Join 3,424 readers in helping fund MetaFilter (Hide)


Focus on the potential.
March 14, 2012 10:53 AM   Subscribe

Lytro, the 6-year-old[1] consumer plenoptic camera start-up started shipping their first light field camera to end users last week. Reviews have been mostly positive with regards to the technology and industrial design, but also warn users of specialized hardware and software that is difficult to use, the "poor quality"[2] display, and low resulting image quality. However everyone seems to agree that light field technology is the way of the future and is here to stay. Previously.

Is this what the future looks like?
posted by jeffamaphone (70 comments total) 8 users marked this as a favorite

 
Just like a lot of technologies, early adopters get the privilege of funding development to make Lytro competitive with existing cameras. Add me to the list of those willing to wait.
posted by tommasz at 10:57 AM on March 14, 2012


Light-field photography looks like it could be really, really fantastic if the benefits, especially post-shot focusing, can be had outside of perfect conditions. So far they can't, but I'm hoping that's a product of the youth of the technology and not a scientific limitation.

I still have no idea why they decided to make their first production model a telescope, though. You'd think that if this is the device that's supposed to sell people on light-field, you'd make it capable of showing off the wonders of the technology right there and then, but that screen isn't showing off anything, and the form factor is...well, it's unique for a camera, I'll give them that.
posted by Holy Zarquon's Singing Fish at 11:05 AM on March 14, 2012


This is pretty cool. Blade Runner, here we come.

Enhance 224176
Enhance, Stop
Move in, Stop
Pull out, Track right, Stop

posted by calamari kid at 11:08 AM on March 14, 2012 [3 favorites]


Not to be a nay-sayer, because I can see how potentially useful this technology could be...but I can also easily imagine it becoming a ham-handed overused gimmick, a la HDR, that detracts more than it adds to an image in terms of "art" or (metaphorical, not literal) "depth". Some of those images in the light field link illustrate exactly what I'm talking about; they show off the technology but the multi-focus feature really doesn't "do" as much for the picture as they'd like to claim.
posted by Greg_Ace at 11:09 AM on March 14, 2012 [1 favorite]


Sorry, not 'light field' link but the this link at the end of the post.
posted by Greg_Ace at 11:11 AM on March 14, 2012


Greg Ace: I can't see your link. Perhaps my focus is wrong?
posted by lalochezia at 11:15 AM on March 14, 2012


The early stuff is almost always sort of crappy. I'm sure this tech will go very far.
posted by Sticherbeast at 11:15 AM on March 14, 2012


I don't care so much about shifting the focus from one element to another as an artistic statement, but I took roughly seven squintillion pictures of local news and high-school sports when I worked for a local newspaper, and by my recollection, three of them were perfectly in focus. Maybe four. The ability to fix that in post-production (without the Photoshop "sharpen" tool) instead of throwing out a bunch of otherwise well-composed shots, would have been a game-changer.
posted by Holy Zarquon's Singing Fish at 11:16 AM on March 14, 2012 [5 favorites]


I'm also fairly hesitant about this. The technology is undeniably very cool, but I'm somewhat unconvinced about a choose-your-own-focus presentation ever being anything more than a gimmick.

Now, if you are talking about this technology as a way to be able to fix focus issue in post-processing, I'm really excited about the possibilities (who hasn't had a shot they otherwise loved end up just slightly mis-focused?). I just think that even if I had this technology in a camera, I'd want to set a static focus for the final presentation.
posted by tocts at 11:19 AM on March 14, 2012 [1 favorite]


I ordered one. Looking forward to its arrival. I'm sure it will be interesting to learn how to work within its constraints, and to figure out how to make the best of what it can do.

I spent several years using a Rollei TLR passed down from my grandfather. No zoom, no electronics, no interchangeable lenses, just a square frame and the classic shutter/aperture/focus controls. It cost a few bucks a frame and you only got 12 frames per roll, so you had to approach photography with a slow, careful mindset.

I'm not sure what kind of mindset the Lytro will demand, but that's part of the fun of it: I'm looking forward to playing with it and finding out how I'll have to think about pictures in order to get the most out of it.
posted by Mars Saxman at 11:28 AM on March 14, 2012 [4 favorites]


tocts: I feel the same - the live-refocusable image makes a good web demo, but I'm not intending to do that myself. I just expect that setting the focus will be something I can start doing in post, like cropping and color correction. The pictures which will ultimately go up on my blog will be static JPEGs.
posted by Mars Saxman at 11:30 AM on March 14, 2012


If you think fixing minor focus mistakes in post-production is the best use of this technology, that's basically the same as the technology being worthless. For the amount of technical trade-offs you have to make to get the desired result, you'd be far better off with a specialized high-FPS DSLR with focus bracketing.
posted by 0xFCAF at 11:32 AM on March 14, 2012 [3 favorites]


If you think fixing minor focus mistakes in post-production is the best use of this technology, that's basically the same as the technology being worthless. For the amount of technical trade-offs you have to make to get the desired result, you'd be far better off with a specialized high-FPS DSLR with focus bracketing.

