Shoot first, focus later: Light-field photography

[Marsu42]

"Light-field photography is the future" - or is it?

Here are two reviews on the new Lytro Illum, still seem to be work in progress, but an interesting concept. And you can emulate a f1.0 lens

German: http://www.spiegel.de/netzwelt/gadgets/kameras-lichtfeldkamera-lytro-illum-im-test-a-1000246.html
Foreign: http://www.theverge.com/2014/7/30/5949913/lytro-illum-review

What do you think: Is old-school single-dof photography doomed?

 

[Maximilian]

"Light-field photography is the future" - or is it?
...
What do you think: Is old-school single-dof photography doomed?

It's a very interesting and promissing technology.

But I am sure that every advantage comes with some tradeoff. As this tech is not widely spread yet these tradeoffs are not described or even discovered.
If we know them we can decide what is the better solution.

 

[alexanderferdinand]

Lets see, what the practise shows.

For my POV, using a xxMP sensor, getting information of a certain "lightfield" (always good to use a new term to be cool and different!) and then cut out a slice can only result in a very low portion of the whole information, means notsogood IQ.
Somewhere, must be now 3 or 4 years ago, I read about a sensor around 3 or 6MP, capable to make prints to show online (their words) - didnt sound very promising other then a fun- or lifestyle tool.

I would be surprised if this would be better!

 

[Coldhands]

I expect light-field to follow a path similar to 3D movies. It seems new and exciting but doesn't really add much substance to the fundamental way we experience a photograph. For me, the novelty of manipulating the images in the article from The Verge wore off within half a minute. I also believe that nearly any photographic composition has a single, optimal, focal-point/DOF combination - so then what's the point in recording all others?

It'll probably continue to get quite a bit of buzz but I'd be very surprised to see it supplant "traditional" photography in the long-term . Perhaps it will continue to exist as a niche.

Now I cross my fingers and hope that, 10 years in the future when light-field is ubiquitous, this prediction remains unnoticed, thereby preventing me from being subjected to immense ridicule.

 

[Maximilian]

... It seems new and exciting but doesn't really add much substance ... I also believe that nearly any photographic composition has a single, optimal, focal-point/DOF combination - so then what's the point in recording all others?

I tend to agree to most of what you've said.
But if the consumer has not the skills and knowledge to find this optimal point at that desired moment, he/she might be glad to find that in pp.
Depending on the IQ output I see this as a "let's make it easier for the consumer"-thing and no so much to the pro/enthusiast. But maybe the IQ will get that high that it'll become interesting to them, too.

 

[meywd]

... It seems new and exciting but doesn't really add much substance ... I also believe that nearly any photographic composition has a single, optimal, focal-point/DOF combination - so then what's the point in recording all others?

I tend to agree to most of what you've said.
But if the consumer has not the skills and knowledge to find this optimal point at that desired moment, he/she might be glad to find that in pp.
Depending on the IQ output I see this as a "let's make it easier for the consumer"-thing and no so much to the pro/enthusiast. But maybe the IQ will get that high that it'll become interesting to them, too.

Yes its not a DSLR killer, but its still in the beginning, it will solve AF issues, help with manual focus lenses...

 

[unfocused]

It has the potential to be the most revolutionary and disruptive technological development on the horizon.

Camera companies invest millions in autofocus systems, but that investment would mean nothing if the photographer could refine focus after the shot has been taken.

No need for AFMA, no real need for sophisticated autofocus. Aim the camera at the subject, get the picture generally in focus and shoot. Then, later (or in-camera) pick the exact point of focus you want.

No more portraits where the nose, rather than the eye is in focus, no more bird-in-flight pictures where a wing or tail feather is in focus but the eye isn't, no more sport's photos with the focus just slightly off, no more wedding photos with a slightly missed focus when the bride is throwing the bouquet.

Autofocus systems are one of the main differentiators between various models in the manufacturer's line-up. But, if this is perfected, you won't need 60 cross-type autofocus points. A single point that gets you in the ballpark is all that is necessary. From there, you adjust it in post.

Not to mention the impact on lenses – imagine being able to infinitely vary the depth of field of any lens. Take a picture with your f4 lens and then adjust in post to have the same Bokeh as a f1.2 lens or the depth of field as though you'd shot it at f64.

How soon and how well it gets implemented is the major question, but no other development has a similar potential to revolutionize photography.

