AA filter question - Joe?

Started Feb 9, 2012 | Discussions thread
DSPographer Senior Member • Posts: 2,464
Re: The two sides of "insane resolution"

Joseph S Wisniewski wrote:

Richard Frederick wrote:

Thanks for the answers. Sometime back (film era, budding digital) I did a back-of envelope calculation using lines per mm published in magazine lens tests and concluded that practically, normal photography will not get above about 50 lines per mm (now line pairs per mm). This works out to about 15 megapixels on a "full frame" sensor.

I am talking about commercially available lenses in normal use. I might suggest the following as examples (all Nikon): 35m f2.0 - 50mm f1.4 - 24-70 f2.8 - 105 mm f2.5, etc.

The 105/2.5 will surprise you. I played with that on the bench (and keep in mind that I'm exhausted and hypoxic form bronchitis and my mind is not at peak) I got either 120 or 130lpmm at some quite usable f-stop, probably f4.

120lpmm is 240 pixels/mm monochrome, and 240*sqrt(2) = 339p/mm for a Bayer pattern. That's an even 100mp on FF. And that's one of the reasons it's always been a favorite of mine with 25 speed Tech Pan and even some 6 speed microfilm that Mark Czinski used to spool up.

Now, bear in mind that's a bench test for resolution with an aerial image, and the resolution on film or sensor (due to pixel size limitations) tends to be lower, until you dramatically oversample the image, often by a factor of around 4, which taxes even the microfilm.

And you have no idea what some of the Nikon macros test out to. Several exceed 250lp/mm.

Which is 500 pixels/mm monochrome or 700 px/mm Bayer.
24*700 * 36*700 = 423 Megapixels full frame.

I'm just pointing out this is consistent with your previous pixel limit.

That's the second side of the "insane resolution" approach to camera design.

  • First, you hit the point where you've got enough resolution to make very large prints or enable dramatic cropping. That's "sane resolution", and anything from 12-36mp "fits".

  • Then you've got enough resolution so that diffraction, resolution limits due to spherical aberration and astigmatism, focus errors, camera shake and mirror/shutter vibration limit the resolution to the point you don't need AA filters. That's the first part of "insane" resolution.

  • Finally, you hit the point where the sensor resolution is so high that you essentially experience no limits but the lens, the "aerial resolution" transferred to a sensor plane.

We're approaching the beginnings of insane resolution, but I predict it can go on quite a ways, simply because it does provide some benefit, even if it's a diminishing return, and it gives the camera companies a "one-up" measurable. People love large numbers, even if they never use them all...

None of these lenses approach being diffraction limited to my knowledge (unless, perhaps, stopped way down), so why are these lenses presumed to need and AA filter given the current pixel density cited in the original post?

And thank God for that. If they did approach diffraction limited, they'd need the gigapixel I calculated earlier. As is, I think we can expect the AA filter to go away somewhere around 100-200mp.

I an not trying to start an argument - only trying to understand better.


You're welcome. Hope this helps.

Hypoxic or not, still a good post.

P.S. Get well soon.

 DSPographer's gear list:DSPographer's gear list
Canon PowerShot G7 X Canon EOS 5D Mark II Canon EF 24mm f/2.8 Canon EF 50mm f/1.8 II Canon EF 200mm f/2.8L II USM +4 more
Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow