Applied Science
1 HN recommender · recommended in 4 months, 2022–2026
14:22 · 597K views · Apr 29, 2020
Applied Science shows how telecentric and hypercentric optics differ from normal lenses, producing no or negative perspective. The video builds a telecentric system from a Fresnel lens and demonstrates a lens that can see around objects.
Commenters cite it as the clearest demonstration of hypercentric lenses and of looking around objects with optics.
5 comments from Hacker News
Applied Science has a good video[1] showing how you can look around objects. He is using a regular lens there, but with a lightfield camera you can emulate basically any lens you want in software as long as the captured lightfield is big enough. [1] this video
I wonder why animation hasn't enabled more experimental cinematography unhampered by physical limitations. So many possibilities: Long takes are exponentially more difficult to film IRL with increasing duration, yet come at no extra cost when the actors and camera are untiring and deterministic. Ray tracing allows for arbitrary virtual lenses, from prohibitively expensive to physically impossible. I saw this video demonstrating a lens that can continuously vary from endocentric to telecentric to hypercentric, where several comments are along the lines of, > Now I really want to see a scene in a horror movie, where in camera, using a MASSIVE one of these lenses, we get to see the killer is behind the person in shot, by adjusting the camera like this. Can you imagine how freaky it would be, at they seem to appear and then grow to loom over them, out of nowhere? ( this video ) Since telecentric lenses must be as large as their subjects, that won't be happening in the foreseeable future. Yet even Blender can render a working ray-traced camera.[1] So why don't we see these things in CGI? Even if most films stick to conventions, there should be some art films, right? Why isn't there a single one-take animated short? Addendum: CGI also allows exposure times longer than the time between frames (i.e. shutter angles greater than 360°). So, 60fps can have the same motion blur as 24fps if desired, which those who very vocally condemn high-framerate animation seem to very much. I find it qui
Wow, essentially an extraordinarily large view camera built around a repurposed lens, then converted into a depth-of-field adapter[0] in lieu of a digital back. There's actually a known technique for taking wide angle photographs with depth of field this shallow using only a typical long lens:[1] stitching together frames to build a wider field of view is all that's needed; long lenses all have shallow DoF by nature.[a] This interestingly can also be used to explain why the episcope lens has such shallow DoF: its focal length of 432 mm would make it a super telephoto lens, yet its large image circle means it has a wide-angle field of view. If it were replaced by a telephoto lens with the same focal length and aperture but a smaller image circle, and that lens were swung around in a spiral to 'paint' the same area as the episcope lens image circle, the two lenses would produce the same result, only the episcope lens can do it all at once instead of needing multiple exposures. [0] [LINK] [1] [LINK] -- [a] Annoyingly, I still don't know a satisfying explanation of why depth of field depends on subject/background distances, aperture, focal length, etc. The formulas exist, but every attempt at intuitive explanation I've seen has been incomplete or incorrect. Here's some videos to show those who say that longer lenses have shallower DoF solely because reframing by standing farther away from the subject causes the background and its blur to become relatively larger: [LINK] [LINK] (O
I think next-level would be a hypercentric lens that can see around / behind objects as build buy Applied Science: this video
That's sort of what a telecentric lens does. [LINK] [LINK] And you can go further: the hypercentric lens. this video