New "No Focus" Camera?

[JDDavis]

http://www.lytro.com/cameras

So I just ran into this today. Not sure how I feel about it. The company is said to be releasing a camera and software that you will not have to focus (or select a focal point). In post you will be able to change the focus to anything you want. They're calling it a light field camera. For such a revolution in camera technology they kind of have a crappy website. At least it's not working in IE.

I'm sure many people decried the DSLR as killing art and skill in photography but I'm not sure I want a camera with witch I can never make a mistake focus wise. I suppose the subjective art part is moved to post processing where you select your focal point (or leave everything in focus). It would be great to have in some situations I suppose. There is no real info on the camera so I don't know if it's geared towards SLR users or compact users.

It will be interesting to watch though and see exactly what this is.

 

[firestarter]

Colour me doubtful.

If each sensor pixel records light direction as well as colour and intensity, then:

• It's going to have to store a whole lot more data than a conventional imager - more processing, more storage.
• It's going to have worse noise than a conventional imager (since you'll presumably have to split each pixel into smaller direction-sensitive sub pixels) unless you have bigger sensors
• You'll need to shoot with fast wide aperture lenses, wide open, to get the 'light direction' stuff to work.
• All this spells 'expensive'

It's a solution without a problem. At the low end, noone cares about focus since they use small aperture P&S cameras with wide depth of field. At the high end, DSLR shooters know what they're focussing on when they use selective focus.

What consumer level photographer would want to spend a lot of time picking focus point back home anyway?

At best, I can see this going the same way as Sigma's Foveon sensor... something that looks good on paper, but fails to be implemented in any market-changing way.

 

[Phrasikleia]

I'm not sure I want a camera with witch I can never make a mistake focus wise.

Why not?? Do you prefer to shoot JPEGs because you like being able to make white balance mistakes?

I welcome anything that makes the technical side of photography easier. There is plenty to master in the realms of timing and creativity to offer a lifetime of challenges. Focusing in the field is one of those tedious aspects of photography that I could easily do without.

More info here, by the way: http://www.nytimes.com/2011/06/22/technology/22camera.html?_r=2&pagewanted=1

 

[mackmgg]

Colour me doubtful.

If each sensor pixel records light direction as well as colour and intensity, then:

• It's going to have to store a whole lot more data than a conventional imager - more processing, more storage.
• It's going to have worse noise than a conventional imager (since you'll presumably have to split each pixel into smaller direction-sensitive sub pixels) unless you have bigger sensors

There's another option you're missing: Less pixels. The sensor won't be huge, but it probably will only be around 5 or 6 megapixels.

• You'll need to shoot with fast wide aperture lenses, wide open, to get the 'light direction' stuff to work.
• All this spells 'expensive'

It's a solution without a problem. At the low end, noone cares about focus since they use small aperture P&S cameras with wide depth of field. At the high end, DSLR shooters know what they're focussing on when they use selective focus.

What consumer level photographer would want to spend a lot of time picking focus point back home anyway?

At best, I can see this going the same way as Sigma's Foveon sensor... something that looks good on paper, but fails to be implemented in any market-changing way.

I think it's aimed for consumers. They will probably use software for autofocus later, but there's no delay for focusing when you're actually taking the photo. I know I've had one occasion where a point and shoot had focused incorrectly. Also, autofocus is slow on point and shoots now, something this would fix

 

[firestarter]

There's another option you're missing: Less pixels. The sensor won't be huge, but it probably will only be around 5 or 6 megapixels.

How many different 'direction sensing' elements do they need per pixel? 4? 8? 16?

Presumably you'd have different directions sensed by neighbouring pixels and interpolate.

I still think 6 megapixels is high in this context... if you sensed 8 directions that would multiply up to 48 million sensors in total.

 

[JDDavis]

Why not?? Do you prefer to shoot JPEGs because you like being able to make white balance mistakes?

I welcome anything that makes the technical side of photography easier. There is plenty to master in the realms of timing and creativity to offer a lifetime of challenges. Focusing in the field is one of those tedious aspects of photography that I could easily do without.

More info here, by the way: http://www.nytimes.com/2011/06/22/technology/22camera.html?_r=2&pagewanted=1

Thanks for the link, it explains it a little better.

Very true...I shoot RAW because I want to be able to fix not getting exposure just right. Valid point. And it's a valid point that an uninteresting picture is still uninteresting even if it's perfectly in focus.

