« Updates & Followups: October 21st | Main | Photography vs. Image-making »

Friday, 21 October 2011


Feed You can follow this conversation by subscribing to the comment feed for this post.

I am not as optimistic as you regarding this camera: I say it will fail, and fail big.

The technology and idea behind it, on the other hand, will likely be incorporated into other picture-taking devices, making money for Lytro through licensing. But a kaleidoscope-like camera? Dead on arrival. If it doesn't fit in a pocket or disappear into a purse, the general public don't want it.

The activity we now think of as "photographing" might take place increasingly at a desk, at leisure, after the fact, rather than in the field.

Only after they commercialise the drones.

Really, really stupid industrial design. Like something "inspired" by Steve Jobs/Jonathan Ive, then implemented by Mattel.

I'm not sure people do view focus as a problem or obstacle. Most people squash the button, the camera's AF does something, and lo, a picture results. Sometimes it's "fuzzy" but that's increasingly rare.

"Have you noticed that every innovation these days seems to be directed at making fewer demands on the photographer's decision making?"

Maybe I'll be in the minority, but I don't think that's it. Indeed, like white balance in RAW processing, it simply postpones the moment the decision needs to be made. And, like white balance, it can be a lifesaver, instead of it being baked into the image once and for all. I'd say it's even a bit similar to what RAW processing vs in-camera JPEGs: instead of having all the choices made for us by an image processor, we get -indeed: have - to choose our own later. The demands are not fewer - if anything, RAW processing is more complicated than copying JPEG files from memory cards :)

(I do agree with Miserere the form factor seems very wrong, as well.)

One huge fail I see is that they are still squashing a 3d space into 2D. Just clicking somewhere on an image to focus seems very awkward to me, because where the pixels are in 2D space have little to do with where they are in 3D space. How they aren't implementing as simple a control as "Depth Slider" is beyond me. Maybe that comes in the proprietary software, and just isn't included in the website?

Anyhow, I agree. Like a lot of things, just cuz you can, doesn't mean you should.


Essentially, in addition to the color Bayer filter, this guy added a "vector filter". Which reduces resolution to about a quarter of the sensor resolution, but lets you "focus" later by re-interpreting the vector data from the raw file in the same way as you re-interpret the color data.

Really, wow! And that does appear to make a lot more sense than having physically moving focus parts in a cell phone.

The demo images in the gallery are too cool. How fun is that!?

For some reason all of this makes me want to dig my Nikon F2 out of the bottom of the closet.

I want to be positive about the Lytro camera, because the tech is cool, but I think it fails in two major ways.

1) Artistically - Uploading the "living pictures" and allowing viewers to change the focal plane removes that decision from the photographer entirely. Yes, there will be an export option, but Lytro is pushing the sharing and interactivity of their new file format.

2) Economically - Economic predictions are always flimsy at best, but I don't see this selling well beyond curious early-adopters. "Serious" photographers will skip this iteration and wait for something that doesn't cause hand cramps from just looking at it. Average consumers just don't pay for $400+ point and shoot cameras anymore. Also, the prime selling point of the light field camera is focusing after the fact, but most people don't even delete their bad photos after shooting, much less choose a focus point. People don't want to think about focusing at all, but the Lytro camera forces them to pay attention to focus much more than a small-sensor digicam or a cell phone camera would. A better consumer use for this technology would be combining face detection, and spitting out a JPEG with all of the faces in focus.

There's tons of potential for the light field technology, but I don't think this first camera fulfills it. I'd want to see a good tool for extracting a RAW file from a light field image with options to customize DOF, focus point, and bokeh. Then I can process that RAW file myself and spit out something resembling a photograph.

With regards to recent innovations being aimed at putting less demands on the photographer, that's a common trend across industries. When Apple brought the mouse and GUI to the mass market, a fair number of experienced computer jockies thought it was unnecessary and/or "dumbed down" the computer. Same for auto-exposure and then auto-focus on cameras. For that matter, I have a friend who's leading a team to implement fly-by-wire on U.S. Army helicopters which will make them easier to fly and decrease accidents -- something that's been possible for decades but resisted up until now because of an attitude that "helicopters are supposed to be hard to fly."

