[First off, you’re absolutely right that Winer is obnoxious and in this case pretty clueless. That said...]
> Optics is, more or less, a solved problem. We hit hard physical limits long ago.
This is only true for cameras which are designed around a principle of light passing through a “generalized pinhole” (i.e. go through a lens and are collected as a two-dimensional image in a single frame at the back, without doing any fancy math; the kind of set-up required by, for instance, a process based around film or our current types of digital sensors).
Look to biology though for inspiration and it quickly becomes clear that visual information can be captured in all kinds of ways, with all kind of crazy “eye” designs.
There’s a huge amount of information stored in the light rays passing through a camera which is currently not captured, and the cameras of a few decades from now will blast through all of those “hard physical limits”. We’re going to get much better approximations of actual light spectra coming in, along with information about polarization and distance to the subject; simulate any lens, any aperture, any focal length in software because all of the relevant data will be captured. (In my dreams, anyway, or even maybe my somewhat-informed guesses; we’ll see how well technology actually advances.)
Anyway though, our typical current digital cameras don’t even hit the limits given the traditional/obvious camera design approach: for instance, we currently have three different colors of pixels each of which filters out a different set of wavelengths, meaning we are only capturing a “third” of the information we need to make a metameric match from each point, and just tossing out a bunch of the other photons. That right there is worth a something like 2x improvement in data captured per area.
One last thing: even before we get crazier camera designs, improvements in computer processing speed are going to make it possible to create high-resolution images out of an image captured over time (i.e. a movie processed into a still): we can see the beginnings of this already with those panoramic cameras I’ve been seeing TV commercials for.
Optics is a solved problem. Optics. I said optics, not "cameras", or "photography".
Plenoptic cameras are cool, and have a bunch of neat features, but each camera element still uses traditional lenses, made of specially shaped transparent materials, whose fundamental design has not changed in four hundred years. There's been a lot of neat hacks on the basic idea (achromatic lenses, etc) but nothing new. Except superlenses, but you'll never see a macroscopic superlens, so it's useless for photography.
My main point is that it’s quite possible that we’ll have cameras in 50 years which are the size of today’s compacts, at the quality (or better) than today’s full-frame DSLRs. Indeed, I’d be surprised if we didn’t.
Comments
[First off, you’re absolutely right that Winer is obnoxious and in this case pretty clueless. That said...]
> Optics is, more or less, a solved problem. We hit hard physical limits long ago.
This is only true for cameras which are designed around a principle of light passing through a “generalized pinhole” (i.e. go through a lens and are collected as a two-dimensional image in a single frame at the back, without doing any fancy math; the kind of set-up required by, for instance, a process based around film or our current types of digital sensors).
Look to biology though for inspiration and it quickly becomes clear that visual information can be captured in all kinds of ways, with all kind of crazy “eye” designs.
There’s a huge amount of information stored in the light rays passing through a camera which is currently not captured, and the cameras of a few decades from now will blast through all of those “hard physical limits”. We’re going to get much better approximations of actual light spectra coming in, along with information about polarization and distance to the subject; simulate any lens, any aperture, any focal length in software because all of the relevant data will be captured. (In my dreams, anyway, or even maybe my somewhat-informed guesses; we’ll see how well technology actually advances.)
Anyway though, our typical current digital cameras don’t even hit the limits given the traditional/obvious camera design approach: for instance, we currently have three different colors of pixels each of which filters out a different set of wavelengths, meaning we are only capturing a “third” of the information we need to make a metameric match from each point, and just tossing out a bunch of the other photons. That right there is worth a something like 2x improvement in data captured per area.
One last thing: even before we get crazier camera designs, improvements in computer processing speed are going to make it possible to create high-resolution images out of an image captured over time (i.e. a movie processed into a still): we can see the beginnings of this already with those panoramic cameras I’ve been seeing TV commercials for.
Optics is a solved problem. Optics. I said optics, not "cameras", or "photography".
Plenoptic cameras are cool, and have a bunch of neat features, but each camera element still uses traditional lenses, made of specially shaped transparent materials, whose fundamental design has not changed in four hundred years. There's been a lot of neat hacks on the basic idea (achromatic lenses, etc) but nothing new. Except superlenses, but you'll never see a macroscopic superlens, so it's useless for photography.
Here are some lecture slides from a neat course at MIT:
http://stellar.mit.edu/S/course/6/sp09/6.815/courseMaterial/...
http://stellar.mit.edu/S/course/6/sp09/6.815/courseMaterial/...
My main point is that it’s quite possible that we’ll have cameras in 50 years which are the size of today’s compacts, at the quality (or better) than today’s full-frame DSLRs. Indeed, I’d be surprised if we didn’t.