The Circle of Confusion, 'Do Sensors "Outresolve" Lenses?' and Our Digital Demands

BBW

Legend
Location
betwixt and between
Name
BB
At José's request, I've started this thread as a jumping off place for continued discussions regarding a link that Marcus posted over on his thread: https://www.photographerslounge.org...e-digital-camera-built-film-ones-existed-411/ in which he referred to an article, dated June 2008, that has been published in Luminous Landscapes: "Do Sensors "Outresolve" Lenses?, written by Rubén Osuna and Effrain Garcia.

Marcus -

Thanks for linking to an amazing article. I just "read" it, and I think after a couple more tries (maybe this weekend) I might actually be able to follow it.

I think this link is worthy of its own thread. If several of us try to read it and report our understanding of the content, we may be able to (together) do a better job of understanding all of the concepts involved.

(And I thought I already understood what a Circle of Confusion meant, based on DOF discussions. Now I realize I'll have to rethink some of those assumptions too.)

Here is the introduction from the linked article, to whet your appetites:
We read everywhere that new high resolution sensors put pressure on actual lenses. These comments arise copiously each time a new sensor with higher pixel counts appears. It happened with the 22 millions of pixels of the Canon 1Ds Mark III, and it will happen again when the 25 MP sensor by Sony comes to life into a new camera. Are this kind of comments accurate? There is no short answer to the question, because the subject is complex. However, we will try to summarize several basic rules and results, closing a previous discussion at The Luminous Landscape.

Now we have a thread devoted to this article as a reference. Enjoy the article, and the continued cogitating!
 
Most digital cameras use Mosaic filters, "Bayer Filters" being the most popular, for color. Most cameras use anti-aliasing filters which basically "blur" the image slightly to get rid of color aliasing due to these filters.

A camera with a 5micron pixel size "would" resolve 100lines/mm if it were monochrome, but is probably closer to 50lines/mm with the anti-aliasing filter. I have a 1953 Jupiter-3 that can match that resolution. And at F4, does much better.

The M8 does not use an AA filter. Uses 6.8micron detectors. That is 72lp/mm, and aliasing can creep in. The lens is out-resolving the Mosaic Filter's 36lp/mm.
 
Here's a real-world example (Lumix G2, 1/50sec ISO100, Panasonic/Leica 45mm macro open to f/2.8)
5000132838_80843627c1_b.jpg

Cropped and exported to jpg, no other post.

In this case, I suspect the lens outperforms the sensor, in traditional terms. Both the sharp highlight detail in the foreground and the soft bokeh in the background (it's a nice lens :) get lost in the sensor's noise.

Of course, after I try to read the post a few more times, I may find that this reply comment is not related . . . :/
 
Full image scales to 1024p across, Leica M8, ZK Sonnar 5cm F1.5:

picture.php


Tight crop, full-res: 1949 ZK 5cm F1.5 Sonnar at F4:

picture.php


The color artifacts are introduced because the detail is more highly resolved than the Bayer pattern mosain filter. The M8 does not use an AA filter.
 
I am quite happy with the image quality from the M8. I shoot ISO 160 when possible, but have used it at 2500 without problems.

My EP2: I use at ISO 200 whenever possible. I have not noticed a noise issue.
 
Thank you both, Arpoador and Brian, for providing examples of lenses out-performing sensors (even though, on my screen, I am not able to see that much noise in the G2 photo either: it would be nice to see the same image taken by both cameras, with the same ISO setting, by some rich fellow ... :) ; by the way: you both omitted to "declare" the ISO setting you adopted for those photos ....; was it 160 for Brian's ? and what for Arpoador' s ? ). Hard time for those unlucky sensors, though, both having to deal with Leica lenses....
But what would you expect to see when the opposite were true ? I mean, if a sensor out-performed an objective ? Would there be any tell-tale symptoms, or the resulting image would just look "out of focus" ? Based on Osuna&Garcia conclusions, I suspect some camera might already have been driven into the desert by the pixel race, even though - of course - it would be stupid to put a good (= costly) sensor and processor behind a poor lens ... (by the way:extra-pixels may be used to improve the image anyway, as it seems; see Not So Fast - Luminous Landscape ).
 
Hi Marcus -

The ISO is 100 on the G2 pic. Thanks for calling that out. I added the ISO & time to the pic info.

Thank you both, Arpoador and Brian, for providing examples of lenses out-performing sensors (even though, on my screen, I am not able to see that much noise in the G2 photo either: it would be nice to see the same image taken by both cameras, with the same ISO setting, by some rich fellow ... :) ; by the way: you both omitted to "declare" the ISO setting you adopted for those photos ....; was it 160 for Brian's ? and what for Arpoador' s ? ). Hard time for those unlucky sensors, though, both having to deal with Leica lenses....
But what would you expect to see when the opposite were true ? I mean, if a sensor out-performed an objective ? Would there be any tell-tale symptoms, or the resulting image would just look "out of focus" ? Based on Osuna&Garcia conclusions, I suspect some camera might already have been driven into the desert by the pixel race, even though - of course - it would be stupid to put a good (= costly) sensor and processor behind a poor lens ... (and that extra-pixels may be used to improve the image anyway, as it seems: see Luminous landscape ).
 
Back
Top