Fixing focus in post is not necessarily for the kind of people who will use (and afford!) a high-FPS DSLR with focus bracketing.
posted by Sticherbeast at 11:35 AM on March 14, 2012


We're talking about a $500 1.2MP gimmick camera and the word 'afford' is in play now?
posted by 0xFCAF at 11:37 AM on March 14, 2012 [3 favorites]


Only money and time are making me hesitate; this is a great piece of technology. The one thing I dislike is that the storage is built-in rather than removable, making it a little unhackable for my taste. Making sure there's enough light for a decent picture appears to be the main challenge. I can't wait to see this paired up with a fast sensor recording video; once there's enough computing horsepower thrown at it, this will allow focusing to become part of the edit, and that's going to make film production an order of magnitude faster and cheaper, not to mention more friendly to directors and actors. This has already happened with audio and lighting; 3 people with a DSLR and a pocket digital recorder can shoot a good-looking feature film. I crewed on such a project last year. To be able to fix focus in post means two things: you can rehearse your actors and shoot in many more locations without being noticed, and you can alter or remove unwanted background elements (people, vehicles etc.) much more easily.

The form factor is genius. It's easy to hold, less intimidating than a regular camera to children and animals (and people who hate being photographed), and above all else, distinctive - like a land camera, Holga, or other famous firsts, the shape is part of the branding and will do more to sell this than any amount of banner advertising could. As a practical matter, it's also easy to clamp and trigger remotely.
posted by anigbrowl at 11:39 AM on March 14, 2012


The base model is $400, the larger memory version is $500[3].
posted by jeffamaphone at 11:39 AM on March 14, 2012


Yeah, I can't help but think that if this had been around since the 80's and invented by Kodak it'd probably be mostly forgotten. I mean why don't people use stereoscopic cameras? 3-D is much cooler than changing depth of focus. Because its old technology.

The Lytro pictures I've seen have been a bit fun to play with, for a small bit of time, but thats mainly because of the novelty of it. Novelty wears off very quickly, especially these days.
posted by vacapinta at 11:39 AM on March 14, 2012


The early version is of course crappier and more expensive than where this technology will lead.

And even then, $500 is much cheaper than a high-FPS DSLR (plus lens) with focus bracketing, plus you can fix the focus in post without having to rely on your DSLR's FPS. Are you telling me that you've never had a shot where the AF missed focus?
posted by Sticherbeast at 11:40 AM on March 14, 2012 [1 favorite]


Yeah, I can't help but think that if this had been around since the 80's and invented by Kodak it'd probably be mostly forgotten. I mean why don't people use stereoscopic cameras? 3-D is much cooler than changing depth of focus. Because its old technology.

Panasonic and probably some others are trying to sell 3D cameras nowadays. Lord only knows how well that's going.

(BTW, Nishika 8000 stereoscopic film cams are fun, but yes, only as toys.)
posted by Sticherbeast at 11:42 AM on March 14, 2012


If you think fixing minor focus mistakes in post-production is the best use of this technology, that's basically the same as the technology being worthless.

I don't think that's entirely true, but yes, at this point I do think it's about the best use of the technology as far as still photography goes.

The presentation of the Lytro photos linked here is fantastic for showing off what the technology can do. However, I remain unconvinced that such a presentation really adds anything to the resulting shots. It's a great tech demo, but it'd feel awful gimmicky if that's how someone was publishing all their photos.

We're talking about a $500 1.2MP gimmick camera and the word 'afford' is in play now?

No, we're talking about the technology behind a $500 1.2MP gimmick camera, which over time will improve and become cheaper. If this sort of thing makes it into a consumer P&S (or a smartphone camera), it would be a huge deal. I can imagine it being a big hit if, for example, you could just tap your touch-screen where you had meant focus to be, and suddenly the photo re-focuses. But, I don't think many people are going to be that interested in permanently presenting photos with dynamic focus points.

