Hello,
For an upcoming event I was asked to make the way a bee views the world that we see.
This includes:
The color adjustment is a post-processing volume on the camera so that part is fine. The question i am stuck on however, is how I could make the vision faceted. I have to make this work in VR (Ignore the motion sickness implications of this…)
That’s tough…
Like if you wanted each facet to be independent… that might be really costly on the hardware rendering each facet independent of one another.
I’d suggest maybe faking it… set up a projection matrix so that the view frustum sees a huge FOV. Render this to a RenderTexture, overlay a grid on this (a hex grid?) And then do some post processing on the edges of the grid to give a slight bulbing effect to each node (facet). Display all of that to screen.
What does a bee see with each facet, though? I’m pretty sure that insects have compound eyes where each facet is not just a facet on a lens, but rather a complete eye in-and-of-itself, with separate optic nerves and all, but how well does each of these eyes work? Does each sub-eye see it’s own complete picture or does each sub-eye just see a color? How sharp is it’s vision compared to a person?
The classic “bug’s eye” demonstration involves showing a bunch of slightly different images side-by-side to represent all the different facets of their vision, but this has always struck me as fairly ridiculous. If someone asked you to represent a “human’s eye” view, would you use 2 side-by-side images to represent the 2 human eyes? That doesn’t remotely convey the actual subjective experience of looking through human eyes.
If I wanted to try to give a cyclops an impression of what binocular vision was like, I’d try to encode depth information into another channel (e.g. color) so that you can “see” it with only one eye, and then try to work out some way of faking the effect where you can “see through” an object if it’s close enough and thin enough (like how holding one finger vertically in front of your face doesn’t block your view of the scene behind it).
I don’t know what the actual advantages of bug eyes are, though, so I’m not sure what would be analogous.
While I agree with the ridiculousness of the “bug eye” view…
I believe OP isn’t here to debate the rationale of the visual design they were requested to accomplish. They just have to accomplish it.
Then again though, I may be wrong… maybe OP has more to include about requirements?
AFAIK we don’t really have any idea how the compound eye is processed by the brain so have fun with that.
I’d clarify what they expect it to look like and get them to provide some better requirements or offer some acceptable solution options of what it might look like. You’re not a bug scientist (are you? lol).
What my boss asked me to make it look like, is the vision of compound eyes. As mentioned we have no idea how the brain processes it, but for the project all thats important is the visual aspect. At least thats what was asked.
This assignment doesn’t need to show it the way the insect sees it in their brain after processing but rather before processing.
I did find some hexagonized shape overlays for cameras but I didn’t know how to make the hexagon effect on postprocessing for a camera. Using a render texture could work to use a single camera image and change it, got no exp with that tho.
Get some definition of exactly what a compound eye should see. Should each facet see its own image? Or should each facet only see a single colour? The second is essentially a fancy way of pixelating an image.
As nobody really knows how a bee actually sees, I would go for an “artistic” solution.
Maybe a variation of this asset combined with fisheye/distortion would do the job, I don’t know how it would work in RV.
Please, share screenshots if possible. 
According to the research paper I received from the University that asked to make it, bee vision would be a different color spectrum, this spectrum being yellow → ultraviolet (this means they do not see red) The compound eyes that I have to create will have to be a wide FOV with a distorted image (Fisheye lens as stated above) and would look hopefully similar to this. I would buy this pack and do that way, but the creator is not responding and spending money on something that “might” work is generally a bad idea.
An update to what I have (Camera with postprocessing to filter out red and give more depth to purple since ultraviolet is something I dont really know how to represent):
A problem I am worried about is not the motion sickness per se, but the fact that I have a hexagon pattern and you won’t even be able to see anything in the end if I distort it and give it a pattern. I have never done this kind of processing so I dont even know how to overlay this and distort the image.
I am aware of how postprocessing works but my knowledge of shaders is very lacking to produce the result. Could anyone point me in a direction to distort the image and add a hexagon pattern (similar to the image I posted above)?
I would try to render the scene to a texture, and map that to a plane which is made out of a lot of (disconnected) hexes.
Then i would try to make a shader which colors each of the vertices of those hexes according to the texture.
Could do color correction in that shader as well.
This would not result in distortion (which i wouldn’t want), but a sort of pixelated image.
Perhaps a silly approach if you’re skilled in postprocessing effects, but that’s what came to mind.
Here’s a good tutorial to help you get started with regards to rendering hexagons in a shader:
https://www.youtube.com/watch?v=VmrIDyYiJBA
Perhaps you could slightly distort the image with a fisheye or zoom-in effect around the center of each hexagon cell?
