On Tue, Nov 01, 2011 at 04:47:38PM -0500, Darren Addy wrote: > What follows is a post about Lytro technology from another list I > subscribe to... (informative, so thought I would share).
Yes - that's the research paper out of Stanford, when they were using a 16MP Contax camera. When I saw their technology demo (a little later, IIRC) they were using a DSLR based on a 35mm camera body. > My comment: the relatively low number of pixels (or conversely, the > HIGH number of megapixel sensor you need to get a decent number of > finished image pixels) probably makes this a technology that will need > a lot more time before it is really practical for imaging larger > prints. True, but it's not quite as drastic as it appears in that paper. Later enhancements to the post-processing software allow the recovery of some of the underlying resolution (at the cost of not being able to do so many clever tricks with the image). So while you won't ever get the full resolution you could get with a straightforward 16MP sensor, you can do a lot better than the 1MP or so you might naively expect from the sensor geometry. Back when I first started dabbling in digital imaging I picked 4MP as what I felt I needed for an 8x10 image. In retrospect I was being too conservative - I got some decent 8x10s from a 3MP Canon PowerShot G1. From what I saw a few years ago that quality of image was already possible from the plenoptic technology. I'm sure that the current generation of sensors can do even better (although as the current target market appears to for creating images to view on facebook or a mobile phone, I suspect that they aren't optimising the post-processing for large prints). -- PDML Pentax-Discuss Mail List PDML@pdml.net http://pdml.net/mailman/listinfo/pdml_pdml.net to UNSUBSCRIBE from the PDML, please visit the link directly above and follow the directions.