Hence, my comment that this is interesting as far as focus correction goes, but in my opinion unlikely to be a huge hit as a presentation method.
posted by tocts at 11:42 AM on March 14, 2012


I agree with the above commenters that this could easily add to the list of things that's "neat-but-not-art".
What it could do, however, is provide a really cool way to create an interactive website. You take a great, giant picture of a restaurant-in-action, and let viewers zoom in and out of the photo, and click on aspects of the picture, and have it lead them to different about-pages.
posted by FirstMateKate at 11:44 AM on March 14, 2012


Are you telling me that you've never had a shot where the AF missed focus?

Well, sure. Some of those I was able to "fix" by just scaling them down to 1080 pixels. :D
posted by aubilenon at 11:45 AM on March 14, 2012 [1 favorite]


vacapinta: I mean why don't people use stereoscopic cameras? 3-D is much cooler than changing depth of focus.
It's a developing field, actually. The main issue with 3D cameras has been that displaying the pictures is so inconvenient, but that's apparently less of a problem now.
posted by Western Infidels at 11:46 AM on March 14, 2012


I mean why don't people use stereoscopic cameras? 3-D is much cooler than changing depth of focus. Because its old technology.

No, because you need a special viewer to look at the pictures, or 3d glasses at the movies...which you may have noticed are making a comeback. Panasonic sells a range of 3d camcorders, from consumer to professional.
posted by anigbrowl at 11:47 AM on March 14, 2012


Yeah, I think people will use this technology to strike JPG "prints" of photos, not to distribute photos where you can fix the focus in post. There will be specialty applications for light field cameras, especially in surveillance, law enforcement, and probably wildlife photography as well, but mostly it will be used as a thing you can do to avoid ever having an unfocussed image again.

This technology will not necessarily replace conventional focus technology ever, but that's not the point. It's just a different tool. iPhone cameras didn't put Hasselblad out of business.

No, because you need a special viewer to look at the pictures, or 3d glasses at the movies...which you may have noticed are making a comeback. Panasonic sells a range of 3d camcorders, from consumer to professional.

3D pictures are fun as animated gifs. I have some examples here, here, and here. Follow the directions in the comments to get to the animated bit.
posted by Sticherbeast at 11:50 AM on March 14, 2012


If you think fixing minor focus mistakes in post-production is the best use of this technology, that's basically the same as the technology being worthless.

So, given otherwise comparable options in terms of price and featureset, you wouldn't pick the camera that lets you fix focus in post-production?
posted by Holy Zarquon's Singing Fish at 11:52 AM on March 14, 2012


Damn, if only there was technology to write a review in such a way that it doesn't crash Safari multiple times.....
posted by c13 at 11:52 AM on March 14, 2012


No, because you need a special viewer to look at the pictures, or 3d glasses at the movies...which you may have noticed are making a comeback.

I was actually thinking of those wiggly gifs which require nothing special. Why aren't more of my friends posting wiggly gifs?? Actually, that would get annoying too. Just like the Lytro pictures everyone has started posting to advertise they own the latest gadget.
posted by vacapinta at 11:53 AM on March 14, 2012


I just shoot everything at f32 and blur the unwanted stuff out.

(parks self in the "gimmick" camp, surprise surprise)
posted by ShutterBun at 11:55 AM on March 14, 2012 [3 favorites]


It's a great tech demo, but it'd feel awful gimmicky if that's how someone was publishing all their photos.

I disagree. What you're seeing here are (mostly) poor quality pictures which are only interesting because of the focus 'gimmick'; if you had 10 excellent pictures and could also play around with the DoF that's be a nice plus. By 'excellent' I mean in terms of composition, lighting and filtering etc.

If this sort of thing makes it into a consumer P&S (or a smartphone camera), it would be a huge deal.

This is a P&S camera as far as I'm concerned, albeit a rather expensive one. I think this will take off like wildfire after a slow beginning, and be the norm for within 10 years. Think of the utility for photojournalism and so forth. The underlying technology can be paired with existing lenses and camera components, which makes for some very exciting possibilities. If I had a few thou$and and nothing else to do right now I'd be trying to build a stereo rig with this and a pair of wide zoom lenses.

Not just for art; this will have massive implications for robotics, brain science, and post-processing of existing photos (because there will be an endless supply of new blur kernels allowing us to better infer what the blurry shapes in existing photos would have looked like if they were in focus).
posted by anigbrowl at 11:57 AM on March 14, 2012


I received mine a week or so ago and think it's very interesting. It's fun and the tech-geek side of me is fascinated.

From a practical perspective, I think that their best future use will be for security cameras - 1 camera to cover a large area where the police or security team can focus on the area they want to see after the fact.

