Representing a limited angular range in maps

[maclariz]

I know I can get colours for whole stereographic triangle in IPF maps, which is very nice for polycrystalline materials. But my one application is a large single crystal with a lot of deformation in it. So, what would be nice would be something that focuses in on just one area of orientations around a specific direction to show up the subgrains and deformation more obviously. How do I do that? Maybe @hakonanes you might know of a workaround for this?

[argerlt]

Short answer: Here is an MTEX example discussing one solution to a variation of your question

Longer answer:
Zooming out a bit, most IPF colormaps use the method described in This Paper, which requires having exactly three vectors that together define a unique barycentric coordinate for every possible rotation in the fundamental zone.

This falls apart for what the application you are talking about, where I think you want either 2 or 3 orthogonal axis, which is either 4 or 6 unique vectors. The three suggestions I would make would be:

1. plot misorientation angle relative to the per-grain mean (What OIM calls GROD).
2. calculate the misorientation relative to the per-grain mean as a quaternion, then use the q1,q2,q3 components to give an RGB color value. maybe something like (mis.data[...1:]+0.5)**0.5.
3. calculate the misorientations, then convert them into HSV format, where H is alpha, S is beta, and V is 1.

[hakonanes]

We don't have a solution for this, currently. As @argerlt points out, MTEX has a nice solution.

I belive we can achieve what MTEX is doing by updating the calculation of the polar angle inside the IPF when coloring: https://github.com/pyxem/orix/blob/develop/orix/plot/direction_color_keys/_util.py#L80-L94. We should be able to say that the maximum polar angle is something else than 90 degrees, so that colors are saturated at a lower angle, say, 7.5 degrees as in the MTEX example.

Feel free to submit a PR!

[maclariz]

I had an idea that you could calculate misorientations from some specific direction and then decompose those into rotations about two axes and represent that somehow. All a bit hazy and I'd have to play with the maths on some real data and then once I get something I am happy with, then colormap it. The last step I am pretty comfortable with and am doing for my own projects using things liks hsv_to_rgb

[argerlt]

I believe we can achieve what MTEX is doing by updating the calculation of the polar angle inside the IPF when coloring

Oh, you're right, I was making this harder than it needs to be. Yes, we can do that, just not maybe any time soon.

@maclariz , if you get your solution or something like what @hakonanes is suggesting above working, I'll happily write some tests and help get it implemented.