 therein gives the ellipse data, but when plotting the x-y locations of the filter pairs, their connecting lines, and the average theta settings, I'm getting all but a coherent image from which one can deduce an ellipse:

ma1

ma

How does everyone draw those at all?

KelSolaar commented 6 years ago

I have never looked at that beyond saying would be great to have :)

nschloe commented 6 years ago

Got it:

d

Turns out that you have to use the delta y/delta x and delta s data (plus a curve fit) to reproduce the ellipses. I still have no idea how to correlate this with the actual measurements, but I guess that goes to show that you'll have to be very cautious when using MacAdam data for any quantitative studies.

KelSolaar commented 6 years ago

Nice! PR Welcome ;)

iinnovations commented 6 years ago

You’re going to have to parametrize these so that you can feed them into a geometry point in feature function anyway. I’ve done this in other languages and will run into this shortly here in python. I’ll send a PR when I get to it.

C

On Feb 28, 2018, at 11:55 AM, Thomas Mansencal notifications@github.com wrote:

Nice! PR Welcome ;)

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.

KelSolaar commented 6 years ago

Thanks Colin (@iinnovations)!

KelSolaar commented 6 years ago

Gave it a stab as I needed to plot them in another diagram than CIE 1931 while using ellipse canonical form coefficients. We have a few new definitions to deal with that now, conversion from canonical to general form coefficients, ellipse fitting using Halir and Flusser (1998) method and a few plotting definitions.

image

Closing this one!

vvclin-git commented 4 years ago

You can calculate MacAdam ellipse using FMC equations. https://www.osapublishing.org/josa/abstract.cfm?uri=josa-61-1-118