Other than that, it's main appeal will be for those who can't figure out how to focus an auto-focus camera so they will be able to take a shot without worrying about focusing.
posted by johnn at 11:57 AM on March 14, 2012


I think the most interesting thing that this could provide is a Z-buffer with the image, which has all sorts of cool robotics and sensing technologies application. As a consumer technology I'm gonna toss myself in the "gimick" camp, but if you fed this through a pipeline that'd use local contrast at different depths to pull out depth information, and did image processing on that, I'm thinking that the self-driving car folks could reduce their use of LIDAR sensors.
posted by straw at 11:58 AM on March 14, 2012 [2 favorites]


I got into a discussion the other day with a photographer friend who said the Lytro was "cheating." She didn't seem to see the irony in her statement when I pointed out that she shot RAW (which allowed her flexibility in white balance and a myriad of other options), had a nearly infinite roll of film in her camera at all times, had earlier lusted after the Nikon D800 with it's 51 point AF system, and super high ISO shot ability.

I didn't even mention that my friend who has worked in film and photography for a long time, shooting and developing his own prints (but now uses digital because it's cheaper) thought the Lytro was neat, and was the first person to tell me about it when they started hyping it.

Eventually I think this tech will trickle down into the P&S market, and will probably make its way into mobile phones. The casual shoot market will adopt it first. It may eventually make its way up into the high end DSLR market or maybe the EVIL class systems, as have previous consumer only tech has (i.e., camcorder modes, etc.). I remember reading that Jobs was excited about the tech, and I wouldn't be surprised if an iPhone would be the first mobile device to carry the Lytro system once it would be possible to shoot 8mp and refocus on a portable device.
posted by mrzarquon at 12:02 PM on March 14, 2012


Why aren't more of my friends posting wiggly gifs??

Because the speed of the wiggling typically isn't correlated to anything in particular, which makes them annoying after about 10 seconds. To make it look truly 3d they should wiggle at about 48Hz, have a subtle alternating vignette & blur, be landscape, and have an aspect ratio > 2. Virtual camera movement, on the other hand, has become a well-established technique.
posted by anigbrowl at 12:08 PM on March 14, 2012 [1 favorite]


So, given otherwise comparable options in terms of price and featureset, you wouldn't pick the camera that lets you fix focus in post-production?

Why doesn't everyone always shoot with the widest angle lens possible and crop in post? You can do a 3x crop and lose less native sensor resolution than this camera does (11MP sensor -> 1MP output)
posted by aubilenon at 12:08 PM on March 14, 2012


if you had 10 excellent pictures and could also play around with the DoF that's be a nice plus.

I have to continue to disagree. I suppose time will tell, but I just do not buy it as a presentation.

I completely buy it as useful to have in other applications, like security/surveillance as has been posited, and I can imagine how this would be useful for robotics. I can even imagine how, if the technology scales up to higher resolutions, it might spark a focus-in-post revolution in filmmaking (similar to digital color correction, and likely to be abused just as much at first, but nonetheless allowing for greater directorial control than has ever been achieved previously). But I'm having a hard time imagining many still photo scenarios where not choosing one focal point in the end would be particularly compelling.

This is a P&S camera as far as I'm concerned, albeit a rather expensive one.

I'm sorry, but a $500 brick of a camera with bad ergonomics and a difficult to use screen that requires special software to do anything with the photos from it is about as far from a consumer P&S as is possible.
posted by tocts at 12:08 PM on March 14, 2012


Lytro isn't the first company selling these camera. Raytrix has sold plenoptic (lightfield) cameras since 2010. They even have video cameras that use this technology now.
posted by MythMaker at 12:08 PM on March 14, 2012


johnn, as you have one am I correct in thinking it's basically a sealed box with only a USB connector? how difficult do you think it would be to, ah, open it?
posted by anigbrowl at 12:10 PM on March 14, 2012


Other than that, it's main appeal will be for those who can't figure out how to focus an auto-focus camera so they will be able to take a shot without worrying about focusing.

It's a wonderful device for my mother, eventually. No matter how many times I tell her how to focus *the most simplest camera ever* she can't figure it. This will make it worth her while taking fotos, as she will know that 80% won't turn up blurred.
posted by Jehan at 12:12 PM on March 14, 2012


a camera with bad ergonomics and a difficult to use screen that requires special software to do anything with the photos from it is about as far from a consumer P&S as is possible.

Come on, consumers have been buying shitty cameras fitting that exact description for years.
posted by ShutterBun at 12:15 PM on March 14, 2012


I'm sorry, but a $500 brick of a camera with bad ergonomics and a difficult to use screen that requires special software to do anything with the photos from it is about as far from a consumer P&S as is possible.

