Jostein wrote: > > Hi, Keith,
I think it was you (Jostein) that said: > > > If light *has* to strike the sensor at an oblique angle, it would > > > logically follow that the chip has a blind spot at the center. So > > > right, Bob. To which I responded: > > I think you knew what he meant, Jostein. You're just funnin' him, right? > > The single photon occupying the lens' axis is dead on, 90 degrees to the > > plane of the sensor, and by definition, all other ray paths out to the > > edge of the sensor must necessarily be at increasingly larger angles... > Are you sure you understood what he wrote? Was it in answer to this paragraph from Mark Erickson, quoted immediately below? I don't like Mark E's words "...adjacent to the sensor..." because I don't think that's what he meant. In HIS definition, the words "the sensor" would imply those pixels immediately on the centerline of the CCD. I don't understand what he had in mind when he said "adjacent to the sensor." The whole darned CCD IS the "sensor!" All the way out to the edges of the working pixel field. * * * >From: "Mark Erickson" <[EMAIL PROTECTED]> > >> It should be fairly clear from the figure at the bottom that light >> striking the CCD perpendicular to the array gets focused on the >> sensor. What is not shown is that light striking at an oblique >> angle will get focused adjacent to the sensor, not on it. * * * > Both you and I know that there's necessarily more than one photon > hitting the sensor at a right angle. It's kinda obvious since any aperture > opening has a physical area larger than a photon. Well, sort of. But a strict interpretation of your words immediately above imply that the light beam is collimated, and we both know it isn't! The only rays that can impinge on the CCD at exactly 90� is that fine stream of photons at the precise, exact center of the lens. All photons passing thru even a .002" radius _outside_ the center are indeed at an angle, and NOT at 90�... > The point was, Bob forfeited that the light _had_ to strike at an oblique > angle. I just wanted to point out that particular breach of logic. > > Could be interesting to calibrate our notion of the term "oblique angle", > btw...:-) No calibration necessary, IMMHO... My online dictionary says: "sloping from the horizontal or perpendicular." The only perfectly straight line of photon travel thru a lens are those photon passing thru exactly on center, which do hit the sensor at exactly 90�. All others outside that MUST be oblique to that center line. > > > IMO, it's pretty obvious that a microlens has an acceptable angle of > > > view. Within which, the light will trigger the right response in the > > > sensor. Beyond which, the light will behave in a way that will cause > > > problems. Such as chromatic aberration, or that the neighbouring > > > responds to light it should not respond to. > > Don't you suppose the lens/sensor manufacturer would be perfectly able > > to figure that out? And provide for it? > And haven't they done so, Keith? Releasing all those digital-optimised wide > angle lenses. Sorry to have insulted your intelligence, mate. Oh, you didn't! Sometimes we all have to talk something out, to come to a common understanding or agreement. Or disagreement, whichever. But, if we can keep the heat out of the arguments, we'll be okay. > Jostein Best, keith