Do you feel that selecting a focal point in the field in order to convey something in particular with the image is a technical step and the "art" part of it could be as easily done in post by selecting the focal point with software? I guess I see both sides of it. There are times, especially in my climbing, that I simply want to grab the image in front of me quickly and not worry about my "artistic interpretation" of what I saw till later. Shooting in RAW helps that and this technology would as well. I'm still intriqued about all the manual and technical parts of photography as well though. It's one of the things that got me interested in it. (Even if it is all digital)

Like I said, it will be interesting to see how this technology works in the wild and how it is implemented.

It's funny though...I think about how much of a pain (ironically) it can be post processing a batch of pics when they are all good. Like when I have used the high FPS feature of my camera to try to capture a moment. It's sometimes easier when I can toss all the out of focus or poorly exposed shots right of the bat and be left with the few good ones. Doesn't say much for me as a photographer. I'd probably rather take fewer pictures and have a higher percentage of keepers. Just an observation.

 

[Phrasikleia]

Do you feel that selecting a focal point in the field in order to convey something in particular with the image is a technical step and the "art" part of it could be as easily done in post by selecting the focal point with software?

I don't think it matters how or when you achieve a certain result; in my view, it's the final image that matters. I do tend to shoot with an idea or visualization in mind at the moment of capture, but I love the flexibility that post-production allows.

All technology starts in baby steps, and I presume it's just a matter of time before all digital cameras are focus-free. Focusing is the source of so much hand-wringing in photography these days. There is so much fuss over focus points--how many a camera has, whether or not they are cross-type, how far they are spread across the image area, whether or not they have assist points, and how they can be selected quickly. There is so much chest-beating over cameras that can track a moving object well. So much concern over whether or not a given camera can auto-focus a given lens or not. Imagine if all of those concerns were to go away. It may not happen in my lifetime, but this new Lytro technology seems to be a step in that direction. I imagine that step will be followed by a stampede eventually.

 

[snberk103]

While I agree that technology that makes taking pictures easier is a good thing, I can foresee four downsides.

1) I predict a lot of boring pictures. Where the photographer can't be bothered to go through every single image in PP to pick the best focus, so they just make every image fully focussed.
2) I predict we will need huge amounts of new storage. I played with some of examples on the Lytos website, and they took longer to load than equivalent non-Light Field images.
3) So much for that bit of my classes where I teach people how to control their DoF. sigh.
4) Sitting around a laptop or computer screen for hours watching someone with a Light Field Camera show you all the possible variations for each of 168 images they shot on a weekend trip.

 

[MisterMe]

Colour me doubtful.

...

Yes, but.... I am impressed by what I see on the Lytro website. The company's explanation reads like hokum, but the pictures are spectacular. If this is real, then it is a revolution.

 

[Ruahrc]

Well the major downside right now is the enormous resolution hit you get. Using a 3x3 array of pixels per shot, you are losing 9 times the resolution. A more optimal capture solution would involve 4x4 or 5x5 which are 16 and 25-fold losses. It scales up from there. I think in this guy's thesis, he was using a 50MP medium format camera and was producing final images of about 0.8MP.

Personally, it still seems more like a solution looking for a problem. Focusing technology is pretty mature at this point, and this method seems to carry a significant penalty for the convenience of touching up your focus in PP.

Perhaps there is some merit in that you are capturing data which can probably used to reconstruct 3D images, or maybe something like macro, where a single shot would be the equivalent of focus stacking, but again, by nature of the design, the resolution drawbacks will always be pretty severe. Would a 200-400MP+ APS or 35mm "light field" sensor still suffer from diffraction limitations? There might be merit in the technique for scientific applications as well, where artistic merit is not important compared to data capture, but that's pretty tangential to lytro.

And just from curiosity why is this one guy's site getting so much buzz? Light field capture is nothing new (demonstrated by Adobe several years ago at siggraph, and existed long before that) and it doesn't look like this guy's approach is any different than what has already been done. It's easy to claim you'll have a consumer-oriented camera available for a few hundred dollars "later this year", a lot harder to back that claim up.

 

[tinman0]

Think of the downside slightly differently.

What is this new technology actually solving? It's solving focus.

Which is a problem that has already been largely solved, and is still actively being improved by all the big names in photography. So, it's solving a solution that already has a solution, that costs a darn site less.

So this is a technology that in this form isn't going anywhere quickly.

 

[jtara]

Colour me doubtful.
It's a solution without a problem.

No, there's definitely a problem it solves. In fact, it solves plenty of problems.

But very specific and technical ones. For example, in microscopy depth of field can be VERY short, making it impossible to focus on an entire subject. OK, so you take a bunch of pictures with different focus. But what if it's a moving subject?

I just don't understand why they are coming out with a consumer product. It doesn't make much sense in that context, except as a novelty.

 

[Flynnstone]

pinhole camera

 

[snberk103]

Yes, but.... I am impressed by what I see on the Lytro website. The company's explanation reads like hokum, but the pictures are spectacular. If this is real, then it is a revolution.