Still, now even the vast majority of "pros" use these kinds of features, at least some of the time. I imagine plenoptic cameras (and your suggestion that in the future photos will be chosen from a frame of video rather than being a specifically-taken shot) will benefit photographers who are in situations where they can't give their full (or even much of their) attention to photography -- e.g. photojournalists in a war zone or sports photographers running alongside marathoners. They'll manage to get shots using this sort of technology that they wouldn't be able to get using traditional methods.

Careful composition, focusing, and exposure will still have their place and will provide superior results for particular types of photography, but technology that lets people "capture as they go" and then select/edit/refine afterwards may well become the norm.

I don't really see how this appeals to the average consumer, who takes In-focus images for granted already. Cellphone cameras have enormous DOF due to small sensors. DSLRs have blindingly fast autofocus. Is someone going to buy this camera for the convenience of saving a millisecond at the moment of capture -- only to be forced to choose his/her point of focus on every single exposure in postproduction? The average consumer doesn't want to do postproduction in the first place.

Selective focus is of interest to pros and bokeh geeks, neither of whom will be attracted to this device.

I'm not too sure about either concept (focus after the fact or stills pulled from video, folks have been predicting the latter from digital day one and it hasn't happened yet), but this does point to one way that film will survive - conceptual and technical process is still important to art. A print made by hand from the photographer's vision will have a certain gallery cachet that a still plucked from ten hours of footage or a Lyotro shot won't.

Miserere echoes my thoughts. The price point is on the high side for the market it's aimed at, it has no flash, it's not ergonomic, it's not inuitive, its... not anything really but a badly designed camera with one really cool trick. And it's a trick that until it gets in the hands of reviewers that we won't know how well it works and how it will effect image quality.

Not specifically about this camera, but an observation about your commentary: isn't that what camera and software makers have been promising for years: "making fewer demands on the photographer's decision making" with features like autofocus, autoexposure, autobracketing, IS/VR, HDR, etc.? This seems to me like a natural progression of that. We can all get off at our own selected stops (my stop was somewhere north of AF, south of HDR, I think), while others continue to roll with the changes, I suppose!

I have absolutely zero interest. I'll take an iPhone 4S please.

If we keep eliminating the technical decisions, then the only thing left for the photographer to do is to decide what he whats to photograph, stand in the right spot and press the button.

And after all, aren't those the three things that are the most important in photography?

When I looked at the samples they provided, it seemed that the price they pay for being able to focus anywhere is the loss of sharpness everywhere...

It seems like the Lytro allows the photographer pretty much zero control.

Nonetheless, I wonder if this technology could be used to create photos with unlimited depth of field using focus stacking. Stiched together, even the Lytro may be able to produce final images with passable resolution.

Cool technology, premature product.

If every smartphone doesn't have a plenoptic "light field" camera by 2016, I'll eat my hat.

I sincerely hope I can watch you eat your hat. I think this is a step backward to the fixed focus cheap disposable type camera.
It will fail for a number of reasons:
- the form factor, as mentioned, while funky, is not practical.
- the target market (non-photographers) don't want to do any post-processing. Auto focus suites this demographic much better, especially with things like face recognition.
- the price is too high for a gimmick
- the resolution is too low. It may be fine for facebook, but even as a full screen desktop image, it is going to be weak.

I may be wrong, but if this becomes mainstream I will eat my hat....

What I noticed in that demo as well as the first time I saw the examples on the Lytro site is that there doesn't seem to be a way to get everything in focus. IOW, you can change focus with shallow DOF, but can't get deep DOF. And "everything sharp" is what amateurs mean when they say "focus." And again, cellphones and your average tiny-sensor pocket cam are already focus-free enough for them so I'm not sure who the market for this is.

I don't know about this camera in particular, but the technology seems like it would be very useful for video stuff. Instead of adding all those rack-focus gizmos to your DSLR for focus pulling during shooting, you just shoot away and then set your focus in the comfort of the editing suite.

"...have you noticed that every innovation these days seems to be directed at making fewer demands on the photographer's decision making?"

You might actually say that some of the more recent advances have been focused on delaying the photographer's decision making, as opposed to removing it. For instance, setting white balance in RAW after the fact frees you up from worrying about it in the field, so you can make that decision later.

