A smartphone with a fluid lens | Hacker News Reader

A smartphone with a fluid lens

economist.com

I worked at a startup that considered liquid lenses from a French outfit for optical applications back in the early 2010s. Problems with the technology include reliability (not great having fluid near electrical components), repeatability, and focus time. It's also hard to expect that the lens can easily 'snap' to focus, since there will be some play in the polymer / motor drivetrain. Mammalian eyes require a complex biological support system to keep them functional over years and decades. An unfortunate aspect of polymers (like the skin of this lens) is that they do not maintain flexibility in the long term.

Perhaps there's been a leap in the technology, but I'd avoid the first generation if possible, like I do with all newly implemented tech, since I consider the camera a critical function of my smartphone.

Changing lens shape is one thing, but it's also possible to just directly change the index of refraction of the material with any moving parts.

How do you do that, i thought this is a property of the material itself?

There are materials in which refractive index can be changed significantly by electric field/temperature/mechanical strain

Considering people now tend to change phones every couple of years, do you mean by "long term" something even shorter than that?

Classic HackerNews to have the top comment be an anecdote from 10 years ago.

What's with news stories without any visuals? It's a story about cameras, lenses and pictures without any pictures!

It's honestly absurd. On a slight tangent this reminds me of the fact that millions of extreme high quality photos are taken each day - but still most news outlets have poor quality photos, and very few of them.

Something is clearly not working licensing wise.

I remember the Boston Globe's "Big Picture" going viral often almost 10 years ago on Reddit with amazing high quality photos - then they disappeared because Reddit banned them for some reason - and it seems few has taken up the idea besides them since.

Remember that The Economist is in print, and while they do have color the quality of impression and the available space is probably not sufficient to get good enough details to meaningfully compare the technology.

Besides, it's not Wirecutter, they're reporting on tech, not making a buying guide. As such the text is really what matters.

I'm a bit surprised by your claim that "something is clearly not working" when I have a basically the opposite take: images and video take over most of the web's contents. You mention reddit, at any given time on the frontpage you'll have mostly images and small videos and very little text. And of course you have social networks like TikTok and Instagram that are only about images.

Given how ultra-comoditized photos and videos have become, it makes sense for The Economist to focus on the thing they actually do better than any random internet blog or youtube review video.

I actually mostly agree with you - our culture is saturated with images. Also i appreciate longreads and dense text.

I just don't agree with the comparison to tik-tok or other kinds of moving images. To me beautiful high res images, well curated and well presented in a layout or standalone is still pretty rare to see done well.

It's like the difference between well curated photo exhibition and a torrent of commercials, weird formats, fast moving images, and sounds / animation / overlays instead of simplicity and just letting images speak for themselves, slowly or completely without motion.

> video take over most of the web's contents

That’s only true because ad industry is paying more for video content

To avoid the paywall, read it here: https://archive.is/NoJ1F

This extension bypasses it effectively.

https://addons.mozilla.org/en-US/firefox/addon/bypass-paywal...

And for Safari users, the built-in Reader Mode bypasses paywalls.

Firefox has a reader mode also, but sometimes the reader icon doesn't appear. I'm not sure if this is based on some kind of "no reader mode" preference from the site, or something else.

If you are not offered reader mode, you can force it by prefixing the url with "about:reader?url="

For the time scale, I was working for Nokia in 2003 and this was already researched back then.

I think this type of advancement is really interesting and would be great to see improve and expand. This company in Canada is looking at liquid crystal lenses controlled by an electric field, I wonder how that compares to the Xiaomi tech?

[1] https://www.scopephotonics.com/technology [2] https://www.cbc.ca/news/canada/kitchener-waterloo/university...

The membrane is a mechanical element.

While LC lenses will no change their shape.

The problem with liquid crystals though is that they are birefringent.

The effect look like as if different focal planes have different (and quite large) apertures.

I believe this is where the company in questions uses software to mask it some way.

The benefit of liquid crystal lenses over other electrically tunable lenses is their thickness.

Other tunable lens materials don't have birefringence, but their electro/thermo/piezi-optical effect is much, much weaker, and large size.

The later is improving with each year. Usable solid TAG lenses may get below 1cm.

Whatever happened to plenoptic cameras like Lytro? Why aren't they included in cell phones?

Been interested in plenoptic cameras for a while. Here's my guess:

1. Focus after the fact was a nothingburger to consumers. Likewise only niche enthusiasm for 3d.

2. Plenoptic cameras require a lens array or equivalent in the optical path. Phones are extremely constrained on depth.

3. Phone makers realized that if they wanted different focal lengths, the most simple route was to put two compact cameras on it.

4. You can't really go below 8x8 pixel "subtiles" with a plenoptic lens array. So you're giving up a factor of 64 resolution at least. That's a heavy cost for features users are very apathetic about.

In particular, fake computer-generated bokeh turned out to be good enough for a lot of users. You get much better composed pictures without any additional hardware. A photographer won't be fooled, but Instragram and your scrapbook probably will be.

The Lytro was incredibly long. It was like a small telescope. My phone is insanely thin (please make the battery bigger and phone fatter). But the optics wouldn't come close to fitting.

The Lytro also had a low resolution. It was fun to click and shoot.

Yep. I have a 1st gen nonfictional one. The website shutdown and required special software, so it's not usable anyhow.

Assuming sensor resolution is fixed/limited due cost or size constraints, switching to a plenoptic lens will provide an image with a much lower resolution.

https://www.bbc.com/news/business-56237991

There was this article in the BBC about development in future smartphone cameras/lenses

Its a company in Norway called Polight that is working on a lens called tlens. They have already announced that they are working with a oem that is deciding about when and which models that they’re lens is going to come in with. This will be a cameramodul for selfie cam, that will give ultra fast autofocus and also will be the best option for under screen camera or small notch. They’re also stated that many oems is interested in they’re lens. Polight.com

I wonder if the attempt to "replicate" our eyes in this approach is not misguided. Why not have a set of small electromagnetic wires, a magnetic ferrous-fluid, some distributed optical sensor + orientation and allow a NN to manipulate these "optics" until it "sees" a reference picture. Given enough training, the optics should be able to out-wiggle nearly all traditional optic setups without moving parts and expensive lenses.

https://en.wikipedia.org/wiki/Ferrofluid_mirror

Laws of physics

The moment you will get a number of wires needed to provide an equivalent quality to a lens, you will be better off with a lens

Adaptive optics will work if you already have a good wavefront to work with, and you can remove more distortions from it with AO than you add

Yes, interesting idea, reminds me of fluid filled glasses - https://www.wired.com/2009/02/20-self-adjusta/

Am I missing something? Why isn't there a clear example of the lens in action?

Some technology will make this cheap and with enough granular control to become a staple lens technology.

Reply on news.ycombinator.com