I don't know about revolutionary, but impressive yes. I don't think it's going to change traditional photography much. At least not the stuff that ends up on paper. But it very well may lead to a different kind of photography meant to be viewed only on screen. These 'non-static' images could either be ones that viewer interacts with, or images that moves the focus around on their own according to the photographer's pre-sets. Sort of like a mini-movie where the viewers eyes are moved through the image as the focus changes. Ok, maybe it is a little it revolutionary.

Well the major downside right now is the enormous resolution hit you get. Using a 3x3 array of pixels per shot, you are losing 9 times the resolution. A more optimal capture solution would involve 4x4 or 5x5 which are 16 and 25-fold losses. It scales up from there. I think in this guy's thesis, he was using a 50MP medium format camera and was producing final images of about 0.8MP.
...

Simple technical hurdles. So, yes significant.... but solvable. Moore's Law would suggest that the resolution has (more or less doubled) since their first prototypes, and will double again by late next year. Sorta Kinda. I'm just saying that the technical limits on the resolution will be solved in a very short time.

...And just from curiosity why is this one guy's site getting so much buzz? Light field capture is nothing new (demonstrated by Adobe several years ago at siggraph, and existed long before that) and it doesn't look like this guy's approach is any different than what has already been done. It's easy to claim you'll have a consumer-oriented camera available for a few hundred dollars "later this year", a lot harder to back that claim up.

To answer your question.... Good publicist. It may mean that they are ramping up and needing money to actually build the thing. If they get the funding, we can expect the early versions of the cameras soon. If they don't get their investment funding, this will be a thread that gets picked up in 2 years when someone doesn't check the date, and we'll have a "remember when?" & "Whatever happened to?" discussion.

Think of the downside slightly differently.

What is this new technology actually solving? It's solving focus.

Which is a problem that has already been largely solved, and is still actively being improved by all the big names in photography. So, it's solving a solution that already has a solution, that costs a darn site less.

So this is a technology that in this form isn't going anywhere quickly.

It doesn't really solve a problem (at least for you and me... ) as far as printed photos go.... but it does create a new type of photography for on-screen viewing, where the photo does not have to be static.


... all of this is imho, of course ...

 

[mackmgg]

All technology starts in baby steps, and I presume it's just a matter of time before all digital cameras are focus-free. Focusing is the source of so much hand-wringing in photography these days. There is so much fuss over focus points--how many a camera has, whether or not they are cross-type, how far they are spread across the image area, whether or not they have assist points, and how they can be selected quickly. There is so much chest-beating over cameras that can track a moving object well. So much concern over whether or not a given camera can auto-focus a given lens or not. Imagine if all of those concerns were to go away. It may not happen in my lifetime, but this new Lytro technology seems to be a step in that direction. I imagine that step will be followed by a stampede eventually.

I have to disagree with you there. While this will probably be going towards the consumer market soon, I don't think focus-less imaging will ever come to the professional market. The reason is that when the technology is good enough for 10 megapixels of the final image, most professionals would rather have to worry about focusing then and get the full image of a normal sensor with that technology. It's the same as CompactFlash vs SD. SD is ages ahead of where CompactFlash was a few years ago, and it great for the consumer market. But every time SD gets better, CompactFlash gets better as well with the newer technology. So professionals/enthusiasts will almost always go for the more developed technology.

 

[Ruahrc]

Simple technical hurdles. So, yes significant.... but solvable. Moore's Law would suggest that the resolution has (more or less doubled) since their first prototypes, and will double again by late next year. Sorta Kinda. I'm just saying that the technical limits on the resolution will be solved in a very short time.

I'm not entirely sure it is just a brute force Moore's law kind of thing though. First, Moore's law applies to transistor size which has little to do with camera sensors. There are many other aspects much more important to imaging sensors such as pixel pitch, quantum efficiency, dynamic range, etc. Second, we're already enchroaching upon the hard limits of optical diffraction with modern imaging systems. Now I'm not entirely sure how this "light field" imaging is affected by diffraction, but if it follows the ramifications of optical physics in the traditional sense, a 400MP DX-sized sensor is going to be severely diffraction limited. Even a 400MP FX sensor (which would only be producing a 16MP final image, assuming a 5x5 array behind each microlens) would be losing a lot of resolution to diffraction.

 

[JDDavis]

I don't think it matters how or when you achieve a certain result; in my view, it's the final image that matters. I do tend to shoot with an idea or visualization in mind at the moment of capture, but I love the flexibility that post-production allows.

