Contour plot -- Greyscale gradient seems wrong

Summary -- Greyscale gradient is significantly different than grayscale banded and is likely wrong.

The attached model (H10a.liml) is a modal analysis of a 125 mm diameter cylinder (1/4 section). For mode 4 the following two images show the Displacement_z of the top surface (Horn_face) using a greyscale banded contour plot with 20 divisions. Note that the white area (highest displacement) extends along approximately 1/4 of the radial distance.






The average Displacement_z of the Horn_face can be calculated as follows. "\Solution\Surface integral" with F(X,Y,Z) = Displacement_z[m] and D = Horn_face gives 2018.



The area of the Horn_face is 3068 mm^2. Thus, the average Displacement_z is 2018/3068 = 0.658.

For confirmation, the average Displacement_z can also be determined by looking at the histogram of the banded greyscale (second image above).
  1. Load the image into GIMP.
  2. "\Select\Select by color" and click on the green.
  3. "\Select\Invert" to instead select the banded grayscale.
  4. "\Colors\Info\Histogram". The mean value is 0.657 (essentially the same as the above calculated value).




The following two images show the Displacement_z of the Horn_face but using a greyscale gradient contour plot. Note that the white area now extends along approximately half of the radial distance.







Then, using the above method for this contour plot, the mean histogram value is 0.851 -- too high compared to either the surface integral calculation (0.658) or the banded histogram (0.657). Thus, both visually and by histogram, the greyscale gradient seems wrong.



Comments

  • Oh, that's certainly wrong. It looks like this bug affects the other color scales too but it's most prominent with greyscale. I'll fix it for the release after next. Thanks for finding and investigating it so thoroughly.
  • edited October 2022
    I guess if the reason is not the lightning effect more than the interpolation process ?¿?.
    For two homogeneous spheres with uniform solution the final tone is different too.




  • edited October 2022
    This bug is a separate effect from the lighting since it appears with a flat surface where the lighting is uniform. It's caused by a different separate light source for the gradient plot, I think ambient or generated from the surface color itself. Not sure exactly.
  • The gradient, "second image" seems to have a more volumetric feeling probably by an additional external light.
    ¿Shouldn’t the lighting effects be reserved for the modeling window?
  • Directional lighting's still useful for the solution even though it interferes with the colors. Without it, 3D shapes are very confusing - they just look like featureless 2D projections.

    The other light sources are there to reduce the extremes between lit and shaded areas, so that helps keep the colors more correct - when it's done right!
  • You always learn something new :)
    Thanks
Sign In or Register to comment.

Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!