Limits of resolution, a simple demo

[Erik Kaffehr]

A question asking what pixel pitch would be needed to resolve the best lens available fully.

Here follows some analysis.

An obvious starting point is to find credible data for truly excellent lens up to very high frequencies. Lensrentals has shared some data on some lenses going up to 240 lp/mm in this article:

More Ultra High-Resolution MTF Experiments

GEEK ALERT!! Let’s be absolutely clear; this is not a practical or useful article. It won’t help your photography or cinematography become better. It won’t help you choose equipment any time in the next couple of years. It won’t provide any fodder for your next Forum War. It’s just a geek...

www.lensrentals.com

I used the data for the Sigma 135/1.8 at f/4 for this example:




The black line shows the MTF data for the Sigma 135/1.8 at f/4 from Lensrentals.

In the figure above, we have a blue line which represents the MTF of the pixel, assuming 100% fill factor at 4.5 mm pixel pitch.

The MTF of the lens and sensor combined is the product of black and the blue curves shown as bottom blue line.

Just as control, I happened to have a slanted edge MTF from one of my better lenses. It is a bit 'too good to be credible'. It is shown in red and it is very close to the calculated system MTF for the Sigma 135/1.8 on the 4.5 micron sensor.

Now, let's look at that Sigma on a 3.8 micron pitch sensor, like the GFX 100 models or the X2D:





Here we can see that at 132 cy/mm, which is the Nyquist limit, we have around MTF = 0.31, which is quite a lot.



Going to 2.1 micron pitch, which is about as long that Lens rentals data goes we would get:





At 2.1 micron pitch we will still have like 18% MTF (that is 0.18), this will still be able to yield aliasing.

Reducing the pixel size will have an effect at low frequencies, but that effect will be very small. Looking at 40 cy/mm we see an increase from 0.803 to 0.825 goinf from 3.8 microns to 2.1 microns.

My take would be that we need far more pixels than we have now to fully resolve the best lenses we have. But, extracting the optimal image quality from lenses may be hard enterprise.

Best regards

Erik

--
Erik Kaffehr
Website: http://echophoto.dnsalias.net
Magic tends to disappear in controlled experiments…
Gallery: http://echophoto.smugmug.com
Articles: http://echophoto.dnsalias.net/ekr/index.php/photoarticles

 

[Jared Willson]

While I agree with you in principle, your last sentence is key. The conditions under which one can extract all that the lens can offer--even if one could arbitrarily pick an optimal pixel pitch--are rare. You need a near absence of motion--either camera or subject--over ridiculously small angles which means surprisingly high shutter speeds to ensure there is very, very little motion. You need to fully address vibration in the system. If I recall, Jim, in at least one of his tests, found that the self time beep in one of the GFX's produced vibration that was visible as motion blur with at least one GFX lens. Was in the 110? I can't remember. Resonant frequencies would likely come in to play depending on the lever arm of the lens (assuming the tripod mount was to the camera). Heaven forbid there be a breeze blowing over your camera/tripod. You would certainly need to an account for atmospheric conditions, possibly even indoors. I know telescope lens makers occasionally debate the importance of having one's lens pointed vertically on the lab bench rather than horizontally when performing interference testing on the lens--just to ensure the column of air involved even indoors is not impacting the results. Even glass flexure under gravity might be relevant. Lenses are only supported at the edges, after all. And what about depth of field and the associated circles of confusion? If you are really trying to get the most out of a lens anything that isn't right in the plane of focus is going to show less resolution than the lens is capable of.

Basically, we are already at the point where I suspect our cameras would provide more resolution if only we improved our technique or could organize our compositions and weather conditions to our convenience.

 

[Schrodingers cat]

if only we improved our technique

I'm pretty sure i'n not going to be improving my technique enough for anybody to notice.

That's not a comment on the high level of my technique, it's a comment on my lack of ability

 

[kciuk aka Ken60]

I quite agree, and your explanation reminds me of my attempt to explore the multi exposure mode of the GFX 100S2 . I was forced , after several failed attempts outdoors, to produce a near working usage with a very weighty tripod standing on a flag stone floor , photographing a cast iron hearth . There were queries over any vehicular activity on a near bye road causing vibrations to this rigid setup through the earth beneath.

