As a physicist I carefully read the PhD dissertation of Ng, inventor of Lytro. There is nothing inherently bad in his invention.

Unfortunately, available technology doesn't always keep pace with great inventions.

Regarding dishonesting comments addressed to Lytro, I fully agree with statement of Rabbi Shlomo Riskin:

"When you're one step ahead of the crowd you're a genius. When you're two steps ahead, you're a crackpot."

@Wisniewski:

Alhough my first comment was response to original post, it seems that you are offended my opinion which is so strongly different from yours.

So be assured, that I have no intention argument with you, and I only repeat the simple fact that in Ng invention physical laws are not broken, but available technology is not mature enough to take advantage of this invention.

So was in past with many other inventions, which are today common things...

Joseph, you've made some MASSIVE assumptions, and they are informing your entire argument.

What if the plenoptic feature ... could be turned off? Maybe it's possible; maybe it isn't. Also, there is a large professional market for 3D video cameras (for which there was also no 'need', until the technology was invented, btw) ... what if the second or third generations of these sensors could create single-lens 3D video?

I could come up with a long list of famous inventions for which there was no need, but sold like crazy. Flat-panel TVs and every single non-portable advancement in audio since 1970 or so, for instance. Oh, and almost every personal hygiene product ever.

Maybe this product is just a flash in the pan. But you'r assuming not only that this technology will never improve, but also that there will never be any more demand than what currently exists.

I'm sorry, but those are both very faulty assumptions.

jeangenie, poor little greenie, I've made no "massive assumptions". I've drawn from Ng's own published papers and my own knowledge of computational optics and volumetric algorithms.

"what if the second or third generations of these sensors could create single-lens 3D video?"

Now, that's a massive assumption. There is something in single perspective point 3D image processing called the occlusion problem, which means that you do not have the information needed to fill in what is occluded by foreground objects in one eye's view, but is not occluded in the other eye's view. Single point 3D looks contrived. You'll notice that most of the 3D examples you see presented by plenoptic researchers are at macro distances. At such distances, you can select rays to create a small stereobase multiple point view, but that base is limited to 1/2 the aperture diameter.

In short, the Lytro can make great 3D images, of stuff about 18 inches away.

"I could come up with a long list of famous inventions for which there was no need, but sold like crazy. Flat-panel TVs..."

Flat panel TVs are an excellent example of an invention for which there is a massive need (space saving, weight saving, power saving, decreased cost, increased size, higher resolution, reduced radiation) so that pretty much defeats your argument. Are you aware of the decades of research that went on at places like Sarnoff or SRI to bring the microtip emitter plasma display to fruition? The billions pumped into LCDs by Sharp, NEC, or Samsung?

"But you'r assuming not only that this technology will never improve,"

No. I'm stating that it will not improve to the point where, as guif asserts "tomorrow all cameras will be plenoptic". Rather than stuff words in my mouth with your version of what I'm "assuming", why don't you read what I wrote?

You can't turn off microlenses (now, that's a massive assumption) so this will never be a general purpose camera.

Joesph, you can Google information all you want. I'm assuming you're not an expert at everything, because you're obviously the sort of person who would have told Zdenek so.

If you were in the television business ten years ago, you wouldn't argue about whether or not there was a need for a flat-panel TV. For the military, or course. For consumers - hell no. But it was sexy, and a status symbol; at the time, an $8,000 LCD had a worse picture than a $1000 CRT, AND had no speakers. In fact, even today certain Sony CRT TVs have a reputation for having a better picture than most current LCD/LEDs. Another ten years before that, it was generally thought that LCD displays could never reach the level of detail that is currently available today, let alone on a $250 TV set.

Microlenses can never be turned off? Bah. A microlens is designed to scatter night in a certain pattern. Many high-end homes feature special windows which act normally, but when a switch is flipped and electricity applied, they turn opaque. This is done by a thin LCD panel between the two panes of glass that rearranges itself with power is applied. Light is scattered, and the glass looks opaque.

This technology could be integrated into a camera. Cameras could 'create' microlenses by drawing extra power. Single-lens 3D could be created by automatically taking two exposures; the first recording data for one eye, and then the lenses reorganize to capture data for the 'second' eye. Assuming shutter speeds are kept above 1/125 of a second, this technology would allow single-lens 3D capture at 1080p60.

You continue to make the biggest, and worst assumption of all: that what is is what will be. Your problem is that no matter what you know, how much you read, how much you consider yourself an expert on ... you lack the ability to dream big. You lack creativity, Joseph-San.

I'm not saying that this will happen now, or ever. I'll I'm saying is that the technology technically exists, and that if you think laterally, it can one day be applied to camera manufacture.

I asked you 'what if?' this things can happen. I did not tell you they would. The Lytro team may fail, but it began with a thought of 'what if?' Many failures begin that way, but so do all great inventions.

"You continue to make the biggest, and worst assumption of all: that what is is what will be."

You continue to stuff words in my mouth, and it is most unpalatable. I commented on what is possible, given the physics of this particular universe.

"I'll I'm saying is that the technology technically exists"

You are incorrect. Your analogy with cloud-gel windows is false, because lenses do not scatter light, they redirect it. Your knowledge of history is incorrect: flat screens were quite common ten years ago in civilian use (they were more common back then, than your own smart windows example is today).

"you lack the ability to dream big. You lack creativity, Joseph-San."

Oh, jean-chan, you really have no idea who you're talking to. I have dreamed bigger than you ever will, and will continue to do so. You're without dreams of your own. You read a bit of highly skewed marketing literature, full of contrived demos, and think that breathing that second hand smoke is dreaming.

"Single-lens 3D could be created by automatically taking two exposures; the first recording data for one eye, and then the lenses reorganize to capture data for the 'second' eye."

You're failing to grasp a fundamental concept. You need to take those two exposures from locations fairly far from each other. For convincing 3D of things 6-10 feet away, that distance is approximately human eye spacing, around 60mm. The standard rule in 3D work is 30:1, you want to shoot 3D of action 20 feet away, you set the lenses up with a 200mm separation.

Lytro can already do as much ray selection as is possible behind its tiny lens. Problem is, with a 22mm lens, you've got the ability to create your "pair of eyes" 15mm apart, which is a fraction of what is needed for general photography. That's why the examples are shot so close.

There is good reason why humans do our 3D with two small eyes, instead of one giant eye the size of our faces.