In the same way, plenoptic cameras may be able to free you from making focus decisions in the field, since it can be made later. Ultimately, maybe this frees you up to focus on the primary things about taking photos - where and when. (I.e., where is the camera located and when do you trip the shutter.)

Eventually, due to video frame grabs, the "when" is potentially going to get blurrier and be decided after the fact too. But I think we're a long way from that being the norm yet. (Although the number of pictures made this way will rise, there are lots of situations where current video streams aren't suitable for taking photos from. For instance, very fast or very short shutter speeds.)

I guess the ultimate would be a plenoptic video camera shooting at least 4K video in a RAW format at very fast shutter speeds and high frame rate. Then you'd only have to make choices in cases where you wanted a slow shutter - just about anything else could be a video frame.

Tarnation sonny! The whole process of 'real' photography started going downhill with the introduction of the gelatin dry plate in 1871. I mean really. If you can't coat your own glass plates how lazy can you be.

just wondering: is this little thing shooting wide open all the time, i.e. with f/2.0?

Drawing has never been the same since the invention of the pencil. No real artist would ever touch one. There are so few of us left now who prefer to bang rocks together.

I quite like the form of the Lytro camera but not for the way it was demonstrated. Give it a rubber eyepiece and you've got a telescope you can take pictures with. A binocular version would be pretty neat too, even without the light field technology.(or perhaps especially without it if there is a substantial resolution hit)

I wonder if they will make a monochrome sensor for it...

Does not look to me like it takes a much better photo than a pinhole camera, which at least has the benefit of being cheap.
Think I'll stick to my Rolleiflex, which has the added benefit of looking gorgeous, unlike this thing.

I was under the impression that I was reading The Online Photographer, not The Luddite Photographer.

When I first watched the demo vidoes last night, and played with the online software demo, I was fascinated by the possiblitiies; not for my own still photography (I don't yet have clue how I'd use it, or if I'll even want to), but for scientific, medical, military, law-enforcement, geeky-hobbiest, video, photojournalist applications, etc.

It's a first generation product, and regardless of how clumsy or ergonomically challenged it might appear to be, I think it's pretty remarkable that the first commercial iteration of technology this innovative can be put into such a small package and sold so cheaply.

Give it a break. This device is a proof of concept, very wisely aimed at a segment of the population that can afford to finance further development of the technology. By enabling and emphasizing its social networking potential, the company is going to be able to move ahead.

I look forward to reading the comments a few years down the road when more mature offshoots of this same technology make their way into "serious" cameras.

I happened to trade one of my digital cameras for a Nikon F5 to replace my long-dead SLR. The F5 is so removed from my memory of shooting film it's embarrassing. Its designers built an idiot proof machine which is smarter than the person who uses it and insures that the photographs it makes are correct (and thus similar to most others). Alas, the whole joy of chemical photography is the soft, organic error margins. I want my film camera to be an idiot, not idiot proof. It dawned on me that the F5 is a very big, very sophisticated point and shoot camera. Is that what "Pro" really means? Operation rather than creation?
The plenoptic camera is an extension of the "Pro" philosophy. It's another victory to technology over human prowess, skill, art. The wider the field of possibility the narrower the choices made. Yes, it may be useful to most (well perhaps not in its first iteration), but it will make genuine seeing, singular expression and true creativity even scarcer than it is now.

What is the actual size of the picture?

Thom said one thing and you said another. I wonder it is not as the camera capture 11M ray and manipulate the ray information (phase etc.) afterwards. Hence, it should 11M sample points throughout, shouldn't it?

Not sure about the exposure, it seems all well exposed. Is the ray (light wave) manipulation can handle the exposure as well?

As regards to hat eating, I think it is the closed system that would kill it. Mind to pose the video when you are eating one. (I suggest a tiny chocolate one would do. Next time, remember to bet on something easier like I would buy you a cake of coke in person. )

These cameras are just toys, but the technology looks very promising. I hope it will be somehow implemented in 3D video and will allow viewers to focus their eyes anywhere in the scene, not just where the camera was focused. This would make 3D video look like reality, not like a stack of flat layers as it looks now.

