Using Gamutvision: Part 2: Displays

Selections available in the display selection box on the right are shown on the far right, with links to descriptions.

The Display options area, shown on the right, is different for each display. These areas are described along with several of the displays, below.

1. 3D CIELAB L*a*b*: wire input, solid output
2. 3D CIELAB L*a*b*: solid input, wire output
       (often with some transparency)

• The most detailed graphic display of input and output gamuts. Derived from a pattern that contains all values of HSL H and L with S = 1 (maximum saturation). Optionally displays the gamut volume— the single number that best characterizes the gamut.
• Can be zoomed or rotated to any viewing angle.
• Can be displayed with input, output reversed.

Display options

• Reset scale resets the plot scale for the current plot: it is enlarged whenever large gamut color spaces are displayed.
• Reset view resets the original viewing angle.
• Top view sets top view. Clicking it again restores the previous view.
• Light turns on lighting to highlight the gamut surface texture. The light position rotates when the diagram is rotated. It is reset by pressing View.
• The popup menu to the right of Light sets the wireframe resolution and gamut surface display. Low res wireframe/smooth is the preferred default.
• Transparency (slider) sets the transparency of the wireframe surface. Can be used to emphasize cases where the solid (usually inner) surface protrudes outside the wireframe surface.
• Rotate turns on auto rotation. This allows the gamut surface be examined thoroughly. The image continues to rotate until the next time you press Rotate. It’s best to turn off rotation before switching to other views, otherwise the image may be misaligned when you return to the 3D L*a*b* view: you may need to press Reset view then Reset scale to restore it.. The light rotates along with the image. To change the Light angle, stop the rotation, then press the appropriate View button.
• Vectors that show the transformation from input to output colors appear when the slider is set between 0.02 and 0.98 (representing the HSL Lightness of the input image).
• Saturation sets the HSL Saturation of the input image. It is highlighted (text in bold burgundy, slider in pink) when set to values other than its default of 1.

• More opens a dialog box with additional display options for the 3D L*a*b* plot.
  • 3D Plot Axis Setting controls the geometry of 3D plots. Affects plot size and aspect ratio. Settings are described here.
  • Background controls the background grayscale color (0 (black) to 1 (white). Set it darker for more dramatic contrast; white for some publications.
  • Light Left/Right sets the direction of the lighting.
  • Wire frame sets the wire frame color/lightness.
  • Wire frame transparency color  (affects only image analysis)
  • Display Gamut volumes  causes input and output gamut volumes (in L*a*b* volume units (ΔE³)) to be displayed in the lower right of the main gamut display area. It is optional because its calculation slows the display refresh.
  • OpenGL renderer is normally checked (turned on), but it can sometimes cause program instabilities. Turning it off makes wireframe display black or transparent (the transparency slider has no effect). Less of an issue in 2022 MATLAB.
  • Reset restores the default values (vis3d, etc.).

3. 3D HL  or h*L* Color difference plots

Plots color differences using H and L (from HSL) or h* and L* (from CIELAB L*a*b* transformed to L*c*h*) as the independent axis. S = 1 (the gamut boundary) is the default, but S can be set to values less than 1 to display color differences inside the gamut boundary. The image on the right is the 3D ΔE*₀₀ color difference for S = 1 displayed on the HL (from HSL) plane.

Display options

• Light turns on lighting to highlight the color difference surface texture. The light rotates when the diagram is rotated.
• The popup menu to the right of Light selects the color difference metric. Choices are the same as the 2D a*b* color difference, above.
• Rotate turns on auto rotation. This allows the color difference surface be conveniently examined. The image continues to rotate until the next time you press Rotate. It’s best to turn off rotation before switching to other views, otherwise the image may be misaligned when you return.
• Image lightness (slider) sets the displayed image lightness (results are not affected). Darker (< 0.3) is best for 3D plots; lighter (0.3 – 0.7) is best for 2D plots to keep contours visible.
• Invert (3D plots only) inverts the display. Useful for Chroma difference, which is mostly negative.
• The popup menu to the right of Invert selects the display type.
  3D HL(HSL) independent axis   (right)
  3D h*L* (L*c*h*) independent axis  (below, right)
  The two displays contain the same information, presented differently.
• Saturation (slider) selects the pattern saturation (S in HSL color representation). The default value of 1 displays the gamut boundary. Smaller values (<1) display the interior of the gamut volume. This feature makes is possible to view rare profile defects that appear inside the gamut volume but not on the surface. 

The h*L* (from L*a*b*) color difference is shown on the right, below the settings window.

