I've described what kind of video this camera is intended to capture: https://blog.mikeswanson.com/apples-mysterious-fisheye-proje...

I just want to say that your analysis is great. I reference it whenever the discussion comes up about the technicalities of Apples solution.

Thank you! I really appreciate the feedback.

I guess the next step here is a video encoder that is natively recording spherical videos ?

Any projection is bound to separate areas which could be compressed more efficiently together.

A native stereoscopic spherical video encoder could improve compression even more, since side by side views are quite similar in general.

Now that's an interesting problem to solve ! (and a very hard one probably)

> A native stereoscopic spherical video encoder could improve compression even more, since side by side views are quite similar in general.

Existing video formats already support this for interlacing, although you could also let inter-prediction refer to earlier parts of the same frame and get most of the benefit.

It is developed in the mpeg standard already, there is a whole group of people for that from some years ;)

Hey! Nice writeup, Just something is missing, some MPEG formats can encode this kind of video in OMAF specification.

Edit: I'll certainly read the rest of the articles!

Thanks! Indeed, there are other formats (like OMAF) that describe some of this. In fact, I helped to author one a long while back called OPF.

Did WWDC end up filling in any additional blanks from your article?

Thanks!

Unfortunately not. I hope we learn more before WWDC25!

You'll find it interesting that people with aphantasia (I have total, multi-sensory aphantasia) counterintuitively score better on mental rotation tasks: https://www.sciencedirect.com/science/article/pii/S105381002...

I do indeed find that interesting. Checking the paper, it does make some sense to me. It’s similar to what I said in another comment [0], the people with aphantasia do a slower, more methodical approach. It then fails for me when there are too many steps required, which is not the case for simple rotations as in the paper.

[0]: https://news.ycombinator.com/item?id=42439783

My one-person indie company released many apps, and one of them (Halftone) had over 6 million users by the time I shut it down. It's definitely possible.

Interested to hear why you shut it down when you had over 6 million users?

Likewise, I run my Nikon Coolscan on my Windows 11 machine with no special adapters. Works better than my Epson 750 Pro and Sony A7RIII with the right lenses, mounts, and lighting.

I also run a Nikon Coolscan and Epson 750 Pro on my Windows 11 machine. I also shoot slides using my Sony A7RIII with appropriate lighting and mounts. The Coolscan consistently gives me the best results.

It is. Or at least it was for some of us. I didn't care if the candidate ever got the right answer. I cared about the thinking, the questions, the strategies, and the conversation.

And if some interpretations lead to trivial solutions, but one leads to a complex problem, it's likely that their intention is the latter. A kind of tacit communication

Actually, it may just as likely be that the interviewer is looking to see if you over complicate things. So _ask_.

One of my all-time favorites, and wow, this brought back some fantastic memories! What great sounds!

Indeed. And this is where arc-length parameterization would help by linearizing t so you can then overlay it with various timing and easing functions.

Exactly, differential geometry is the easiest when using arc lengths as parameters.

Now, if we just add Sloot's compression technique, we can include an entire movie library on a single cartridge! https://en.wikipedia.org/wiki/Sloot_Digital_Coding_System

Interesting mystery!

> The Sloot Digital Coding System is an alleged data sharing technique that its inventor claimed could store a complete digital movie file in 8 kilobytes of data — violating Shannon's source coding theorem by many orders of magnitude. The alleged technique was developed in 1995 by Romke Jan Bernhard Sloot …

> just days before the conclusion of a contract to sell his invention, Sloot died suddenly of a heart attack. The source code was never recovered, and the technique and claim have never been reproduced or verified.

8k for a prompt perhaps

Thanks! Yeah, I think it'd be considered fair use too, but I didn't want to go there.