It's $400. the $500 is if you want more storage. The ergonomics are excellent: make a tube from your fingers and hold it up to your eye. This is the same principle as a director's viewfinder. Everyone with hands knows how to hold this thing and point it in the right direction. I know pro photographers taht have been crying out for years for a digital camera to be shaped like a barrel instead of having a giant box stuck on the end of the lens.

The point of P&S is not that it's cheap or has lots of features, it's that you can just point and shoot. That's exactly what this does. No more waiting ot focus, just point it at what you're interested in and push the button, the end. You don't need to check the picture in any great detail because you know it will be in focus. Of course what people will be interested in will be a specific focus point, just as autofocus proved ideal for casual snapshots of people's night out. But since there's no great barrier to publishing with the focus information, many people will do that by default rather than going through the additional step of striking jpegs - which in turn will lead to an expectation that all photos should be refocusable, just as Photoshop and its ilk led people to think that all images could be editable, rather than requiring a darkroom and years of training.
posted by anigbrowl at 12:20 PM on March 14, 2012 [1 favorite]


Come on, consumers have been buying shitty cameras fitting that exact description for years.

Most P&S cameras I've interacted within in the last decade or so, you basically plug it in and you're done, as far as software goes. If you're on a Mac, iPhoto likely just opens right up, and you import your photos. No effort required. And you can get one that's good enough for people who aren't photography buffs for under $100. (Hell, I bought one for like $70 for my mother-in-law a few years back, it does everything she needs and she loves it!)

As for ergonomics, they may not be great, but come on. Even the worst designed P&S typically is something that fits comfortably in a pocket or purse, and is generally obvious as to how to hold. The Lytro, on the other hand, looks like a boxy telescope. The reviews indicate that it's got a really badly designed zoom control, and the thing looks more like it's supposed to be mounted in a high corner of a convenience store than handheld.

This is an early adopter model. There's nothing inherently wrong with that. As I've said, I think the tech is interesting, and I think particularly when it is cheaper and included in a more traditional camera body, it's going to be a big selling point. But this is not a ready-for-primetime, mass market consumer product.
posted by tocts at 12:25 PM on March 14, 2012


Why doesn't everyone always shoot with the widest angle lens possible and crop in post? You can do a 3x crop and lose less native sensor resolution than this camera does (11MP sensor -> 1MP output)

I said comparable features. 11MP and 1MP are not comparable. I'm not interested in buying this camera either - I'm interested in where the technology can go once it's been perfected and made cost-effective to the point where it can be included in a camera with the price and specs of a PowerShot.
posted by Holy Zarquon's Singing Fish at 12:25 PM on March 14, 2012


Has anyone spotted a specification for whatever is the light field equivalent of minimum focusing distance or maximum enlargement? What about limits of "virtual F-stop"?

What I want to know is: can Lytro replace a traditional 100mm 1:1 macro lens? Can it provide arbitrarily large depth of field? If it did, it would be relevant to my interests.

Actually, I found an example from lytro but it's quite underwhelming. here For some reason, you can't focus all the way "out" to the hand in the background, and I also can't find a spot to click on the ladybug so that I'm satisfied with the focus on it either!

Why can't I move the focus incrementally with e.g., the mouse wheel?
posted by jepler at 12:34 PM on March 14, 2012


and you have to have a mac? I'm surprised nobody else has noted that so far.
posted by jepler at 12:35 PM on March 14, 2012


"...the thing looks more like it's supposed to be mounted in a high corner of a convenience store than handheld."

Yes. That will be, eventually, the point. The street finds its own use, etc.
posted by digitalprimate at 12:43 PM on March 14, 2012


I said comparable features. 11MP and 1MP are not comparable

You also said comparable price. But the way this technology works means you're 'spending' 91% of your pixels on depth information - you aren't going to see lightfield cameras output comparable resolution images at a comparable price to other cameras until the cost of the electronics in the cameras falls enough to be a negligible part of the price.

And to me, letting viewers choose the focus seems more limiting than liberating. This is how I feel about QTVR panoramas too. Focus and framing are important tools a photographer can use to create an impactful image. I don't want to cede that control to the audience.

I don't mean to say there's no value to those interactive experiences, but those, and video as well, basically fill a different role than traditional photography, and they aren't likely to replace it.

It's interesting to consider all of this at its extremes. If technology were advanced infinitely, would it make sense to shoot high-resolution, wide-angle, high-speed lightfield video? You could just go somewhere, shoot for a bit, and then decide aperture, exposure length, timing, and framing later. But unless I could at least record my 'suggested' values for all these at capture time, I would find this unsatisfying. I want to spend more time in front of a camera, and less in post processing, not the other way 'round.
posted by aubilenon at 12:46 PM on March 14, 2012


and you have to have a mac? I'm surprised nobody else has noted that so far.