As expected the HL color difference for the Epson Hot Press matte paper (similar to other matte surfaces), shown on the left, displays a larger difference for the dark areas (low L*= values) than the Luster because it can’t reproduce the deepest black tones (with density > 1.5) Click on the image to enlarge it.

4. 2D a*b* Gamut map (S = 1)

Plotted on the CIELAB a*b* plane for L(HSL) = {0.1, 0.3, 0.5, 0.7, 0.9} with maximum saturation (S = 1). The petal-like dotted concentric curves are the twelve loci of constant hue, representing the six primary hues (R, Y, G, C, B, M) and the six hues halfway between them. They follow different curves above and below L = 0.5.

Display options are shared with the 2D L*a*b* saturation map (below).

• Reset scale resets the plot scale for the current plot: it is enlarged whenever large gamut color spaces are displayed.
• The popup menu controls the display. Markers (circles indicating 12 evenly spaced colors, including R, Y, G, C, B, and M) can be switched on an off. The 2D a*b* Saturation map (below) can be displayed with light lines on a dark background.

• Vectors connecting the with the output colors appear when the Vectors slider is between 0.02 and 0.98 (representing the input L(HSL) for the 2D Gamut map (above) or input S(HSL) for the 2D Saturation map (below)).

5. 2D a*b* Saturation map

Plotted on the CIELAB a*b* plane for saturation S(HSL) = {0, 0.2, 0.4, 0.6, 0.8, 1.0}.

The default value of L(HSL) = 0.5 is the middle lightness level where hues reach their greatest saturation, e.g., pure red [255,0,0], green [0,255,0], and blue [0,0,255] all have L(HSL) = 0.5. L(HSL) may be varied between 0.05 and 0.95 using the HSL Lightness (L) slider at the bottom of the display options area. It can be reset to its default value of 0.5 by pressing the Reset L=0.5 button.

The 2D a*b* saturation map and color difference map (below) are the best displays for visualizing printer gamut as a function of Saturation S: At low S-values the output (solid lines) tracks the input (dotted lines), but the output clips strongly when S reaches its limits: around 0.4 in the green and blue-magenta regions. This plot is especially valuable for comparing rendering intents, which don’t behave as the textbooks indicate.

Note that the value of S at clipping depends on the input color space: the larger the color space gamut, the more saturated the colors corresponding to S, i.e., the larger the chrominance (c* = (a*² + b*² )^1/2 ).

Despite its evident limitations, the Epson R2400 is an excellent pigment-based printer.

6. 2D a*b* color difference

Similar to the above 2D a*b* saturation map except that it displays color differences.

The Display options are the HSL Lightness (L) slider, Reset scale, Reset L=0.5, and a popup menu for selecting the Color difference metric. Choices include ΔE*_ab, ΔC*_ab, ΔE*₉₄, ΔC*₉₄, ΔE*_CMC, ΔC*_CMC, ΔE₀₀, ΔC₀₀, ΔL*, ΔChroma, ΔHue angle, and Δ|Hue distance|, L* (input), L* (output), chroma (input), and Chroma (output). (The last four are not actually differences.) Color difference metrics are described in the Colorcheck/Color/Tone Appendix. 

7. 2D CIE 1931 xy saturation map

Plotted on the xy plane for L(HSL) = 0.5 with saturation S = {0, 0.2, 0.4, 0.6, 0.8, 1.0}. A D50 (5000K) white point is implied.

The CIE 1931 chromaticity diagram is explained in the Wikipedia and in EFG’s Chromaticity Diagrams Lab Report.

Display options are shared with the u’v’ saturation map (below).

• The popup menu allows you to select several view options.
  Lightened view  |  Normal (saturated) view
  Lightened/gamut boundaries  |  Normal/gamut boundaries
  

Lightened view (shown on the right) enhances the visibility gamut lines.

Normal view (shown in the u’v’ diagram, below) is about as close as it’s possible to get to the actual colors of the chromaticity diagrams on a monitor, which has a gamut comparable to sRGB.

Gamut boundaries means display only gamut boundaries to reduce clutter. Otherwise, display gamuts for saturation S = {0, 0.2, 0.4, 0.6, 0.8, 1.0}.

• Wavelth: selects spectrum locus wavelength display. The wavelengths in nanometers can be displayed (u’v’ plot, on the right) or omitted (xy plot, above).

8. 2D CIE u’v’ saturation map

The CIE u’v’ plane is a transformation of the xy plane with improved perceptual uniformity. It is less familiar than the CIE 1931 xy plot, but still widely used. A D50 (5000K) white point is used.

