Can lenses get much better??

CheshireCat wrote in post #18213370
The physical limit is diffraction, but software could go a bit beyond knowing the PSF of the lens.

The good thing is that diffraction is limited by the aperture, therefore f/1 lenses would help.

Oh boy, the razor thin DOF fanatics would love f/1, but what about the landscape shooters who would like to use f/11 or f/16 for deep DOF?! So on a more practical level...

1. If a 24mm tall image is enlarged to 48" tall, we have a 20X image.
   
2. If the print could deliver 10 line-pairs/mm of detail, we need about 20 line-pairs/mm to resolve that.
   
3. So we have a need for about 9600 pixels in the picture height, and
   
4. a lens that can deliver 4800 line-pairs/picture height

Wilt wrote in post #18217120
Oh boy, the razor thin DOF fanatics would love f/1, but what about the landscape shooters who would like to use f/11 or f/16 for deep DOF?!

Landscape shooters usually focus at infinity and stop down for image quality, not DOF.
A 21mm f/1 wide angle lens focused at infinity resolves details as small as 21mm, which is not usually a limitation, and in the few cases it is, there's focus stacking.

Wilt wrote in post #18217120
...we already have lenses that deliver about 3500 line-pairs/picture height, so we have room for another 35% improvement in lenses. Since we started with the assumption of 10 line-pairs/ mm, most pro usage would not need much more.

It is not only about print/display size.

Today we are carrying around huge and heavy lenses for the sake of IQ.
What about a small 21mm f/1 prime lens resolving a 1 GigaPixel sensor ? We could then crop a 100mm equivalent shot with still great IQ. No mechanical zoom, no lens changes, no long and heavy tele lens.

Lenses and sensors can get much better. And they will.

CheshireCat wrote in post #18212977
A couple of them were even radioactive.

Yes, I have few such, a sister of the Kodak Aero Ektar 6" and another one I can not remember the name.

CheshireCat wrote in post #18217252
...
A 21mm f/1 wide angle lens focused at infinity resolves details as small as 21mm, which is not usually a limitation, and in the few cases it is, there's focus stacking...

21 mm details where? For sure not at infinity, and likely it is better than 21mm at the sensor.

ZoneV wrote in post #18219062
21 mm details where? For sure not at infinity, and likely it is better than 21mm at the sensor.

21mm details in the world.
If you focus a 21mm aperture at infinity, the smallest "world pixel" the lens can resolve is a circle with 21mm diameter, no matter how far it is from the camera.
So, if your "world pixel" is 1 meter away, it will be mapped on many sensor pixels (blurred), but if it is - say - 100m away, it will be smaller than a pixel.

CheshireCat wrote in post #18217252
Landscape shooters usually focus at infinity and stop down for image quality, not DOF.
...

Lenses and sensors can get much better. And they will.

Thanks for the tip.

Seems I've been doing it wrong for over 40 years, with pretty decent results. I stop down, in landscape shots, specifically for DOF. I don't believe I've ever focused a landscape shot at infinity, or used a wide open aperture. There is absolutely no point in focusing a wide angle lens at infinity. Example: 6D/17-40 @ 17mm & f/11: focus at 5' and everything from 3' to infinity is sharply in focus. I can shoot my Rokinon 12mm FE all day and never touch the focus ring. At f/5.6-8, and focused at 4', everything from about 18" to infinity is in focus.

If you focus wide-angle lenses at infinity, in my mind, you don't know how to use them.

I agree with you that lenses will get better.

Being slightly less... errr... "to the point"... than Bassat: I would agree that focusing at infinity for landscape shots would be a really bad idea (you're throwing away a lot of your depth of field). Hyperfocal distance is usually your friend; though I also suspect many hardened landscape pros may well have invested in a TS lens; which will somewhat alleviate worries about diffraction from needing to stop down for DOF.

CheshireCat wrote in post #18219420
21mm details in the world.
If you focus a 21mm aperture at infinity, the smallest "world pixel" the lens can resolve is a circle with 21mm diameter, no matter how far it is from the camera.
So, if your "world pixel" is 1 meter away, it will be mapped on many sensor pixels (blurred), but if it is - say - 100m away, it will be smaller than a pixel.