Better get out the hot sauce, Mike. As Alan pointed out, consumers don't want to do post-processing. "Hey, watch what I can do with this shot", will wear off in a few weeks, and the Lytro will join Grandpa's Stereo Realist in the closet, beside Dad's 8-track tape player, and soon to be joined by several pairs of 3D TV glasses.

When can I get a camera the automatically points itself to the perfect subject with the perfect composition?

@mark lacey: You know what's going to make you _really_ depressed? In 5 years this camera will probably take better photos (and I'm talking image quality, not just the plenoptic extras) than the Rolleiflex. They could possibly do it even now with a medium-format sensor and an astronomical price. And it's small enough that you can stick it inside a Rollei shell and it'll look the same as your camera :).

Let's get this straight: this camera does NOT take your picture for you. You have all the creative control you always had, and the photograph is a true depiction of the scene you saw. But it does free you a little more to concentrate on, as Joe Lipka says, the really important things in photography: what, where and when to shoot.

Art isn't craft, regardless of how long one has spent perfecting focus-by-feel on one's Leica.

You don't need to focus when you take a picture, you just need to focus when you look at it. Er what?

Sorry, this is about the dumbest consumer product I have ever seen in my entire life.

Photographers are not struggling to learn how to focus (dude) they are struggling with lack of spatial awareness. How does this help?

The tech may be useful in other areas for sure. But in the consumer market this is solving a problem noone has and sacrificing features everyone demands. How does that work?

Why is everyone taking this seriously? I mean, really!

Seb wonders: "just wondering: is this little thing shooting wide open all the time, i.e. with f/2.0?"

It is indeed. The DOF is all in the computation so the lens is a constant aperture f/2 zoom. No AF. No aperture diaphragm. No shutter.

There is a good cutaway on dpreview


Most of that barrel is lens: all of the non-textured bit. That why the camera is shaped as it is and doesn't look like a "mint tin".

Now if someone put a constant aperture f/2 28-90 on a fixed lens m43 camera they could make a lot of people happy.

And I concur with the dry plate comment above. It's been downhill since then.

New technology, bah, humbug!

"Nothing is more directly opposed to the ordinary image of artistic creation than the activity of the amateur photographer, who often demands that his camera should perform the greatest possible number of operations for him, identifying the degree of sophistication of the apparatus that he uses with its degree of automatism."

-- Pierre Bourdieu, Photography: A Middle-Brow Art

...this was written in 1965. Looks like nothing much has changed.

A plenoptic or synthetic aperture camera with a f/2.0 lens isn't so interesting. A camera with a f/0.1 lens would be a lot more interesting.

I was really hoping that the Lytro camera was going to be a lens array camera, not a single big lens with an array of apertures camera.

Of course if the single lens were really big that would be cool , but I don't think that anyone besides me would try to use a camera with a 30 pound lens on it.

I hope someone buys these, because I really want a synthetic aperture camera, but I want at least a 150mm aperture on a focal length somewhere between 20mm and 50mm.

So didn't read all the comments, apologies if someone brought this up, but why in the world would you want this or need this in a cellphone camera? Hard to take a picture with a cellphone camera that doesn't have everything perfectly sharp. I suppose you could use it to simulate a larger format sensor with shallower depth of field but do consumers really want this? Personally I don't think so.

This is a unique product with some novel uses, but I'm hard pressed to imagine any mass market appeal for it.

Looks like another camera that puts the fun back pixel peepers need not apply definitely one for the pixies and fairies gotta love it

I wonder who he sell this to. Not the pro or amateur photographer who likes his (D)SLR or bigger to be complex, his skills to be noted and his pictures to be admired. To me this is a nice piece of kit for the gadget freak and its design is directed to the gadget freak eather. The same old Steve Jobs one button mouse philosophy surrounds this "camera". BUt having said that if this technology comes of age, it sure has potential. It's like a Sony Mavica of the crashing 10th. The question is wether the world is ready to embrase a new technology, time will tell.

