"Light Field" camera

[Collas]

This sounds very intriguing, though whether you can have everything sharp remains to be seen.

http://www.gizmag.com/lytro-consumer-li ... cus/18996/

Nick

[zoikes]

From a brief reading it looks like this technology may be similar to tomosynthesis that is being introduced in X-ray imaging. Utter conjecture, but it makes sense to me.

The X-ray systems take a series of images at slightly different angles. From this you can reconstruct a 'slab' that is in focus and then scroll through a series of such images through the structure of interest. Breast imaging is where much of the research is - because clinicians are looking for such small, low contrast objects (i.e. cancer).
In photographic terms - the depth of field is pre set, but you can change the focus point. However, there is already software to combine a stack of such (photographic) images if you focus bracket (e.g. Helicon focus) so you can get everything in focus in the final image.

The blurb says that the Light Field camera records the angle of the light ray. I wonder if that means the sensor pans across the projected image during the exposure.


Ian
(Day job : Medical Physicist)

[Emmanuel Bigler]

The moderators could argue that we are going off-topic here, but what is claimed regarding the future "revolutionnary camera" is close to some of our requirements in terms of depth of field, and may be could make obsolete all our techniques and discussions about "scheimpflug and extended DoF".

The principle of the Lytro(TM) camera was announced and discussed on the web when the inventor, a Stanford Ph.D. student, presented his oral dissertation. I agree with Ian that the images are probably reconstructed by computer like in X-ray tomography or magnetic resonance (NMR) imaging.

So far the only comment I've read is that "the quality is the similar to an ordinary point and shoot digital camera". If this is so, it is a bit short for us, readers of this LF forum. However, any imaging technique that relies on brute-force computer methods cannot be ignored, since any consumer point & shoot digital camera actually relies on more-or-less secret brute-force image pre-processing techniques.

The point I'd like to discuss is that the new camera pushes forward our dependency, I should say : "slavery" with respect to software and computers, to an advanced degree that I find totally un-interesting as an amateur photographer.
In medical imaging, the situation is exactly the reverse; without the use of the computer, medical images would be poor, conventional X-rays of one century ago were harmful for the doctor and for the patient. "No computer" in medical imaging would mean : no X-ray tomography, no NMR, no ultrasound imaging ...

Now if we are interested in large format cameras, it is because we need something else than a point and shoot camera. Not that we never use point and shoot cameras, on the contrary: we love them, we use them often, we collect them, but we do not mention them here
But we need something else. Something manually operated. Something to directly challenge our hands, our eyes and human skills. Directly. without any computer. Like a hand-crafted piece of furtinure that we dream to fabricate by hand from the tree trunk that we cut in the woods to the final coat of varnish.

If we are interested in large format cameras, it is because the arguments in favor of the "decisive moment" (quoting Cartier Bresson), "press the trigger, we do the rest" "now with his camera you can't miss any shot" "fully automatic" , etc, etc.. are advertising tricks that we've been reading since the Rolleiflex "Automat" was introduced in 1937 (not me, but my father probably heard about this revolutionary camera that won a Grand Prix at the Paris Fair in 1937).
Each new generation of cameras tries to promote faster action, "our technology let you unleash you creative freedom" and other kinds of well-known arguments, used ad nauseam.

Here with the new camera, you won't even have access to the image files. Even with some professional digital cameras, you have at least access to a JPEG file, for which the format is public and well documented, or it is a RAW file for which the technical specifications of the proprietary format are accessible, at least, to 3-rd party software developers.
In the proposed 'revolutionary camera', the reconstruction of images are trade secrets, it is what you pay for in order to get the advantage of the multi-focus image capability.

Moreover, the idea that "you do not care for focusing, you'll do it later on the computer" is something which has proven to be counter-productive in professional digital photography. On the contrary, the more you prepare your lighting, your framing, your focus, the less you have to spend time in costly hours of post-processing. What is the reason why Canon and Nikon have updated their tilt+shift lenses ? Why does Schneider introduce a new line of tilt+shift lenses ? Simply because setting a shift just prior taking the picture takes 5 seconds, correcting the converging verticals by post-processing takes at least 1/2 hour.

If I use a view camera on film it also is because I do not want to bother with the management of a digital archive in the next decades. because I want to be free of the choice of my lens, of my film, of my chemicals. Of course, I may use a computer, if I wish, for post-precessing but I don't want to be the slave of the computer for my hobbies. And, as a general rule, I do not want to be tightened by proprietary software and secret file formats.
And if I use the view camera, it is because properly focusing by hand on the main subject I have chosen is a pleasure, not a punishment. Because selective focusing is part of my aesthetic choices. I want do it in front of the subject. I want to be challenged to be able to take a perfect shot in front of the subject. I know that I can correct some mistakes aftewards. To me this is not real fun.

If some day I want to disguise as Cartier-Bresson capturing the decisive moment, I have several small-size cameras that I can play with. Unfortunately those "old" cameras refuse to die and serve me faithfully ... no need for a new one

Now regarding the commercial aspect of the Lytro(TM) camera, I'm curious to see how this camera could find its place on the market. A strategic decision was apparently made not to licence the product & related patents to the Big Players, the Gang of Four, but to develop an independant product with direct-marketing technques. A real challenge ! We'll see, and re-focus on this new camera .. later on

[Joanna Carter]

An excellent discourse from Emmanuel

I have a somewhat simpler question: how do you print the image at 40" x 32" with everything sharp?

To me, it seems like you will not be able to see the entire image, sharp, without viewing it on a computer, or even when viewing it on a computer.

By the way, as a moderator, I am very tempted to delete this thread; as Emmanuel says, this is far from LF

[Neil Barnes]

While I have no plans to use such a device, with my engineer's and signal processing hat on, I'm fascinated by the sensor.

There's a good description here: http://graphics.stanford.edu/papers/lfcamera/ but the significant issue to us is that it requires dozens of the original pixel sensors under each microlens; you have a high resolution sensor, but it can only produce a low-resolution image. The tech report from Stanford is well worth a read - I just skimmed it, but there's some good theory there.

Perhaps this will lead to affordable digital backs for 4x5? They're going to need that sort of resolution to be able to match a modern SLR sensor.

Neil

[Dave Tolcher]

Look here also:

http://www.bythom.com

His postings this week have been a discussion if this subject so hopefully havent moved to archive.