So this is your version of the diffraction theory?
In the diffraction theory I know the object resolution is given as a aqngular resolution. The further away, the smaller the details. A good lens should have a better resolution than their eintrance pupil diameter at near distances. And no 21mm entrance pupil lens resolves 21mm object details at infinity (with visible light), even not at 100.000km.

ZoneV wrote in post #18219502
So this is your version of the diffraction theory?
In the diffraction theory I know the object resolution is given as a aqngular resolution. The further away, the smaller the details. A good lens should have a better resolution than their eintrance pupil diameter at near distances. And no 21mm entrance pupil lens resolves 21mm object details at infinity (with visible light), even not at 100.000km.

Diffraction has nothing to do with this.
What I am simply saying is that if you have a perfect lens (with no diffraction) with an effective aperture of 21mm focused at infinity, on a perfect sensor, it will be ablle to resolve any detail bigger than 21mm, no matter how far that detail is from the lens.

Landscape photos usually need critical sharpness at infinity (e.g. mountains), therefore I focus at infinity. If I also need blades of grass in the foreground to be resolved, I stop down to an effective aperture of less than a blade of grass width (e.g. stop down my 21mm lens to f/11 = about 2mm). Yes, the grass near the camera will not usually have pixel-level sharpness, but it is not the main subject after all.
The classic hyperfocal method and its scale on your lenses are inappropriate for modern digital sensors, to the point that novices often wonder if their lens is broken because they use hyperfocal yet the background is blurry. If you do the math, pixel-level background sharpness on modern sensors using hyperfocal distance may require stopping down the lens well beyond diffraction aperture.

If you want critically sharp background and critically sharp foreground, you need focus stacking.

CheshireCat wrote in post #18219742
Diffraction has nothing to do with this.
What I am simply saying is that if you have a perfect lens (with no diffraction) with an effective aperture of 21mm focused at infinity, on a perfect sensor, it will be ablle to resolve any detail bigger than 21mm, no matter how far that detail is from the lens.

Bunk. In the first place diffraction is a function of any lens. There is no perfect (no diffraction) lens. It is not possible. Double-bunk. There is no perfect sensor. Again, not possible. While you are playing in theoretical-land, others are using real cameras to take real pictures in the real world. No real camera, using any real lens you choose, will resolve a 21mm object at, oh, lets say 3 miles.


Landscape photos usually need critical sharpness at infinity (e.g. mountains), therefore I focus at infinity. If I also need blades of grass in the foreground to be resolved, I stop down to an effective aperture of less than a blade of grass width (e.g. stop down my 21mm lens to f/11 = about 2mm). Yes, the grass near the camera will not usually have pixel-level sharpness, but it is not the main subject after all.

What? You can't possibly do any better than pixel-level sharpness. If all you are worried about in your landscape photos is focus at infinity, you are taking some relatively boring landscape photos. A critically sharp mountain in the distance is a boring photo. A critically sharp mountain in the distance with some stream-bed rocks, or wildflowers in the foreground (also in focus), will make a much more interesting shot. And that is not possible if you are focused at infinity.

The classic hyperfocal method and its scale on your lenses are inappropriate for modern digital sensors, to the point that novices often wonder if their lens is broken because they use hyperfocal yet the background is blurry. If you do the math, pixel-level background sharpness on modern sensors using hyperfocal distance may require stopping down the lens well beyond diffraction aperture.

Evidence, please. I've been using HFD for 40 years. It works as well on any digital camera I've ever used as it did on 6x6 film. I'm only guessing here, but it seems you have a poor grasp of HFD, and landscape photography in general.

If you want critically sharp background and critically sharp foreground, you need focus stacking.


This is so far from the truth it isn't even wrong. It is absurd. Please, don't take my word for it. Look at some well-done landscape photos.

Internal comments.

.

CheshireCat wrote in post #18217252
Landscape shooters usually focus at infinity and stop down for image quality, not DOF.

I am compelled to disagree with this statement. Emphatically disagree, as it is incorrect on two levels.