I don't know about the technology in the Lytro, but am interested in the optics. 25 years ago in the British Army I was issued with a set of binoculars that had no focus wheel, but somehow those binoculars could focus sharply at any distance from about 10 yards out to infinity. A bright and clear field of view as well, with no apparent aberration. Good binos, and they are still being issued. I lost a set on an exercise once, and was charged about £400 from my salary to pay for the loss, so the cost of manufacture is reasonable. Why don't we have optics like that in photography?

I don't want to repeat the infamous statement and say this is a solution looking for a problem, but really, what is this for?

One, as Bryan Davidson said up there, people don't think about focus at all. They dump their several hundred shots on the innocent public without a second thought about the technical or artistic quality of the shots. So why would they start fiddling with individual shots now?

Two, alex, the eye-flick video focus is still in the realm of science fiction. Theoretically, an interesting idea, but out of the reach of the computing abilities at the moment. What you see demonstrated is one photo. Now imagine that repeated 24 or 30 or even 60 times in a second to get on-the-fly focus selection. That would take some serious computing power, which we still don't have readily available. I don't even know what kind of power is needed, to create a comparison.

On the plus side, shallow DOF is considered much more artistic than the deep DOF on compact cams and phone-cams.

But again, face-recognition, auto focus, intelligent auto mode, artistic filters and such have had success because they make it possible to have immediate good results. Why would something calling for additional effort be successful then? I don't think that "artistic" is enough of a reason.

The first comment to this post (by Miserere) nails it: the camera will fail big, but the technology will be incorporated in other devices.

I would go a step further and guess that Lytro doesn't care about the failure of their camera. It is probably meant as an attention grabber. A mere curiosity and short-lived lifestyle accessory for the users, but a great platform to make the technology known. Provided Mike's hat is safe Lytro should be doing real money in a few years with licensing and use it's own camera to bridge the liquidity gap between now and then.

Lytro is probably trying to follow the example of the Fraunhofer institute which is credited with the invention of the MP3 compression. Wikipedia notes that the licence for the MP3 compression algorithm generated about € 100,000,000 in revenue for the Fraunhofer society in 2005.

The Lytro camera might produce fuzzy pictures, but the company has a very sharp focus on the bottom line (pun intended). They are betting high.

I am a bit afraid.
Imagine a camera that would capture all possible information in a micro-second (all possible apertures, all possible iso, all possible focus, etc.). This will exists one day and this will clearly lead to perfect pictures, but this will be a nightmare for photographers.
I can't imagine having pleasure to take pictures with such a perfect camera. One interest of taking pictures is to capture scenes that cannot be easily captured.

Kudos. This represents the dawn of another way to capture visual information.

I struggle between Meh and that is really, really cool.

As then, I cannot be specific, but, I get the same feeling as when I first saw the Ipod. This is likely going to be big in ways we cannot even imagine. The only thing that seems strange is that Apple is not involved.

I'll probably be dead before I give a crap about these types of cameras. So this will have to become a JOYFUL NUDES visit to TOP.

I disagree with the comments about the design of this thing being a mistake; departing, completely, from the traditional idea of what a camera is supposed to look and feel like may be one of the wisest moves possible. Then again, when I think about it, it's kind of a return, on a much smaller scale, to the real traditional camera.

Imagine you are 60 years old, or older, and your main concern is taking photos of the grandkids. I'm getting one, even if it won't be the only camera in my bag. I know what my mom is getting for her next birthday.

It seems that the comments so far are split along the same lines as the Raw vs. JPEG arguments.

Shooting Raw, the photographer moves contrast, H&D curve, white balance, saturation, exposure, black point, demosaicing/sharpening/noise-reduction, local contrast etc. into post production.

With JPEG the photographer more or less says "screw it , let the camera sort it all out".

Plenoptic photography seems like a move in the direction of raw , where there is more creative potential left in the digital file than if the camera had made all the decisions.

I'm curious what information is in the Lytro file, and whether the exposure information is fixed or not.

Count me among the skeptics, but mainly because I like to think of photography being about optics and materials as well as the resulting image. Optics and materials can be learned, understood, and controlled, where as software can't be -- at least not in the same way. With software and computer systems you're always at the mercy of some product manager who might decide to change how the product works. You do a "required" update, and whammo, everything you thought you understood has changed.

