silverhalide
Offline
•
Upload & Sell: Off
|
 p.37 #7 · 'Un-Official' pre-PMA Rumor Thread | |
Monito wrote:
they would be incompatible with almost every single photographic lens ever manufactured.
I never questioned this; actually, I sort of assumed it. Hence, the need for a new mount, etc.
Tentacle wrote:
Curving a sensor plane? Oh boy, that's not a can o' worms, that's a 40 foot sea container full of worms:
hehe. Now there's a vivid image.
You need dish shaped wafers. Try cutting that out of silicon billet. Then try to use curved shapes in a stepper scanner. Oh wait... your stepper scanner litho equipment needs to project its mask on a curved surface. Not to mention that the whole sensor production line needs to be adapted for non-flat specimens.
This is the kind of response I was hoping for. I knew there had to be reasons this hadn't been done.
Then there are the lenses. They need to be designed and build from the ground up for projecting on a curved plane. That's relatively easy for a prime, but not so for a zoom.
Why is this so? If we can design a lens to focus on a flat plane with a variable focal length, and we can design a lens to focus on a curved plane (hmm, that term is an oxymoron I think), why can't we design one to focus on a curved plane with variable focal lengths.
I had actually been wondering if the requirement to focus on a flat plane was hampering lens/camera design, and if it was an artifact from design decisions made a hundred years ago.
And how are you going to place the shutter? You can't put the shutter right in front of the sensor, because the opening will not be equidistant to the sensor anymore.
So? is there a requirement that the shutter be equidistant from the sensor? or can it be placed anywhere between the sensor and the mirror?
What is the effect of a non-equidistant shutter? All I can guess at is that the shadow cast on the sensor by the shutter curtain would be more or less sharp at different places, but I don't know how this would affect things.
Can you make a horizontal curtain shutter dish shaped? If not, then you need a shutter at the optical node. OOPS, that's in the lens, not in the body. And how about sensor alignment? Now you need to have the sensor parallel and at a set distance. A little shift upwards or downwards isn't a problem. With a concave sensor, all of a sudden you need to get it aligned on 2 more axis.
Ok, that sounds like a good argument against.
Need me to speculate any more on the matter? 
No, but feel free to . Personally, I think it's fascinating, and at least as relevant (or, productive, take your pick) as half the other ramblings I've been enjoying in this thread.
Returning to why I brought up this subject in the first place: light fall-off in the corners. As a variation on the standard design, I assume that the "gain" could be automatically adjusted on the sensor so that the corner sensor elements were more sensitive to made up for the angled light. However, I'm assuming this isn't done because the amount of gain would depend on the angle of the light, which would depend on the position of the rear lens element, which would change from lens to lens, and with focus distance (and, on a zoom lens, with focal length). Is this correct?
E.
|
?