Prev: Which digital picture frame can do RANDOM slideshow?
Next: Life Expectancy of Cheaper Digital Cameras
From: Wolfgang Weisselberg on 7 Feb 2010 13:24 Alfred Molon <alfred_molon(a)yahoo.com> wrote: > We've discussed this matter to death already. Do a Google groups search > for all the details. Yep, we did, and the 4.7 M pixel positions Foveon rivals a 6 or 7 MPix classical bayer filter (RGGB or BGGR) sensor for resolution. As any Google search would show. As any Google search would show, there are plenty cameras with 8 or more MPix bayer sensors. As any Google search would show, Foveon has bad colour separation and no higher ISO qualities to speak of. Shall I go on reading you what you should have googled yourself? -Wolfgang
From: Wolfgang Weisselberg on 7 Feb 2010 13:43 nospam <nospam(a)nospam.invalid> wrote: > In article <hjv99s$no4$1(a)rjf7r.motzarella.org>, Bob Fleischer > <bobfnospam(a)duxsysnospam.com> wrote: >> If having three imagers is desirable for a better video camera, are >> there any three-imager still cameras? If not, why? > keeping three chips perfectly aligned is *very* difficult .... though maybe someone will come up with a method to use a sensor-moving IS technology to allow self-adjusting the chips. > and it would > require a beam splitter which makes most lenses not work, Put the beam splitter where your classical DSLR has it's beam splitter (half mirror) and forego the classical viewfinder (and fast phase detection AF). > there is > three times as much data to move which means storage and cpu > requirements triple, JPEG doesn't care if 2/3rds of the values for each pixel are interpolated or from a real sensor. Only RAW would triple in size. And you'd not need the interpolation routines (which have to be quite clever --- and think how much more time they need if you do RAW processing on your PC) so while I/O speeds would need to triple (by using 3 paths for 3 sensors, instead of one from one) CPU needs might actually drop. And since you don't need interpolation, you'd make do with less pixels for the same resolution, so I/O doesn't need to triple, it needs only slightly more than double. You'd be able to use a much weaker AA filter, too. And with proper beam splitting you don't need to loose more than small amounts of light to the splitting-by-colour, which could increase sensitivity especially for the red and blue channel ... > it's a lot more expensive and it doesn't offer > much benefit. That's probably the problem. -Wolfgang
From: Ray Fischer on 7 Feb 2010 15:49
Wolfgang Weisselberg <ozcvgtt02(a)sneakemail.com> wrote: >nospam <nospam(a)nospam.invalid> wrote: >> In article <hjv99s$no4$1(a)rjf7r.motzarella.org>, Bob Fleischer >> <bobfnospam(a)duxsysnospam.com> wrote: > >>> If having three imagers is desirable for a better video camera, are >>> there any three-imager still cameras? If not, why? > >> keeping three chips perfectly aligned is *very* difficult > >... though maybe someone will come up with a method to use a >sensor-moving IS technology to allow self-adjusting the >chips. The point isn't whether it _can_ be done. It's whether it's worth the trouble and expense. -- Ray Fischer rfischer(a)sonic.net |