I'm no luddite. I've been working in the software industry for 20 years, which is why I have little faith or trust in software. (When you have an insider's view of just how fickle and shaky most software products are, and the kind of bad thinking that often goes into its development, you really appreciate the analog world.)

That said, I can see how this thing might have some specialized uses, as others have said. But I'm not going to trade in my GF1 and its lovely interchangeable optics for this thing.

Not to contradict what I just said, but I will admit to being curious about technological things, even if I don't trust them. One thing that would make this technology a LOT more interesting is if there were an adjustable focal plane. As in, not just forward and back, but if you could adjust for tilts and shifts like a view camera does. I can imagine a nice 3D interface where you can visualize the focal plane -- including DOF -- and can adjust them with a series of sliders. That would be cool, especially if it were matched with a decent lens and a large sensor.

I certainly wish them well; and the technology is awesome, and has loads of practical applications, it seems to me.

I'm doubtful that their current product will fly in the mass market (but maybe that's not what they're intending?). The mass market is notoriously averse to post-processing. And for that matter doesn't have a Mac. And is moving away from separate phones to using cell-phone cameras.

The camera form-factor is interesting; I don't know if it's dictated by needing extreme depth for the light-field sensor, or just a design choice. It looks like a "good" kind of different, anyway; worth a try. I haven't yet found the zoom buttons in the pictures :-) .

I think Richard's featured comment is seriously underestimating the potential of this technology. In theory, given the right algorithms and matching parameters, the data from a captured light field can be rendered in the style — bokeh, spherical aberration, and so on — of any real world lens. Or of any imagined lens. And that certainly will happen, maybe not in the next year but in the next decade. Then the question of style becomes not what equipment matches a personal style, but whether the software is sophisticated enough to represent the style you imagine.

@David Bostedo — you don't need slow shutter, either. You capture continuous light field data over time, and then decide later whether to use a ¹⁄₅₀₀th second single frame, or to integrate over time to create a 10-second exposure. So, that decision too can be delayed until later, just like focus and depth of field.

Hmm. The first that went through my mind after seeing this was "Segway". Another massively overhyped niche product. The second was, they've missed Christmas. Unless it's to spoil someone else's similar product they should have left it 9 months so they could deliver stock early October.

As to the product;

It's not fun enough - the focus effect is interesting for about 5 minutes and then gets old.

There's no way to apply hipstamatic style filters, which actually are fun, they need to address this.

The sample images are so low-res even the sharp bits aren't as sharp as shots from my iPhone 3GS, which ain't cutting edge exactly.

Doesn't shoot video.

From a marketing standpoint, it's a fail as joe public still thinks megapixels are the only measure of quality, there's a bit of desperation in the "11 MegaRays" it looks like they've picked a number that sounds OK in megapixels and worked back.

And to cap it all, it's not even got visual appeal. It looks cheap, pig ugly and seems like an ergonomic disaster, which could easily have been solved with a few curves and a flip out screen.

Is it only me that thinks it looks like a slide viewer?

Less time shooting, more time sitting on your rear making photographic decisions while sitting in front of a computer?

Sounds utterly miserable to me.

The attention to the consumer's perceived need for focus options after the fact makes me wonder about the breakdown of the camera industry. For purposes of discussion, I'd break down camera buyers into two groups. 1: photographers who understand the operation of their camera as a system from shutter trigger to output, and make purchase decisions that reflect same (i.e. informed decisions about features/technology involved, and have the skill to properly use the equipment. 2: Everyone else that wants to take pictures for any reason, including non-photographers who want to take 'good pictures' but can't necessarily articulate what it is that makes a 'good picture' or why one class of camera may be better suited to them than another (and frankly, cameras within comparable classes are probably fungible for nearly all photographers).

I suspect group 2 is a minority of >95% (frankly, I'd have said 99%, but that's a loaded statistic (politically) these days).

I don't want to hazard a guess as to whether Plenoptic technology (or the Lytro specifically) will succeed in the market. I just don't have a good track record for guessing acceptance of new technology. If I don't predict, I won't be wrong.