9. Black & White density response:

These curves plot the grayscale output vs. input. The fourth plot also plots the CIELAB L* values of primary colors. These are the best displays for observing the effects of Black point compensation.

Display options contains a single popup menu with four selections.

• Density vs. Log pixels (log-log scale);  Shown on the right, above. “Density” is Output density. Input density is not referenced. The average slope of the Output density curve (slightly greater than 2) is the gamma of the output device. This plot shows the effects of Black point compensation. When turned off (with Colorimetric rendering intent) the diagonal like will be linear until Density reaches 2.08, then flat.
• Output vs. input density (log-log scale);  similar to the traditional characteristic curve for film (the Hurter & Driffield D-Log E curve).
• Output luminance vs. input luminance (linear scale)
• CIELAB Output L*, a*, b*, and c* (chroma) vs. input L*.   Shown on the right, below. Chroma = c* = (a² + b²)^1/2 is the deviation from neutral gray: a valuable measure of B&W printing performance. The Saturation slider is active for this display.
  

  This plot displays several results.
  –	Output vs. input L* for the gray region is diaplayed as a black line with dots (•) that normally looks like a thick black line.
  –	a*, b*, and chroma (a2+b2)1/2 of the gray region are displayed as dotted blue, red, and green lines (multiplied by a factor of 10). These curves illustrate the deviation from neutral gray. They would be close to 0 for an ideal profile.
  –	The Output vs. input L* response for the primary colors (R, Y, G, C, B, and M) is shown as solid lines. The saturation for these colors (S(HSL)) is set by the Saturation slider. This display is used to advantage in Profile mystery: The case of the smudged pines.

10. Image color analysis display

This display, opened by clicking Read image for analysis in the Display selection box, shows the perceptual color difference of an image before and after gamut mapping. It can be used to preview how images change when printed. The pseudo color display, which gives the color difference in any of several perceptual color metrics, contains far more information than the gamut warnings of image editors, which merely tell you that a color is outside the gamut of the output device, but give no indication of how much change to expect. This example uses a photograph taken in Paris by Virginia Bonesteel.

Read image for analysis must be pressed for each image to be analyzed.

• The color map is selected in the lower popup menu. The map shown, WYRMBK (White-Yellow-Red-Magenta-Blue-Black) goes from white to black through a range of light to dark colors. It offers an intuitive light-to-dark progression and good value discrimination. Other color maps have different properties.
• Probe turns on the probe. When the probe box is checked, crosshairs appear in the Gamutvision window, and the error metrics shown below are displayed beneath the image for the location where the left mouse button is clicked. You can also click on the preview image on the upper-right. Input and output colors are shown to the left of the metric:, input on the top and output on the bottom. The probe is turned off whenever you click outside the images.

  The probe displays

  • Input and output L*a*b* values
  • ΔE*_ab, ΔC*_ab, ΔE*₉₄, ΔC*₉₄, ΔE₀₀, and ΔC₀₀ (The CIEDE2000 metrics replaced ΔE*_CMC and ΔC*_CMC in Gamutvision 1.3.7.)
  • ΔL*, ΔChroma (ΔC*), Δ|Hue distance| (ΔH*), and ΔHue angle (degrees)

• The Invert checkbox inverts the direction of the color map. It is useful for metrics like ΔChroma, which is mostly negative.

11. 3D Vector image color difference display

The four display selections at the bottom of the Color metric popup menu are not, strictly speaking, color differences. They display color changes between input and output images in 3D L*a*b* space.

• Input 3D cluster displays input colors as dots or circles, where larger circles indicate a greater frequency of the color.
• Output 3D cluster displays output colors as dots or circles.
• Input-Output 3D vectors displays vectors connecting the input and output colors as well as output (inner) colors as dots or circles.
• Output-Input 3D vectors displays vectors connecting the input and output colors as well as input (outer) colors as dots or circles. Shown on the right.

A semitransparent wireframe of the output color space (after gamut mapping) is also shown in this display. A great may display options are available.

12. The X-Rite/Calibrite ColorChecker^®

Gamutvision can analyze a simulated ColorChecker, derived from L*a*b* D65 data supplied by X-Rite (transitioning to Calibrite in 2022). This data is transformed to the input color spaces (1 and 3). Display selections are identical to Image color analysis, described above.

ΔE₉₄ pseudocolor plot
Adobe RGB to Epson 2200 Enhanced Matte, Colorimetric intent