Uploaded February 2020 | Updated September 2026, 3 weeks ago
A network of Penrose triangles in Nil.
The non-VR version: youtu.be/lWbrHAYsQk4
See also: youtu.be/YmFDd49WsrY (for some links about Nil)
Stereoscopic videos usually simulate our binocular vision (they are shot from two points which correspond to our eyes). This method does not work in nonisotropic geometries such as Nil, because our visual systems cannot interpret this (in Euclidean space the images in both eyes are always at the same height; in nonisotropic spaces, they are likely to be at different heights and we cannot interpret that).
Instead, we use azimuthal equidistant projection (i.e. compute direction and distance and render every object to a point in the Euclidean space in that direction and distance) and then render that using standard Euclidean stereoscopic vision. Actually, we are not using Euclidean stereoscopic, but ODS projection -- this lets us create VR videos with minor artifacts (mostly when looking upwards/downwards).
The Euclidean binocular vision would not work correctly in Nil.
This is not optimized (and took lots of time to render). Older version: youtu.be/gW8P8C4heaw
A network of Penrose triangles in Nil.
The non-VR version: youtu.be/lWbrHAYsQk4
See also: youtu.be/YmFDd49WsrY (for some links about Nil)
Stereoscopic videos usually simulate our binocular vision (they are shot from two points which correspond to our eyes). This method does not work in nonisotropic geometries such as Nil, because our visual systems cannot interpret this (in Euclidean space the images in both eyes are always at the same height; in nonisotropic spaces, they are likely to be at different heights and we cannot interpret that).
Instead, we use azimuthal equidistant projection (i.e. compute direction and distance and render every object to a point in the Euclidean space in that direction and distance) and then render that using standard Euclidean stereoscopic vision. Actually, we are not using Euclidean stereoscopic, but ODS projection -- this lets us create VR videos with minor artifacts (mostly when looking upwards/downwards).
The Euclidean binocular vision would not work correctly in Nil.
This is not optimized (and took lots of time to render). Older version: youtu.be/gW8P8C4heaw


![Can we simulate spherical geometry in Euclidean space?
Three-dimensional spherical space can be created from 120 spherical dodecahedra. Four of these are filled. In the first part of the video, we see the effects mentioned in https://youtu.be/leuleS9SpiA
Can we simulate these effects using an Euclidean game engine and portals? The remaining two parts of the video show that this does not work, but they should still be fun!
In the second part, we construct the same scene from 120 Euclidean dodecahedra. Some of the spherical effects can be seen if you look close enough. This looks like some cool abstract art, but it does not work very well as a simulation of 𝕊³.
The edges look strange because we have only 349.695° of space around them.
In the third part, we follow the suggestion of Jos Leys [ http://www.josleys.com/article_show.php?id=83 ] we attempt to construct the scene from the stereographic images of spherical dodecahedra.
This concentrates the curvature on faces instead of edges [ http://geometrygames.org/HyperbolicBlanket/ ].
Again, this looks more like abstract art than 𝕊³. Can we simulate spherical geometry in Euclidean space?](https://i.ytimg.com/vi/XUIYga-AfLI/mqdefault.jpg)