What's more concerning (to me anyway) is that Lytro hosts all the photos, subject to their terms of use

Here's what is forbidden:
(1) is unlawful, obscene, pornographic, violent, defamatory, fraudulent, harassing, or harmful to others; (2) violates the rights, including intellectual property rights, or the privacy of any other person; (3) incites or furthers criminal or unlawful acts; (4) constitutes hate speech or a personal attack; (5) contains viruses or other features that can harm Lytro.com or other property; or (6) which may otherwise expose us or our users or visitors to liability.
Understandable restrictions, but still, it stinks that you can't just host them yourself so you don't have to care if someone, somewhere in the world considers your image to be obscene or a personal attack.
posted by aubilenon at 12:52 PM on March 14, 2012 [1 favorite]


The Lytro, on the other hand, looks like a boxy telescope.

You say this like it is a bad thing. I think that is what makes it intuitive.

I also can't find a spot to click on the ladybug so that I'm satisfied with the focus on it either! Why can't I move the focus incrementally with e.g., the mouse wheel?

The sensor in the thing is not that much bigger than an iPhone, so you're running up against the limits of the ahrdware. The best focus is on the ladybug's shoulder or head rather than on its carapace, but you won't be able to get it razor-sharp.

One should be able to focus with the mouse wheel, and eventually will...on the desktop, whether with their software or a competitor's. I'm not sure if Flash recognizes that as an input, but the click to focus/double-click to zoom simplicity is essential at this stage to introducing the product. Every one of those photos is an ad for the product, and must thus be as accessible as possible to casual web surfers who come across them. Lytro is not (presently) aiming at the pro photographer who thinks a lot about the picture, but rather at people who want to understand and use something immediately. The lack of features/configurability is an asset when it comes to introducing a fundamentally new technology like this - just as it was with Polaroid. That technology would never have taken off if users had had to fiddle extensively with the camera before shooting a picture.
posted by anigbrowl at 12:54 PM on March 14, 2012


And I'd like to get the picture I actually want, not the picture the camera decided to take because the autofocus really likes shooting crystal-clear images of the fence behind my subjects.

I think we're coming at this from very different perspectives, though. I have no attachment to the artistic experience of shooting a picture; ultimately, my camera is the tool I use to get the best still image recorded and printed. I really don't care if I end up using it less and Photoshop more, as long as the finished product is better.
posted by Holy Zarquon's Singing Fish at 12:58 PM on March 14, 2012


anigbrowl - Yes, it is a sealed box...not sure I'm ready to try to open it just yet :)

As far as hosting the photos at Lytro, yes, if you want to allow others the ability to "play" with your photo, yes, it must by hosted by them.

But (and it's a big butt!) you are supposed to be able to take photos, set the focus point and then export it as a JPG (or other). I don't see that functionality in the software just yet, but will be keeping an eye on it. To me, once it's focused, then I might want to print it or send it to someone.
posted by johnn at 1:04 PM on March 14, 2012


And, just like that, I found the Export button. It exports as 1080x1080 and is as clear as the original photo. Although it shoots at f2.0, it loves a lot of light.
posted by johnn at 1:15 PM on March 14, 2012


Oh, oh.. Mefi prognosticating that a technology has no future!! I think this is where I get to pull out my backpocket post from 2007.

Behold the predictions of doom and irrelevancy for the gimmicky, poorly designed, inauthentic and positively silly device being discussed.

And indeed, who has heard of the Kindle these days? That thing went nowhere.
posted by PissOnYourParade at 1:57 PM on March 14, 2012 [1 favorite]


Considering people bought thousands of the dismal 640x480 Casio QV-10 back in '97, the Lytro has a long future ahead of it.
posted by scruss at 2:01 PM on March 14, 2012


lalochezia - Maybe; try focusing on the actual link in the O.P., that I was merely referring to rather than re-linking. ;-)
posted by Greg_Ace at 2:49 PM on March 14, 2012


If I understand correctly, the main pieces of technology here are an array of tiny lenses, and the software to extract images from the resultant data. Would it be possible to reverse the technique with a super high resolution screen, to display a light field?
posted by lucidium at 3:08 PM on March 14, 2012


Oh, oh.. Mefi prognosticating that a technology has no future!!

Well, lots of people are prognosticating in lots of different directions, and I don't want to speak for them. My own prognostications are not that this necessarily "has no future" but mostly that I don't necessarily agree that it "is the way of the future" as the FPP says. That's the difference between asking "will people keep making/buying this new type of camera?" and "will people keep making/buying the the old kind of camera?"