Outdoor attempts were failed by a slight wind causing both camera and tripod resonances, defeating the micro pixel accuracy of the registration.

Note the ramblings above never blamed or mentioned the lens used in either the partial success or the failures. I decided that there was still tonnes of juice to squeeze from the 100 I already have, and honestly that is about technique with the GF45 - 100 f4 , and not even moving into the realms of the GF110 f2.

 

[JimKasson]

if only we improved our technique

I'm pretty sure i'n not going to be improving my technique enough for anybody to notice.

That's not a comment on the high level of my technique, it's a comment on my lack of ability

I think good technique is not a function of innate ability. I think it's a function of:

• desire
• acquisition and internalization of skills,
• practice
• persistence
• diligence

 

[kciuk aka Ken60]

Or a lucky error that you learn to repeat and call intentional.

 

[DavidMillier]

if only we improved our technique

I'm pretty sure i'n not going to be improving my technique enough for anybody to notice.

That's not a comment on the high level of my technique, it's a comment on my lack of ability

I think good technique is not a function of innate ability. I think it's a function of:

• desire
• acquisition and internalization of skills,
• practice
• persistence
• diligence

Very much in line with Art and Fear which I have just read for the second time.

At my photography group, the rather low overall standard of image making isn't well explained by a lack of ability, rather it is the result of lack of attention to detail and visual awareness. And that stems fundamentally from a lack of motivation to do the very best you can, and a tendency to dismiss weak efforts as the result of a lack of talent in order to avoid putting in the necessary effort and commitment.

I'm now working with a small group of committee members to attempt to encourage group members to become more diligent and committed to producing better work via a slow and gentle introduction to basic editing skills and a focus on identifying where existing images could be improved. We've got past the first hurdle by persuading people to start properly analysing the strengths and weaknesses of their images rather than trying to sidetrack the group by waffling on about their holidays and where they shot the images. Our hope is that we can eventually lead people to accept that the route to competent photography is primarily effort, not talent, there is no Royal Road, and to encourage them to start putting in that effort.

--
2024: Awarded Royal Photographic Society LRPS Distinction
Photo of the day: https://www.whisperingcat.co.uk/wp/photo-of-the-day-2025/
Website: https://www.whisperingcat.co.uk/wp/
DPReview gallery: https://www.dpreview.com/galleries/0286305481
Flickr: http://www.flickr.com/photos/davidmillier/ (very old!)

 

[John Crowe]

I hope that this is the case. My own experience throwing more and more pixels at my good lenses over the past eighteen years and always getting better results tends to reinforce the point. However, no camera manufacturer seems convinced that more sensor resolution is required, and most photographers don't seem to want more either.

 

[kciuk aka Ken60]

I'm now working with a small group of committee members to attempt to encourage group members to become more diligent and committed to producing better work via a slow and gentle introduction to basic editing skills and a focus on identifying where existing images could be improved.

I wonder with all the dosh ( brit for $$$$$) invested in toys , if there might be an interesting series of posts by the well versed, on post processing paths/tricks/methods or routines that in the belief of the poster adds to the production of their images.

Possibly not limited to even looking at image styles of well known or exemplary practitioners and how that might be achieved. Perhaps a series that could add interest to the wet months ahead.

 

[Schrodingers cat]

if only we improved our technique

I'm pretty sure i'n not going to be improving my technique enough for anybody to notice.

That's not a comment on the high level of my technique, it's a comment on my lack of ability

I think good technique is not a function of innate ability. I think it's a function of:

• desire
• acquisition and internalization of skills,
• practice
• persistence
• diligence

Then I restate my comment :

"I'm pretty sure i'n not going to be improving my technique enough for anybody to notice.

That's not a comment on the high level of my technique, it's a comment on my lack of

• desire
• acquisition and internalization of skills,
• practice
• persistence
• diligence"

Having said that, none of the above seems to detract from my enjoyment of and immersion in the photography hobby.

--
Personal travel snapshots at https://www.castle-explorers.com
1. Making good decisions is generally the result of experience.
2. Experience is generally the result of making bad decisions.
3. Never underestimate your capability for doing incredibly stupid s**t.

 

[Erik Kaffehr]

The thread was intended to be a partial answer to this question:

https://www.dpreview.com/forums/post/68423495

Best regards

Erik

 

[DavidMillier]

The thread was intended to be a partial answer to this question:

https://www.dpreview.com/forums/post/68423495

Best regards

Erik

Good work!

