I'm interested in the new gravitational lensing model used in Space Engine. I know it's based on the Kerr metric and the fact that it can be solved in real time still amazes to me. However, I have noticed some slight differences with the lensing effect produced by rapidly spinning black holes in Space Engine compared with the simulations produced by Double Negative for the movie Interstellar. Specifically, I have noticed that in their simulation, the Einstein Radius (the kind of lensing ring around the black hole) is more lopsided than what we see in Space Engine. Both simulations produce lopsided looking black holes but theirs appears to be more lopsided (both the black hole and the EInstein Radius). I have tried in Space Engine to produce a similar looking lensing effect as seen in Double Negative's video here:
https://cfn-live-content-bucket-iop-org ... DWEfeYA%3D but so far I have returned empty handed. I'm curious as to if there are any theories as to what the discrepancy is if there is any at all (could just be how Double Negative is visualizing their black hole since some of their animations show an equirectangular projection). If anyone disagrees that there is a difference between the two, feel free as I am only judging this by eye.
The lopsidedness depends on the spin of each individual black hole. If you select a black hole and edit it via the planet editor (or make a new one from scratch using SE's scripting), you can set the spin to a custom value. The black hole in the link looks square because of the projection, as you mentioned. Double Negative's test render also does not distort the stars properly, as each star is represented as a 0D dot rather than a 2D distortable graphic - it is only a test render and definitely does not use the same techniques as the final render in Interstellar (in which a star directly behind Gargantua would become a ring, unlike the test black hole)
SpaceEngine's renderer is based on the actual metric. Some accuracy is lost by default because the shaders are otherwise taxing on the GPU, but you can increase the accuracy in the settings using "gravitational lensing precision":
The lag is a fair bit more noticeable than the inaccuracies introduced to remedy it, though. Volumetric objects resolution will affect how blurry accretion discs and nebulae are.