The slider with the range [-1170,900] is the slider that controls the range of the window level slider. There is 1 level of abstraction.

What values for such range would you suggest instead ?

Does it make sense to have the window/level range that exceeds the intensity range of the input?

Yes if you want a ramp that is not pitch black for your lowest scalar value (or that is not bright white for your highest scalar value).

In my opinion, it's very confusing. If you want to control low/high color, why not just add the color popup selectors for the extremes? What is the meaning of the ranges [-1170, image_min] and [image_max,900]?

In any case, Ron thought this is a bug. I will forward this bug report to Ron, and if he is convinced by your argument, I have no problem is this issue is closed.

I agree with you, it is confusing, I(we) just didn't think of a better mechanism.
One thing to keep in mind, is that for a scalar range of [0, 279], it should be possible to have a window that ranges from [-100, 1000] because we want to associate 64 to the scalars 0, 66 to the scalars 1, ... and 128 to the scalars 279
<pre>
255| ____
...| /|
128| ... |
| / |
| / | |
| /| | |
0| / | | |
*-----|<-->|------|------------
-100 0 279... 1000
</pre>

Ron told me he read the comments, and he does not understand your explanation. He says it should be [min,max] for the range.

I think we are thinking of the functionality of this slider in different ways, and your explanation does not come across.

I think what you are trying to say (if I understand you correctly) could be accomplished in a different (IMO, more intuitive) way. You could have one double-range slider for color map LUT (min=0, max=255). In your example above, it would be set to [66,128]. Then you would have a second double-range slider for the range of the image intensities (in our example, [0,279]), or a sub-range of the intensity range, if the original intensity range is very large -- this would be the initial value for the popup slider in question here . Then the second double-range slider would be the actual window/level for mapping image scalar values onto the LUT colors. min/max of this second double-slider would be initialized to the popup slider selected range.

If this is still confusing, why don't you just discuss this issue with Ron at the next opportunity.

This seems like a valid solution, however I can see 2 drawbacks:

• it requires to tweak 2 sliders instead of one (remember, the slider in the popup should rarely be used )
• it is a different feel from the "window/level" radiologists are used to.

Maybe it's a better solution than the current one, I just wanted to raise those limitations.

Have you ever seen a radiologist using a slider to control w/l??? In my experience, "radiologist" and "slider" do not belong to the same sentence.

But jokes aside, I think the main issue is to make the interface intuitive and easy to understand. Reducing the number of controls is secondary. Maybe you can just ask around what other users think about the existing approach. I am confused as it is now. But I am not very smart, so it may be just me.

@andriy: Which approach radiologist are currently using ? Couldn't e just inspire from this ?

Was the diagram reported below by Julien helpful to understand ? May we could refine the widget to propose a similar chart that could be interactive. Instead of having sliders ?

@jc: radiologists I work with use mouse to control w/l.

The primary customer for the functionality we are discussing is Ron. He is using it for some modes of volume rendering. So I think the best would be for Julien to discuss this with Ron to make sure he is happy with whatever changes are proposed (or if it is decided no changes are needed).