 

[emdundronald]

I always wonder why movies look so sharp

Maybe switching a consumer camera to 8K and post processing to assemble a hirez image is a viable alternative to still-image pixel-shift tricks or even to heavy tripods

I'm now working with a small group of committee members to attempt to encourage group members to become more diligent and committed to producing better work via a slow and gentle introduction to basic editing skills and a focus on identifying where existing images could be improved.

I wonder with all the dosh ( brit for $$$$$) invested in toys , if there might be an interesting series of posts by the well versed, on post processing paths/tricks/methods or routines that in the belief of the poster adds to the production of their images.

Possibly not limited to even looking at image styles of well known or exemplary practitioners and how that might be achieved. Perhaps a series that could add interest to the wet months ahead.

 

[Erik Kaffehr]

I always wonder why movies look so sharp

Maybe switching a consumer camera to 8K and post processing to assemble a hirez image is a viable alternative to still-image pixel-shift tricks or even to heavy tripods

My guess would be:

• Long viewing distance
• Lenses that are optimized to achieve high MTF at low frequencies
• Well designed OLP filters
• Good postprocessing

Best regards

Erik

 

[SeeingSomeThingsMissingMany]

My take would be that we need far more pixels than we have now to fully resolve the best lenses we have. But, extracting the optimal image quality from lenses may be hard enterprise.

Best regards

Erik

Nice one Erik.

[ o ]

Perhaps a poser for ProfHankD.

As Gfx100 has "merely" 100MP which lenses (adapted) without breaking the bank would Gfx100 sensor resolve excellently/fully.

--
Photography after all is interplay of light alongside perspective.

 

[tex]

I think good technique is not a function of innate ability. I think it's a function of:

• desire
• acquisition and internalization of skills,
• practice
• persistence
• diligence

I'll add one more (once we all agree on getting as close to optimal as we can as another baseline):

• managing deficiencies

In our realm, photographing in museums and photographing artwork---and this is a wild range of stuff, from paintings, works on paper, transparent things, Dan Flavin sculptures , stuff in the studio but then in situ in galleries with sometimes multiple light sources, outdoor stuff, and more and more, etc., we're often just...confounded.

There are some very strict standards, and then there is "rubber hits the road" and what will look "natural" as a record of an exhibition (archival shots in the studio can be far more dialed in...sometimes!).

All of Jim's bullet points are the best start to be sure. But then...

 

[James O'Neill]

My take would be that we need far more pixels than we have now to fully resolve the best lenses we have. But, extracting the optimal image quality from lenses may be hard enterprise.

Best regards

Erik

Nice one Erik.

[ o ]

Perhaps a poser for ProfHankD.

As Gfx100 has "merely" 100MP which lenses (adapted) without breaking the bank would Gfx100 sensor resolve excellently/fully.

I understood that the image captured could not resolve more detail than the lens, nor more detail than the sensor and as sensor res increased the final image approached the lens res (and vice versa).

I have watched plenty of arguments as to why the 1/output_res = 1/recording_res + 1/Lens_res rule from film either holds for digital or doesn't hold and I don't want to start another one.

But everything I've learnt says in theory to fully resolve the image of any lens needs infinite sensor resolution. You also get to a point where you've resolved - lets say 99% of a very good lens's detail with a large number of pixels, and a lot more only gets you to 99.1%. So on the one hand any increase in Pixels means you can capture a finer detail, but on the other you may not be able to see the difference.

 

[JimKasson]

But everything I've learnt says in theory to fully resolve the image of any lens needs infinite sensor resolution.

Reference? I don't think that's true at all. If that were the case, Harry Nyquist was wrong.

 

[kciuk aka Ken60]

Surely the real answer goes beyond a couple of slide rules playing my number is bigger, and extends into the real world of where the image will be used. Can whatever resolution can be achieved be seen in that rendition, because anything beyond usage is money wasted.

 

[JimKasson]

Surely the real answer goes beyond a couple of slide rules playing my number is bigger, and extends into the real world of where the image will be used. Can whatever resolution can be achieved be seen in that rendition, because anything beyond usage is money wasted.

You forget that there is no lower bound to aliased frequencies.