For instance, that Casio QV-10 scruss mentions, yeah that fits in the "way of the future" category - digital photography has completely replaced film photography in most applications. I don't believe light field cameras will "replace" conventional cameras in the next 15 years.
posted by aubilenon at 3:09 PM on March 14, 2012


I had one of those QV-10 cameras... and the worst thing about it wasn't the 320x240 resolution, it was the fact that it it stored images in RAM, and if you let the batteries die, you lost your 20-40 photos. The fact that it had no removable storage was the second worst problem.

The light-field is cool stuff, I've been doing experiments with the idea for a few years. Here are some of the better results hosted on Flickr.

In the future, waving the camera around may help you get a better image, as the computer can track motion, take lots of very fast exposures, then merge them together in either 3d or superresolution, depending on your needs.
posted by MikeWarot at 5:40 PM on March 14, 2012


One of the founders told me that they were definitely planning to keep people on the website for the foreseeable future, not exporting jpegs or on a format. This made me depressed.
posted by curuinor at 8:05 PM on March 14, 2012


You should be able not just to set the focal plane and DOF in postprocessing, but to play some more interesting tricks. No reason the focal plane has to be perpendicular to the camera's view axis, for example. Or that it has to be a plane.

I've taken plenty of photos where the subject of interest is at multiple depths (two eyes, for example) so I can never hit the focus right, short of an unreasonable tilt/shift lens. (I do wonder: in digital, why not do a big image circle and a tilt/shift sensor?) I'd go for this postprocessing capability.
posted by away for regrooving at 10:56 PM on March 14, 2012


That's an intriguing idea, regrooving. Applying a "true" (albeit simulated) tilt-shift effect in post processing would be intriguing. Although tilt-shift has to do with more that just "what's in focus." The "shift" part of tilt shift involves moving the lens horizontally or vertically, so that the image can be re-framed without moving the camera. (things like parallel lines converging in photos of skyscrapers can be avoided using this method)

In that respect, I don't think a lightfield camera would be of any help.

Or that it has to be a plane.

Technically, I think focal planes (at least in the sense we're using the term here) are not planar, but rather bowl shaped (or the inside of a sphere, if you will) since it's simply a specific distance from the point of focus (the point at which the lens is creating the sharp image; ideally on the image sensor or film)

In reality, a truly planar area of focus might be more novel. (say, for taking sharp pictures of a wide horizontal subject with everything in focus)
posted by ShutterBun at 11:41 PM on March 14, 2012


Technically, I think focal planes (at least in the sense we're using the term here) are not planar, but rather bowl shaped (or the inside of a sphere, if you will) since it's simply a specific distance from the point of focus (the point at which the lens is creating the sharp image; ideally on the image sensor or film)

But the points on the sensor are not a constant distance from the lens either. The 3d shape of the focal plane is in fact a projection of the shape of the sensor. If it were bowl-shaped, you'd focus on a bowl-shaped region.

As luck would have it, yesterday I was shooting macro shots of graph paper, and I can assure you, the place where the razor thin focal plane intersects a planer subject is indeed a straight line.
posted by aubilenon at 1:12 AM on March 15, 2012


Here, I took an extra one just for you. Forgive my imprecision - it doesn't line up exactly with the grid, but it's still quite clearly a line and not an arc (angle of view is 18°, so that would be quite noticeable down near the bottom of the image like that).
posted by aubilenon at 1:34 AM on March 15, 2012


I just shoot everything at f32 and blur the unwanted stuff out.

(Note to self... ask for ShutterBun's lighting equipment from grieving widow "as a momento", after he dies of skin cancer... and buy SPF100 sun block in bulk.)
posted by IAmBroom at 9:29 AM on March 15, 2012


If anyone here got a Lytro and is sharing photos, I would love to see what you've shot.
posted by jeffamaphone at 10:13 AM on March 15, 2012


Aubilenon, yes, in fact back in the days of film, there were cameras specifically made to include (slightly) "bowl shaped" film planes to accommodate field curvature. Nowadays, the correction is simply done by increasingly complex lens design, to the point where many lenses can minimize the effect of field curvature to tolerable levels. (macro lenses are specifically designed with this in mind, since they spend so much of their time photographing flat subjects)

But the effect is always still there, at least somewhat. Here's a pretty good article on field curvature, including some graphic examples.

I'm not sure what camera/lens setup you've used in your example photo, but if you were to try it with, say, a wide open 50mm prime lens with a full-frame sensor looking at a large brick wall, chances are you'd see some obvious sharpness dropoff in the corners.

ask for ShutterBun's lighting equipment from grieving widow "as a momento", after he dies of skin cancer.