I do think the Mac OS only option is odd. Presumably, Lytro must think product acceptance will be at a sufficiently higher rate among Mac owners than would be for Windows owners, since the two architectures are essentially equal (I mean hardware/ISA) and cross-platform development would be a minor portion of the total project development cost. I can't think of a single programmatic advantage the Apple development environment has over Windows that couldn't be duplicated with the application of more software developers.

What I do think is that we are at the beginning of a new phase of explosive growth in camera functions and capabilities, akin to the dawn of AE or AF. But much, much bigger. It amazes me how little camera makers seem to understand the pent-up demand for in camera image manipulation made available by camera phones. Not just art filters, but true manipulations (including workflows). Why cameras don't have wifi (or something like Bluetooth3) built in boggles my mind. NFC (near field communication) would enable simple, near instant, secure pairing between camera and phone, and offer high speed, low power consumption. I WANT THIS NOW (and I doubt I'm alone)!

Folks, cameras are changing, and going to accelerate how different they are than current cameras. Image processing like Plenoptic is just the beginning of the deluge.

Me, I just want to take good pictures and get them out of my camera (g)!

I think CMS is right — this product isn't really meant to be a big success in and of itself. It's a technology-launch publicity stunt. And _that_ answers something I was puzzling about. Surely, for about 10x the price, they could have made a version which might actually be practically useful for some interested and ambitious photographers. But it's not really about that. This is, instead, your chance to own a piece of history. This tech is the future, and you can buy a piece of history right now, while it's still right on the edge of being science fiction.

Does it blow highlights?

My gut feeling on the refocusing is that the camera is essentially doing the same thing as a HDR file (NOT a tone mapped gimmick photo). That is, when you press the shutter it adjusts focus to different points of interest (or just pre-defined distances) and takes a bunch of photos which are then stored in a layered format file (ala HDR format). The "refocus" ability would then be just selecting the appropriate layer.

Otherwise the way this technology is being represented should be capable of adjusting not just focus but depth of field too. Since its a set aperture though and very low resolution it makes more sense for it to just take a series of shots (eg 0.5s worth of video essentially) and move the focus throughout. The display applet just then transitions or interpolates between the shots as you choose.

As I said, that's my gut feeling on this miracle (snake oil) breakthrough.

No, that's not it. I think the best simple explanation of a plenoptic camera is still the short video prepared with the original Stanford paper:


This might help you get a handle on what's going on.


Dear Mike,

DDB and I discussed this at some length and it puzzles me on several points.

This is clearly targeted at the amateur (as opposed to serious hobbyist) market. Not even close to the pro market. The design form, the (lack of) controls, the fairly low resolution (roughly comparable to 1-2 Mpx Bayer array, most likely), all pull it away from the "serious" user. Not counting the gimmick thing, ala pinholes, Dianas, etc. No doubt some pro somewhere will figure out how to love this. But en masse, not so much.

But then, the marketing seems all wrong and the price point absurd. They probably can't do anything about the price at this point, but it's way too expensive for the cheapo market. And the marketing doesn't emphasize the right things. What would be most likely to make this camera a market win, if a win were possible (I have serious doubts) is to emphasize how it does right the things that current cheap cameras do wrong. No shutter lag-- that's actually a biggie, if you're trying to photograph kids, pets, or animated adults. Cell phones have great depth of field, but lousy lag times.

I'd have a web page along the lines of "Does your camera make photos like this?"... "Ours makes photos like THIS!" with egregious examples of missing the critical moment on the left, and corresponding Lytros wins on the right.

I'd be emphasizing the lack of controls-- "Does the back of your camera look like this?" (lotsa buttons and switches)... "Ours looks like this!"

Like the incredibly successful iMac TV ads that told the viewer how they'd get onto the Internet in only three steps with a Mac:

1) Plug it in and turn it on.

2) Launch the browser.

3) There is no step three.

(This was back when self-configuring network utilities were LOTS less common).

I'd be showing some photos that were sharp EVERYWHERE. It's gotta be able to do it. Don't show people just that the focus can be shifted, show them that it's always right... unless you want it not to be.

The underlying technology is marvelous, the product placement seems misconceived.

pax / Ctein
-- Ctein's Online Gallery http://ctein.com
-- Digital Restorations http://photo-repair.com

One of the more puzzling criticisms I keep seeing is that this camera will fail because most people don't want to do post processing. So... What post processing are you guys referring to?

