[Paraview-developers] [EXTERNAL] Re: Color legend presets

Scott, W Alan wascott at sandia.gov
Wed Dec 20 13:25:59 EST 2017 

Adding the ParaView developers back into this thread.

From: Francesca Samsel [mailto:figs at me.com]
Sent: Wednesday, December 20, 2017 11:21 AM
To: Bujack, Roxana Barbara (LANL) <bujack at lanl.gov>; Scott, W Alan <wascott at sandia.gov>
Cc: Nouanesengsy, Boonthanome (LANL) <boonth at lanl.gov>; aron.helser at kitware.com; Moreland, Kenneth <kmorel at sandia.gov>; Patchett, John M. (LANL) <patchett at lanl.gov>; Ahrens, James Paul (LANL) <ahrens at lanl.gov>; Rogers, David Honegger (LANL) <dhr at lanl.gov>
Subject: Re: [Paraview-developers] [EXTERNAL] Re: Color legend presets

Alan,
We have been working on the colormap issue for over four years now.  It is complex.  I would suggest that the main voices sit down together with KitWare, and documentation from folks who have weighed in.  One option - I will be coming out your way in mid-February for the ParaView tutorial.  Would it make sense to put together an in person / skype meeting and have a discussion about color?  We could meet either in ABQ or at LANL. I agree with the ParaView committee that any changes made need to be well considered and made once as opposed to smaller sequential changes.


My two cents… We need the rainbow, de-sat rainbow, jet and Moreland cool warm. Beyond that we should look for a range of maps that produce accurate results on differing fronts - discrimination, uniformity, specialized use, etc. A shortened list is definitely needed but it needs to be carefully selected to include a variety of maps that best addresses a variety of needs.  It might also be helpful to look at the Colormap UI  to see if it could be reconfigured to categorize colormaps beyond the standards via an advanced option.

Terry Turton and Colin Ware have done a good bit of perceptual testing.  Looking at their findings will help guide the way to providing a relevant range of maps.


Recommendations are another topic to discuss.  We have tutorials up on SciVisColor.org<http://SciVisColor.org> and are working on a new series of videos specifically designed to help scientists quickly and easily select a colormap based on their data and task.  Matplot lib has great documentation.  We are updating our documentation.  Check back mid to late January when we hope to have the updates complete.

On a side note, I presented last week at AGU.  The scientific community is clambering for help on the topic.  Figuring out how we can best provide the guidance and broaden our impact is also worth discussing.

Best to all,
Francesca


On Dec 20, 2017, at 11:49 AM, Bujack, Roxana Barbara <bujack at lanl.gov<mailto:bujack at lanl.gov>> wrote:

I agree that we should shorten/reorganize the list, but also that the rainbow needs to stay there.

Best,
Roxana

On Dec 20, 2017, at 10:42 AM, Boonthanome Nouanesengsy <boonth at lanl.gov<mailto:boonth at lanl.gov>> wrote:

Hey,
Not sure if you're on the Paraview developer's mailling list, but there is a discussion of changing the colormaps options in Paraview. Maybe you have opinions on this.

cheers,
Boonth


-------- Forwarded Message --------
Subject:

Re: [Paraview-developers] [EXTERNAL] Re: Color legend presets

Date:

Tue, 19 Dec 2017 17:50:32 +0000

From:

Scott, W Alan <wascott at sandia.gov><mailto:wascott at sandia.gov>

To:

Aron Helser <aron.helser at kitware.com><mailto:aron.helser at kitware.com>

CC:

paraview-developers at paraview.org<mailto:paraview-developers at paraview.org> <paraview-developers at paraview.org><mailto:paraview-developers at paraview.org>


Just to reiterate here (a thread I have presented to a more limited, but very vocal audience), we NEVER will be able to remove Rainbow. There are cases it is not just the best colormap, but the only color map that fills requirements.  If a customer is trying to compare screenshots or movies against historical runs, and these historical runs were Rainbow, Rainbow is the best color map.  Further, if Rainbow is contractually required, we will lose users if we don’t present Rainbow as an option.