I was joking about the f32 thing, but truth be told, my strobes are way overpowered. I'm rarely able to shoot any wider than about f11, even with the lights on their lowest setting. I'll put in a good word with the Mrs. once we get the biopsy results.
posted by ShutterBun at 11:30 AM on March 15, 2012 [1 favorite]


Shutterbun: Very interesting! This is not an effect I was aware of, though I guess it's not really quite similar to pincushion/barrel type distortion (which are also quadratic), just on the Z axis, and so I shouldn't be surprised.

Okay, so now I tried a similar test with Canon's 50mm prime, opened to 1.4. Here is that image. The DOF is deep enough here that I could draw a straight line all of which is in focus, but the red spline indicates where I think the center of the in-focus paraboloid intersects the paper.

I was of course aware that the corners are less sharp as the center on all lenses, but I had no idea that one component of this was just that they focus to a different distance. I'm sure that even if I were shooting the right shaped paraboloid subject, the corners would still be less sharp than the center though - all of the rays hit the glass at a more oblique angle, which I would think would magnify any imperfection.

The lens in the first test shot was Canon's 65mm MP-E 1-5x macro lens (I'm shooting on a 1st gen 5D: a full frame sensor). The focal plane in that first shot it sure looks flat to me, but I guess that just attests to how good they can get the geometry when they sacrifice features like affordability, auto-focus, wider apertures, focus-to-infinity, and 5 stops of light. Plus the 50mm lens's scene subtends nearly 3x the angle of view, which I can't imagine makes this any easier for it. I am curious to try with dot paper instead of graph paper, to see if I can see different parabolas for the two different directions of blur that article mentions, but I'm not going to do that now, because I have some important whiskey that needs drinking.

But now I know something I didn't know before - thanks!

(I was going to also try a shot with the cheaper 50mm f/1.8 lens, because I thought that would also be fun to compare, but the front half of it fell off in my hand. Fantastic plastic, my ass!)
posted by aubilenon at 10:51 PM on March 15, 2012


This is not an effect I was aware of, though I guess it's not really quite similar to pincushion/barrel type distortion (which are also quadratic), just on the Z axis, and so I shouldn't be surprised.

aubilenon, they're actually quite related. If you think about it, if the "best-image surface" is spherical (for a simple imaging system), distortion will be low on that surface itself, but any flat plane "imaging" surface that intercepts the focal surface at the optical axis will record both blur at the edges (from field curvature), and increased magnification at the edges proportional to distance from the center (which is distortion).

Real camera lenses are more complex, and the best-focus surface is more likely wavy and centered around the film plane, not tangent to it at the center, but you get the idea. 1st-order distortion is likely a side effect of field curvature.
posted by IAmBroom at 7:48 AM on March 16, 2012


The "not similar" you quote was an editing error - I meant "similar". (I can think of lots of way better examples of things that are unlike what we're talking about!)

This best-image surface you, ShutterBun, and his linked article are talking about: am I correct in defining that as the locus of the in-focus projection of a flat surface subject?

Could one instead project the sensor surface out into the scene, and describe a manifold with of all the points which are actually in focus on a real camera? That framing makes a lot more intuitive sense to me, assuming it's actually valid - which I'm not sure of. I'm still trying to wrap my head around this. Maybe I'll sit down with a compass, protractor, and straightedge, and do some raytracing this weekend.
posted by aubilenon at 10:54 AM on March 16, 2012


This best-image surface you, ShutterBun, and his linked article are talking about: am I correct in defining that as the locus of the in-focus projection of a flat surface subject?

Essentially flat: the source "plane" has a much greater radius (unless you're doing macro work!). If you're photographing a wall 10m away, the "source" curvature is 10e4mm.

Since the "ideal focus surface" actually has real depth (actually due to the Heisenberg uncertainty principle, if you're working at diffraction limits - which you often are at high apertures), you aren't likely to notice the source surface curvature. The Depth of Focus (DOF, measured at the source) may be >1m at 10m.


Could one instead project the sensor surface out into the scene, and describe a manifold with of all the points which are actually in focus on a real camera?

Bingo!!! Again, remember that the effective focus has a slow curve near its blur minimum - you're aiming for something more like the bottom of a hyperbolic curve than the point of a "V".
posted by IAmBroom at 12:10 PM on March 16, 2012


« Older A new working paper by economists Charles Courtema...  |  As fans of Community get ready... Newer »


This thread has been archived and is closed to new comments