Do you mean adjusting the focus? The photographer doesn't even do that. The viewer does it. It's interactive viewing -- clicking (or tapping) on a spot to set the focus, or double-clicking to zoom. That's not what I call post processing. The result is that the photographer never has to worry about focus: not when taking the picture, and not afterwards. That is, not unless he wants to down-convert it to a JPEG... But why do that?

The other thing that occurs to me is that we now have a use for all those megapixels that manufacturers keep piling into their sensors. They were starting to become pretty useless for ordinary people taking everyday snapshots -- merely a way to consume storage space. (I thought my old 2.1 MP Elph was Just Fine for snapshots -- better than any 110 film camera I ever had, for sure, and more pocketable too.) So now we have a technology that can take all those millions of sensor points and do something new and interesting with them. Seems like a good idea.


Thanks for the link, that makes better sense than the hype I've seen around Lytro. The PDF/video are what I was expecting, in that you can change the depth of field to have the entire image in sharp focus, whereas Lytro gives the same apparent DOF. Which to me means the same effect can be achieved by taking a short video, moving the focus and then using the software to select the frame with the sharpest pixels in the selected area of interest.

Being able to adjust the DOF and to some extent move the observer is an interesting and more (I feel) powerful feature of LF technology. Perhaps Lytro haven't been able to fully make use of the information captured to allow this with the first product.

Something that didn't occur to me (and coming from a 3D background it should have) is that it is a synthetic (rendered) photo. So in light of recent posts does this make LF images a photograph, art or something inbetween?

Looks like the Kazoo of cameras.

Clicking to make one spot sharp and then another is playing. The interest won't last past the first 30 seconds. What is necessary, and I hope possible with that Mac-only host software they say comes with the camera, is to process the image to a normal jpeg that's sharp throughout. At that point, you've got a useful picture that you can use anywhere, and that will look better than your sister's pictures of her kids and cat. But to get there is post-processing, which is anathema to that class of user.

And they don't actually say their software can easily do that (it's obviously possible from the underlying technology, of course).

Limited DOF is an enthusiast / artist interest, i.e. .01% of the market. What the actual market wants is for everything to be sharp, or at least for the entire subject to be sharp. And also imaged at the moment you saw, rather than another moment a while later. The underlying tech can, I think, give that -- but that's not what they're marketing. Looks crazy to me.

Two more comments; I can't resist.

1) I've seen several people say "limited DOF is only of interest to artistic enthusiasts", and estimate that to be a minuscule fraction of the market. Maybe, but does that really explain the millions of people buying entry-level DSLRs and flooding forums and flickr with questions about bokeh and "great depth of field" and whether to buy a "nifty-fifty"? If it's a fraction, I don't think it's one to be underestimated.

2) If they could set up a service to sell high-quality holographic prints, that'd be a pretty nifty niche.

I think the comment about how photography may consist, even more than it already does, of sitting in front of a screen mucking around with some image manipulation tool is important, though perhaps not in the way it was intended. I'm probably ahead of most people, but at this point I have spent pretty much every working day of my life for the last twenty-five years doing just that, and I really would like to spend a lot less time doing it in future, because it has got just a little bit dull by now. If photography has to involve doing more of that, than I will find something else to do.

Matthew, I'm talking off the top of my head about impressions, so you have every right to prefer your own impressions :-) . I would point out that, on the one hand, there are new enthusiasts joining the ranks every week, and that may well be enough to explain the obvious new enthusiasts showing up in forums (I do agree that they very clearly are new enthusiasts showing up with new DSLRs pretty frequently). And that the number of DSLRs sold is far greater than the number of people showing up in such forums (isn't it? They're scattered and don't publish reliable figures, so I'm really guessing here).

And Lytro will be competing against the Nikon 1, which may have the delay problem solved (at least compared to entry DSLRs; a D3 is probably still faster) and may have a small enough sensor for DOF to be not too bad a problem. But is more expensive.

Well, the technology is fascinating and will have all sorts of down-stream uses, regardless of how this product and this company fare.

The comments to this entry are closed.



Blog powered by Typepad
Member since 06/2007