 

[Oceo]

I rented an Illum for a week to see if it would improve my "keep/shoot" ratio for my specialty - butterflies in flight (BFIF). It didn't. In truth, operator technique likely was partly responsible for that un-improvement. The post processing for re-focus is limited to the equivalent DOF of an f/16 for whatever equivalent focal length was being used at time image capture. Biggest problem was the time needed for post-processing of the images. About 40 seconds per image on a shoot of 300-500 images is a looooong time to wait for the re-focusable images to show up. I used a PC with Win 7 64-bit OS, Intel i5 processor. The camera begs for a viewfinder, at least for BFIF.

The camera's fine print notes that the PC is an extension of the camera, so the Illum isn't a $1600 camera, it's really a $3,000 camera.

 

[iron-t]

It has the potential to be the most revolutionary and disruptive technological development on the horizon.

Camera companies invest millions in autofocus systems, but that investment would mean nothing if the photographer could refine focus after the shot has been taken.

No need for AFMA, no real need for sophisticated autofocus. Aim the camera at the subject, get the picture generally in focus and shoot. Then, later (or in-camera) pick the exact point of focus you want.

No more portraits where the nose, rather than the eye is in focus, no more bird-in-flight pictures where a wing or tail feather is in focus but the eye isn't, no more sport's photos with the focus just slightly off, no more wedding photos with a slightly missed focus when the bride is throwing the bouquet.

Autofocus systems are one of the main differentiators between various models in the manufacturer's line-up. But, if this is perfected, you won't need 60 cross-type autofocus points. A single point that gets you in the ballpark is all that is necessary. From there, you adjust it in post.

Not to mention the impact on lenses – imagine being able to infinitely vary the depth of field of any lens. Take a picture with your f4 lens and then adjust in post to have the same Bokeh as a f1.2 lens or the depth of field as though you'd shot it at f64.

How soon and how well it gets implemented is the major question, but no other development has a similar potential to revolutionize photography.

I agree. This technology is in its infancy now but has the potential to render DSLRs completely obsolete. It also has the potential to further push the art of photography into a process of editing rather than capture, as capturing all information needed for any given scene becomes easy. I see advantages and disadvantages. Advantages: raw materials for making great images is easy to come by; cameras are small enough to take to remote locations even when traveling as lightly as possible (e.g. hiking or running); less need for a range of top quality optics. Disadvantages: capture process is no longer involving, creative, or skilled; more time spent stuck in front of a screen post processing.

 

[AcutancePhotography]

I can also see utility for security cameras too.

 

[MickDK]

"Light-field photography is the future" - or is it?
...
What do you think: Is old-school single-dof photography doomed?

It's a very interesting and promissing technology.

But I am sure that every advantage comes with some tradeoff. As this tech is not widely spread yet these tradeoffs are not described or even discovered.
If we know them we can decide what is the better solution.

It's just a gimmick. The tradeoff is (*much*) less resolution that would otherwise be possible for a sensor of the same size.

 

[mnclayshooter]

If you've not seen it and are interested, the founder of the company published his doctoral thesis on the technology... It's an interesting read if you want to digest some of the technical explanation of the technology.

There's pictures that explain it pretty well for the rest of us.

https://www.lytro.com/downloads/resources/renng-thesis.pdf

 

[KrisK]

Panasonic and others have filed patents for their own light-field technologies, some of which circumvent the imaging trade-offs mentioned above. (In on system, existing lenses can be used, if I recall correctly.)

Here's one article from Imaging Resource: http://www.imaging-resource.com/news/2014/04/16/panasonic-patents-light-field-sensor-with-high-res-capture

Frankly, I'm skeptical about the whole concept. People want to passively 'consume' photographs, they don't want to actively 'interact' with them.

 

[Tabor Warren Photography]

What I would gladly accept would be a system designed to modify the focal point within, say, 3 inches.

I can only imagine how many shots would be saved if the focus of the image could be brought just a touch closer/further away. It would be incredible for bridals, families, even our little canine buddies. I also think it would help with the time needed to capture, process, and move on to the next image in camera.

-Tabor

 

[mnclayshooter]

What I would gladly accept would be a system designed to modify the focal point within, say, 3 inches.

I can only imagine how many shots would be saved if the focus of the image could be brought just a touch closer/further away. It would be incredible for bridals, families, even our little canine buddies. I also think it would help with the time needed to capture, process, and move on to the next image in camera.

