freakin awesome telescope lens distortion
Mazyar told me to make my frame messier which was some of the best news I have ever heard. I LOVE lens effects. Glass and its distortions are beautiful in their own way and vintage lenses are amazing. I’ve got the chance with this project to research what distortion telescopes have, and its a whole other world, and they love to make their charts for it. This one below is beautiful. I got really into astrophotography for the past few years, so here I get to show off some of my knowledge.
Mazyar was absolutely right, and all of these together make it feel like a “real” fisheye. More than any lens distort node could do! Critique from artists and growing from that feedback is my favorite.
There is coma, which occurs closer to the edges of the frame, and gives stars all sorts of oblongs shapes. it can be useful to tell you about sensor tilt, but makes your stars ugly. There is Atmospheric dispersion, which is where heat from the atmosphere makes objects through a telescope, specifically bright stars and planets, have their own directional chromatic abberation. This is from the same effect that makes stars twinkle when you look up.
I’m accomplishing some of these effects with different methods;
For the coma, I use a directional zoom blur, and then radial blur the brightest parts of that. I used two radial blurs, one in each direction (-2, 2) and I have a theory that if I just did 4 and then rotated the whole image -2 it would have the same effect and save on a heavy blur node. Maybe, maybe not.
The atmospheric dispersion is a chromatic aberration affect whose direction is determined by a camera shake, which is on the red shuffle and then mirrored on the blue shuffle. Green stays still.
Then we have a slight vignette with a grade for .9 gamma and a radial mask, and a defocus closer to the edges with a blur using the same radial mask.
After that image I also added some stationary lens dirt with a mask from the internet which is inverted and overlayed onto the end result, behind every single effect right between merge10 and merge13. It is also slightly defocused.
and the result! look at that delicious procedural coma right at the edges.
This script was getting quite heavy, so I would like to look into properly using precomp’s - as well as precomp tools that automatically “condense” finalized or near-finalized nodes into rendered EXR’s.
I also think one last touch I might need to do is adjust the noise of the plates, as you can see above the BG is pretty noisy/low-res and makes the cashew ring stand out as CG. I will have to re-grain the CG or denoise the BG.
Another element of this week;
https://docs.google.com/document/d/14SLS-RPBxG5qlfqRiXAr4lDNiDKUVmnQ4N3YdhzkJWs/edit?usp=sharing
I’ve started a “VIRTUAL PRODUCTION CAMERA TRACKING DATA HOW-TO & NOTES” for motion tracking data coming off of SCAD’s XR stage specifically, and the ways it can be utilized.
I did some test and I hope to get visuals here soon, but it looks like the tracking fidelity is just a bit too low to be used for a perfect key when there is camera movement. However, it is still useful maybe for when the motion isn’t needed to be as accurate. It also is a neat write up for how to import FBX animation into any program anyways.
