We have an image in which the voxel values are the statistical effect sizes ranging from about -0.3 to 0.9. Most of the image is 0s (voxels that didn’t reach significance). I would like to render this image in a way that negative values are a blue to green ramp (-1, 0), positive values green to red (0, 1), and values of 0 are not shown at all.
I am trying creating color tables but have no luck. I appreciate if someone can guide me.
attached is a sample nRRD file: Box
To give a visual example, this is with my modified color table which supposedly have symmetric values from -2 to 2 with 256 colors.
For whatever reason, the scale bar is displaying values only from -0.1 to 0.3 even though the max and min values for this volume is clearly reported to be -0.3 to 0.9 range
I made progress up to a point. Min/Max window level adjustment is really not intuitiuve to work with. So the remaining issue, how can I supress 0 values to be shown in slide views. Since I cannot enter a split range into the threshold value, threshold setting is not very useful.
In volume rendering I can hack something with manually adjusting things. But I need to be able to do this more consistently with more control over colors, and particularly in 2D as well.
Please provide pointers/suggestions.
Any suggestions to hide 0 values (green background)? @pieper
I tried setting 0 to NaN, but then in slice views it switch to color of minimum value in the dataset (even though the data probe shows it as NaN)
To specify a color transfer function I would recommend using a procedural color node (where you can define any color for any absolute value) instead of a color table (where you specify color values at regular intervals that you later map to some intensity range). See full example in the script repository.
Any suggestions to hide 0 values (green background)?
Display threshold can be used for hiding background.
my values range from negative to positive, zero being in the middle not at the lowest value. How can I just threshold 0, or a range like (-0.0001, 0.00001) ?
There are values close 0 on either side, and they are not discernible if 0 is rendered in the overlay
For example this is the full range
negative only (-0.3 to -0.01)
positively only (0.01 to 0.9)
I want the last two together in the same image.
Here are two solutions, by modifying an existing discrete color table to be a transparent in the center or a continous mapping function to be black in the center.
Colors → DivergingBlueRed (a color function that that uses just 3 colors: blue = lowCopy button (to create a modifiable copy), accept the proposed default name DivergingBlueRedCopyopacity → 0 (to make the center of the color map transparent, to avoid adding a white haze for values around 0 in the overlay)Active volume → Exp-02...Lookup Table → DivergingBlueRedCopyL → 0 (to make the center at 0)Exp-02... volume as foreground, set its opacity to around 0.5final volume as backgroundIf the abrupt change from transparent to opaque is not desirable then set more colors around the center semitransparent by decreasing their opacity values. It may be easier to do this by editing the color table in a text editor or by Python scripting.
Opacity cannot be specified for continuous color mapping, so we will use black as center color and use additive image blending.
Colors → RedGreenBlue (a color function that that uses just 3 colors: red = lowCopy button (to create a modifiable copy), accept the proposed default name RedGreenBlueCopyPoint → 1 above the color function editor)Active volume → Exp-02...Lookup Table → RedGreenBlueCopyL → 0 (to make the center at 0)Exp-02... volume as foreground, set its opacity to 1final volume as backgroundBlend mode icon and choose AddCreating the color table is somewhat tedious, so it could make sense to add a transparent/black centered color table to the presets.