-Tabor

This is precisely what intrigues me about it... sort of an on-the-fly AFMA.

How many times have you pressed the button, saw something on the screen and said, "that looks perfect," only to get home and discover that it's slightly out of focus? This technology could, once far enough developed, help us with that. Think of the tweaking you could do on macro shot? or virtually eliminate the need to do focus stacking manually... To me, it's not exciting for landscapes etc... but some of the other more-DOF-specific types of photos, it does raise an eyebrow.

I've been following along with Ng's proposed technology for a while... basically since I heard of the "lightfield" or whatever word you want to describe basically simultaneous/array multiple-image capture. To me it has some merit... imagine what someone said when the idea of a flip phone was proposed after a show like Star Trek had shown the concept decades earlier?... technology starts somewhere... sometimes it doesn't end up being used the way it was originally invented... it takes time and innovative thinking to figure that part out though.

 

[Marsu42]

It's just a gimmick. The tradeoff is (*much*) less resolution that would otherwise be possible for a sensor of the same size.

At the moment, that is. A decade from now, people might evaluate current dslr tech as ancient as we do with film cameras. And back in the film days, lots of people would have said "digital is a gimmick, look at the resolution"...

Frankly, I'm skeptical about the whole concept. People want to passively 'consume' photographs, they don't want to actively 'interact' with them.

I understand the concept as a benefit to the photog, not for playing around with the focal point for the viewer. If you can adjust the dof a bit after the fact and get some 3d representation, this would be terrific.

 

[Mt Spokane Photography]

If you've not seen it and are interested, the founder of the company published his doctoral thesis on the technology... It's an interesting read if you want to digest some of the technical explanation of the technology.

There's pictures that explain it pretty well for the rest of us.

https://www.lytro.com/downloads/resources/renng-thesis.pdf

He tried to sell the idea to all the camera companies, but none of them were interested, not even Sony

Then, he sold the idea to some financial backers and kicked out his first model, which flopped at several levels. Then, he went back to the well and found more financing and turned out the current model which has had better acceptance, but has not really managed to generate a lot of interest.

Its a big question as to any further financing, if this product does not sell, its likely dead. Innovation is a good thing, but so far, the wild claims are not backed up by facts. The variable focus range is very limited.

 

[KrisK]

Frankly, I'm skeptical about the whole concept. People want to passively 'consume' photographs, they don't want to actively 'interact' with them.

I understand the concept as a benefit to the photog, not for playing around with the focal point for the viewer. If you can adjust the dof a bit after the fact and get some 3d representation, this would be terrific.

Agreed; the ability to tweak or correct focus after-the-fact would be nice.

But Lytro does seem to be pushing the interactive element pretty hard in their "Lytro Living Pictures" gallery page. https://pictures.lytro.com/

As a viewer, I'm not sure what I'm supposed to make of these things as I play around with them (e.g. "Ball, Fuzzy Boy; Boy, Fuzzy Ball".) Am I supposed to make up a narrative for these pics? Is there some dramatic tension between the boy and the ball that I don't know about? I'm guessing not, so in each case I find myself wondering what's the point?

Ironically enough, the novelty of the interaction trivializes the actual photographs, some of which are very nice.

 

[Marsu42]

As a viewer, I'm not sure what I'm supposed to make of these things as I play around with them (e.g. "Ball, Fuzzy Boy; Boy, Fuzzy Ball".)

I agree changing the dof seems to be rather childish (like I laughed my a** off as a 5-year old when seeing a movie backwards). But I am impressed by the interactive 3d effect you can create when "focus-stacking" the whole dof region of the camera. Imho ( ) It's not only nice to play around with, but you can even change the composition in post.

https://pictures.lytro.com/lytro/collections/41/pictures/894396

 

[meywd]

As a viewer, I'm not sure what I'm supposed to make of these things as I play around with them (e.g. "Ball, Fuzzy Boy; Boy, Fuzzy Ball".)

I agree changing the dof seems to be rather childish (like I laughed my a** off as a 5-year old when seeing a movie backwards). But I am impressed by the interactive 3d effect you can create when "focus-stacking" the whole dof region of the camera. Imho ( ) It's not only nice to play around with, but you can even change the composition in post.

https://pictures.lytro.com/lytro/collections/41/pictures/894396

I can see a use case in interactive Ads