.

Here's one of the photos Bassat defines boring

https://upload.wikimed​ia.org …s_and_the_Snake​_River.jpg

About the rest, do your own math homeworks

Tom Reichner wrote in post #18219786
.
I am compelled to disagree with this statement. Emphatically disagree, as it is incorrect on two levels.
.

Fair enough, I was probably too harsh to make my point.
Let's just say that if all the important elements are at infinity, there is no need to stop down for DOF.

CheshireCat wrote in post #18219742
Diffraction has nothing to do with this.
What I am simply saying is that if you have a perfect lens (with no diffraction) with an effective aperture of 21mm focused at infinity, on a perfect sensor, it will be ablle to resolve any detail bigger than 21mm, no matter how far that detail is from the lens.

Landscape photos usually need critical sharpness at infinity (e.g. mountains), therefore I focus at infinity. If I also need blades of grass in the foreground to be resolved, I stop down to an effective aperture of less than a blade of grass width (e.g. stop down my 21mm lens to f/11 = about 2mm). Yes, the grass near the camera will not usually have pixel-level sharpness, but it is not the main subject after all.
The classic hyperfocal method and its scale on your lenses are inappropriate for modern digital sensors, to the point that novices often wonder if their lens is broken because they use hyperfocal yet the background is blurry. If you do the math, pixel-level background sharpness on modern sensors using hyperfocal distance may require stopping down the lens well beyond diffraction aperture.

If you want critically sharp background and critically sharp foreground, you need focus stacking.

Diffraction is a matter of physics, with light rays trying to bend thru the small openining of the diaphram. Even with f/2 selected for a lens there is some diffraction effect; diffraction simply INCREASES as f/stops get smaller.

The problem of Hyperfocal distance is simply that the calculators almost all use the very flawed standard of human vision that is assumed by the 'manufacturer standard'...about only 1/3 as good as someone with 20/20 vision. That is why Hyperfocal use seems to fail all so often! For example, if my human subjects were 50' away and the Washington Monument were 550' behind them...

• for 135 format with 35mm lens, 'manufacturer standard' calculation assumptions result in Hyperfocal distance of 22.4' at f/5.6, with a calculated DOF zone of 11' - 3533'
  
• for 135 format with 35mm lens, '20/20 vision' assumptions result in Hyperfocal distance of 67.4' at f/5.6, with a calculated DOF zone of 33.7' - 70697'
  

About diffraction: of course it is a physical thing and there is no such thing as a diffraction-free lens, but in optics it is often useful to talk about a "perfect lens" when discussing other matter not involving diffraction.

About the HFD scales displayed by our lenses, they are simply useless for digital photography.
The right equation for HFD is the one that keeps the circle of confusion smaller than the sensor pixel size. A HFD scale would then vary depending on the sensor density.

Your examples about the viewer with 20/20 vision will be even more extreme if the viewer is pixel-peeping at 1:1 on a computer display. That is what I am talking about when I say "critically sharp": pixel level sharpness.
And my next point is that if you want pixel-level sharpness at infinity using HFD, you may need to stop beyond diffraction (and at that point, diffraction becomes painfully real).

When I have the time, I'll shoot some example with the 24/1.4 to show you guys how ridiculous HFD scales are nowadays.

CheshireCat wrote in post #18219790
Here's one of the photos Bassat defines boring

https://upload.wikimed​ia.org …s_and_the_Snake​_River.jpg

About the rest, do your own math homeworks

This photo is a prime example of what you say NOT to do. Had AA focused on the mountains, there would be no detail in the trees/river. There isn't a ton of foreground here, but what is there adds to the photograph.

You also tout shooting landscapes wide open. The evidence you present proves you wrong. I am not familiar with this particular photo, but let us suppose AA was using one of his 8"x10" plate cameras. A film that big is capable of INCREDIBLY thin DOF. Notice the INCREDIBLY DEEP DOF in this shot. My guess is that AA was stopped down considerably for this shot.

Wide open and focused at infinity is not how folks who know what they are doing shoot landscapes. Your evidence is contrary to your point. Guess again.