Light divergence per wavelength

[aomanchuria]

I'm working on a simple 1 or two strong lens plus light source combination and noticed a few things with beam divergence.

The source is composed of 16 x 1.5unit squares that emit on the same plane with a 60degree half angle divergence. I modeled the lens and found that the image produced in real life would require the lens to have an OIR of 1.34 or have a Z scale of the lens at ~.78 . I literally just placed a light, a lens and a screen both on my desk and blender and calibrated th image size on blender by adjusting the IOR. 1.34 doesn't correspond to any glass but gets close to quartz maybe.

My light source is a UV LED at 385nm and the lens is definetly not quartz but just regular sodalime glass. The beam in blender was white using power at 100W with default 17L/unit.

So that sparked my curiosity... Could I just use a blue or purple color to get the right divergence or image size? Nope. I change the light color and the image stays the same size.

In reality, high frequency light like UV should bend more thus producing a small image while Ir should produced a larger image. It would be great if we could adjust beam diffraction per unit color/wavelength. And if the color/wavelength actually made a difference in image size.

[aomanchuria]

I guess also it would be cool to make a monochromatic beam that one could set the wavelength, divergence, power, M^2, and power distribution of. Power distribution for example for a square emitter could be Gaussian and using false color one could look at the distribution on a screen.

[CodeFHD]

Hi,
LuxCore is currently not a full spectral rendering engine, which is one thing that would be needed to get a result like you describe. There is a dispersion feature for glass, but it does not take enough coefficients of dispersion to model complex systems well, and it has other implications as described in this issue: #262

Silica Glass has a refractive index of around 1.47 at 385nm (https://www.corning.com/media/worldwide/csm/documents/HPFS_Product_Brochure_All_Grades_2015_07_21.pdf). Soda lime would be higher. So unrelated to the spectrum you might want to check the model of your lens. You can use my other addon OptiCore to help you visualize ray paths, it has a feature to display ray fans traced separately through the lens. The addon is specifically created to help model lenses according to optics design specification. https://github.com/CodeFHD/OptiCore

About your second message, you can already use a spot light to have a source with controlled divergence. I agree that other types of light sources would be interesting (I could use this myself, like 2D-gaussian beam profiles), which is a matter of manpower to actually write this code.
M^2 is going to be more of an issue. You already (but incorrectly) used the term "diffraction" (dispersion is what you were talking about before). LuxCore does geometric ray tracing, diffraction is a while other technique and problem.

[aomanchuria]

Here's what a friend of mine is doing, (who is totally not me at all) https://www.photrio.com/forum/threads/projector-cyanotype-and-chiba-python-digital-negative.209082/page-4

I guess you can close my report. I'll ...my friend'll just change the emitters to be spotlights so the angle and distribution can be adjusted/calibrated. In the experiment the goal is to have a light source that behaves like the real thing just to adjust the distances until the image is blurred enough. That would give me precise measurements to update and print new parts for the experiment. The software is great by the way and I've started using opticore. The lens in this case is something I modeled in UGNX. I took a photo of the lens with my phone and then copied the curvature in NX. I have no other means to actually measure the lenses, specially if its a concave lens. Similarly for the IOR of the lens, the only way is to make a beam and measure it at several spots and try to make the blender version do the same. It was just then when I realized that I don't know what is best to do...change the IOR of the glass to match the real beam or change the geometry to match the real beam.