Ansco 130 works for me!
Ansco 130 works for me!
"I would feel more optimistic about a bright future for man if he spent less time proving that he can outwit Nature and more time tasting her sweetness and respecting her seniority"---EB White
The issue is, at what point are just oversampling the information that the lens can resolve?
Current sensor on dslrs are able to resolve around 80 lp/mm
The very best MF digital lenses can resolve 80 lp/mm at around 50% modulation on axis, and down close to 0% modulation in the corners. Large format optics can't come anywhere close to this.
So how much more pixel density can we really benefit from?
Ok. I just looked at the 32mm one. That's certainly better. I wonder at what point they consider the lens to be at its practical limits. In film photography, 30% modulation is generally thought to be the lowest contrast that's worth anything. Possibly with a digital sensor and the ability to sharpen, you can get useful detail with less than that.
Wide lenses are weakness of MF digital. Lenses built for MF DSLRs have to be retrofocus, are huge, have many elements leading to flare, have barrel distortion and are usually not as good performers as non-retrofocus designs (except in falloff).
Even lenses built for digital view cameras need to be retrofocus designs, although less extreme. Too high incident angle on digital sensors leads to color problems.
A Rodenstock HR-Digaron-S 28mm lens has a flange focal distance of 53mm, considerably more than its focal length. It also shows the typical retrofocus problems, distortion and performance well below the sharpest lenses for the format.
A Rodenstock Grandagon-N 90mm has a flange focal distance of 94-98mm (F6.8, F4.5), nearly the same as the focal length. Less retrofocus, less distortion, better performance relative other lenses for the format.
First - Thanks Emmanuel for the good explanation!
Since others have asked about pixel size and QE note from OVT website:
http://www.ovt.com/applications/app_mobile.php
They have pixel size of 1.4 microns. Some sensors are back side illuminated (BSI) which means ~100% of photons impinging on the silicon surface are absorbed so QE is high.
Having done both film & CCD astrophotography, there is no comparison, CCD wins out due to ability to stack images and remove noise via dark frames. The best film was hypered TechPan. For large high performance sensors one can look to Fairchild with their BSI models although the prices are truly astronomical.
Best Regards,
Tim
Sorry, I have been unable to reply before now.
Emmanuel says at the end of his excellent reply that to get the sensitivity (which relates to both ISO and bits per pixel does it not?) we need bigger pixels. And that I believe is where I started from.
As for getting more photons out than went in ...
I hereby declare my invention ... Erbium doped CCDs. Each pixel (post RGB filter could be Erbium doped ...)
Ha!
Thanks guys. I'll stick with it (the LF thing) for now.
I'm no digital whiz, so this may be a stupid question; nevertheless I'll stick my neck out and ask it. Given that a digital sensor is totally a different matter from a piece of film, and is a static, integral part of the camera: why do they not work on hemispheric sensors? This would have at least two major advantages, removing the angle-of-incidence problem and allowing the lens designer to forget about spherical aberration! Maybe at first this would work only for fixed focal lengths with a sensor specially designed for the focal length in question. But perhaps further developmental work would result in greater flexibility.
Of course, the ultimate gain of resolving power would probably involve eliminating glass optics entirely in favour of some kind of laser holographic system!
Oversampling is not a bad thing, or even pointless. Nyquist theorem says you want 2x over-sampling, so for an 80lp/mm lens you would want a 160lp/mm sensor with no AA filter to maximize resolution without artifacts. Actually with Bayer-filtered sensor I guess there would even be some potential benefit from going higher than 2x, because each sensel is R, G, or B and therefore not a "full" sample.The issue is, at what point are just oversampling the information that the lens can resolve?
Jeff Kohn
The Majestic Landscape
I definitely don't think 16mp full-frame is anywhere near optimal. A year or two ago I might have said that 24mp is the sweet spot for full-frame; but the 16mp APS sensors coming out since then have shown impressive resolution and dynamic range, which leads me to believe there's still some headroom for full-frame sensors. I expect Canon will have a 30-32mp 1Ds3 successor before too long, with Nikon following at some later point.Is it not true that going too much smaller than the size of pixel on a say 16M full-frame DSLR leads to worsening ISO performance and increasing noise ?
It's true that as you go higher, pixel-level noise at high ISO's increases. But for a given print, pixel-level SNR is less an issue than sensor size. I suppose if all you care about is ISO 6400 performance, there's no much point in high-density sensors. But a D3x ISO-100 shot will easily beat the D3s in both resolution and dynamic range, so I don't buy the argument that 12mp or 16mp is all you need.
Jeff Kohn
The Majestic Landscape
Bookmarks