A plenoptic camera differs from a regular camera in that it has lots of micro lenses mounted directly on top of an imaging sensor. This has the effect of providing a continuum of possible viewpoints (bounded by the main lens aperture). While a depth map can be extracted from the data captured from a plenoptic camera it isn’t clear to me that early versions of these type of cameras will have enough viewpoints for all types of shots since the width of any 2 viewpoints would be limited by the main lens. Also, sensor resolution has a ways to go before the amount of data will afford super high quality photographs.
However, what makes this type of camera so interesting is that it would be the perfect match to a micro dot lenticular lens (where the lenticular lenses are dots as opposed to long vertical cylindar shaped lenses).
Don’t place an order for one just yet because the available software is pretty primitive and… oh yes, they have only been hand assembled in the laboratory at considerable cost!
The other thing that I find discouraging is that early versions of the camera, should someone decide to manufacture one, will not have the capability to provide source imagery for life size printouts since the technology for micro dot lenticular lens arrays is still too primitive and everything simply lacks the resolution for large size images.
But why wouldn’t 8×10 photos be ok?
Ah, well they would be ok but I think small 3D images are very much akin to high quality doll toys made from digital scans of actors. Wow, they look a lot like the actors but they look like toy dolls. This is a big problem with 3D because once you start creating realistic imagery that is more fully processed by the brain you get imagery that mimics real life possibilities. A person only 10″ tall in a photograph is going to look like a Ken or Barbie toy doll. Or it is going to look like a regular photograph with an optical gimmick that provides depth. Either way, it would be lacking and probably not have enough gravitas to be demanded by the general public.
I know, this is a returning theme in my blog but I think it is so important to the ultimate adoption of 3D photography. After all, what makes any photograph desirable is what the image evokes both overtly and subliminally. Once you open the 3D Pandora’s box you open a higher level of scrutiny that doesn’t exist with regular photographs and the “ken and barbie” effect is not desirable.
Given the expense and the rapid improvement of sensor technology I would guess that plenoptic cameras will become feasible in about 5 more years and the ability to take advantage of the technology will be a lot further along.
So my answer to the title question is that plenoptic cameras WILL be good for 3D in about 5 years or so. In the meantime, there are workarounds that are quite amazing when great care is made in their use.
Want to know more about Plenoptic cameras? Check out Wiki and also search google where you will find numerous papers from MIT and Stamford Universities (among others).