All technology starts in baby steps, and I presume it's just a matter of time before all digital cameras are focus-free. Focusing is the source of so much hand-wringing in photography these days. There is so much fuss over focus points--how many a camera has, whether or not they are cross-type, how far they are spread across the image area, whether or not they have assist points, and how they can be selected quickly. There is so much chest-beating over cameras that can track a moving object well. So much concern over whether or not a given camera can auto-focus a given lens or not. Imagine if all of those concerns were to go away. It may not happen in my lifetime, but this new Lytro technology seems to be a step in that direction. I imagine that step will be followed by a stampede eventually.

Again, great points. I think....maybe....for me I place more value on the artistic effort that's done in the field with the camera, lens, filter, the light and all of that. A bit ironic I guess since I shoot only digital and pump everything thru Aperture and Nik. Maybe I'm just a traditionalist when it suits me. I would never condemn any technology but personally I don't want to stand in a spot in Yosemite, snap the shudder, and then push the Ansel Adams pre select in the post software when I get home and get a perfect pic. Not that I don't want a perfect pic, I just want to earn it. I would never tell someone that did that there picture was somehow less becuase of it either. Each to thier own.

I do think there is a lot of art and skill that goes into digital post production and it would be nice to have a truly nuetral image to start with and turn into your vision. I guess I'm hesitant to welcome the trend of everything becoming automated.

Sorry for turning my comments more towards the philisophical. I just can't hang with the excellent technical discussions that are going on as well.

 

[JDDavis]

I don't think it matters how or when you achieve a certain result; in my view, it's the final image that matters. I do tend to shoot with an idea or visualization in mind at the moment of capture, but I love the flexibility that post-production allows.

Had another deep thought

When does the final image become more the result of the camera / software presets and less the product of the photographers skill and artistry?

I've struggled with that the more I've gotten into photography. My Nik plug ins are fantastic and often times the pre set filters look great. I often adjust them a little to make myself feel better or to make it more my own. At the same time I think why is the polarization filter in Nik any different than putting a polarizer on the lens? I have more options doing it in post.

Just rambling. I realize I'm detouring a good bit from my original post. It's Friday

 

[Chappers]

I don't think this is anything to worry about - a badly framed photo looking down on something instead of getting down to the subject - or just a crap photo - will remain a crap photo - just it will be an in focus crap photo. Admittedly you will be able to chose a point to focus on, but it will still be ....... well you get the idea.

Generally I like the idea.

 

[designguy79]

My biggest question is if it can create dreamy backgrounds like these?

(not mine, just some fine shots I found on flickr!)

http://www.flickr.com/photos/vnhan/5380473504/

http://www.flickr.com/photos/hkdollar/4145203275/

http://www.flickr.com/photos/polvero/3466964233/

 

[Phrasikleia]

I don't want to stand in a spot in Yosemite, snap the shudder, and then push the Ansel Adams pre select in the post software when I get home and get a perfect pic...I guess I'm hesitant to welcome the trend of everything becoming automated.

It's not an issue of automation any more than shooting in raw means you're getting an auto exposure. It's simply one more option you have to adjust in post.

Sorry for turning my comments more towards the philisophical. I just can't hang with the excellent technical discussions that are going on as well.

No apology necessary. Make that two of us.

Had another deep thought

When does the final image become more the result of the camera / software presets and less the product of the photographers skill and artistry?

I've struggled with that the more I've gotten into photography. My Nik plug ins are fantastic and often times the pre set filters look great. I often adjust them a little to make myself feel better or to make it more my own. At the same time I think why is the polarization filter in Nik any different than putting a polarizer on the lens? I have more options doing it in post.

Just rambling. I realize I'm detouring a good bit from my original post. It's Friday

One of my pet subjects, but alas, one that would quickly derail this thread. My short answer, for what it's worth: so long as the photographer is making the decisions and taking responsibility for the results (however they were obtained), the photo is a "product of the photographer's skill and artistry," as you put it.

 

[Ruahrc]

At the same time I think why is the polarization filter in Nik any different than putting a polarizer on the lens? I have more options doing it in post.

Not to get too OT but a "polarization filter" digital effect will never be able to accomplish what a real polarizer does, since a true polarizer actually filters out light based on its polarization. Macroscopically, in each case you get what seem to be "richer colors and deeper skies", but in reality with a CPL you are suppressing the reflections off of objects which just means you are able to better see the true colors underneath (hence it looks "richer").

A good example is shooting a water surface. With a CPL, you can adjust the filter to block out any reflection of the sky, allowing you to see under the water surface. That is an effect that cannot be replicated digitally.

That's why things like CPL and GND filters have uses in the digital photography era. Things like the skylight or tobacco filters can be recreated digitally and are not really needed anymore.

Ruahrc