But, having said that, let’s make it easy to do the right thing, and harder to do stupid stuff (such as use Rainbow as default).

Here is a good writeup on the four matplotlib colormaps: https://bids.github.io/colormap/.  Note that they are both perceptually uniform (not necessarily with regards to brightness), and are also very good for color blind folks.

Here’s another good writeup on colormaps (from the good folks at matplotlib):https://matplotlib.org/users/colormaps.html


Alan

From: Aron Helser [mailto:aron.helser at kitware.com]
Sent: Tuesday, December 19, 2017 9:07 AM
To: Scott, W Alan <wascott at sandia.gov><mailto:wascott at sandia.gov>
Cc: paraview-developers at paraview.org<mailto:paraview-developers at paraview.org>
Subject: [EXTERNAL] Re: [Paraview-developers] Color legend presets

Sounds like a great idea to me - the long colormap list gives no guidance to users.
I agree that rainbow and other perceptually troublesome colormaps should be advanced (or removed).
IIRC these are designed to have uniform or linear perceptual brightness? What about the others?

  *   Viridis
  *   Plasma
  *   Inferno
  *   Magma
I think we'd have to decide what to do with categorical colormaps as well.
Regards,
Aron

On Mon, Dec 18, 2017 at 6:37 PM, Scott, W Alan <wascott at sandia.gov<mailto:wascott at sandia.gov>> wrote:
I would like to propose that we reorder the color legends presets as follows:


  *   Add an advanced icon.  Default will get rid of everything other than the first page.  Increase the size of the Choose Preset to show all of the default list.
  *   In Advanced, add sub menus.  These should be Common, Rainbow variant,  double ended, dark to light, misc., etc.
  *   Default will look like this.  Note that for the most part I didn’t invert any maps, since there is already a button to do this.

     *   Cool to Warm
     *   Cool to Warm (Extended)
     *   Viridis
     *   Plasma
     *   Inferno
     *   Magma
     *   Black-Body Radiation
     *   X-Ray
     *   Gray Scale
     *   Cold and Hot
     *   Black, blue and white
     *   Black, orange and white
     *   2hot
     *   Rainbow Desaturated
     *   erdc_rainbow_bright
     *   erdc_rainbow_dark
     *   Jet
     *

Get rid of Rainbow Blended White, Gray and Black.  These were things I tried, and didn’t work, with our lighting model.
Get rid of the second erdc_rainbow_bright/dark
Get rid of coolwarm.  It is Cool to Warm, which we already have.
Notice that I pushed normal Rainbow to be advanced only.

Thoughts?

--------------------------------------------------------
W. Alan Scott
ParaView Support Manager

SAIC
Sandia National Laboratories, MS 0807
Org 9326 - Building 880 A1-K
(505) 284-0932<tel:%28505%29%20284-0932>   FAX (505) 284-5619<tel:%28505%29%20284-5619>

The most exciting phrase to hear in science
is not "Eureka!" but "That's funny..."  -- Isaac Asimov
---------------------------------------------------------


_______________________________________________
Powered by www.kitware.com<http://www.kitware.com/>

Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html

Search the list archives at: http://markmail.org/search/?q=Paraview-developers

Follow this link to subscribe/unsubscribe:
https://paraview.org/mailman/listinfo/paraview-developers

<Attached Message Part.txt>


Francesca Samsel M.F.A.
Research Associate
Center for Agile Technology
University of Texas at Austin
figs at cat.utexas.edu<mailto:figs at cat.utexas.edu>
512-232-5487

www.francescasamsel.com<http://www.francescasamsel.com>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://paraview.org/pipermail/paraview-developers/attachments/20171220/c4f14e23/attachment-0001.html>

More information about the Paraview-developers mailing list