[Blue]

I don't care about the amount of light that comes in the lens.

I care about the amount of light that is collected by the sensor... and of course, the FOV, DOF, Bokeh, SNR, Resolution...

I didn't say you did. All I have done is stated that regardless of sensor size, the physical focal length and aperture of a lens does not change.

[Ikarus]

How would this best be explained? CoC? This is a good case for FF then, as apparent IQ would be better for all FF lenses on an FF camera vs the existing APS-C cameras, which mean better quality images from the same lenses.

It still all boils down to cropping as the key factor. When judging the quality of the lens on a cropped format, the part of the image circle outside of the sensor surface is completely irrelevant, so as far as this use case is concerned, it might as well be pitch black. On the other hand, the part of the lens inside the sensor surface is now under increased scrutiny. If the number of pixels is kept constant between formats, there is a higher center resolution power asked of the lens in the cropped case. If the lens still outresolves the sensor in this area, its performance will tend to be better on the cropped format than on FF, because the weaker corners will be cropped. If it no longer does, the lens will perform poorly on the cropped format.

[lytrytyr]

If the number of pixels is kept constant between formats, there is a higher center resolution power asked of the lens in the cropped case. If the lens still outresolves the sensor in this area, its performance will tend to be better on the cropped format than on FF, because the weaker corners will be cropped. If it no longer does, the lens will perform poorly on the cropped format.

The paradoxical feature of the Nikon lens I referenced
(https://www.pentaxforums.com/forums/pentax-slr-lens-discussion/211066-field-v...ml#post2231021)
is that the corner performance is worse on 1" than on 24x36mm!

The center resolution holds up on all the formats.

[tacticdesigns]

A buddy of mine keeps asking me about different "next" cameras for him. He's a Canon user, so I dropped the idea about maybe the next camera he should get is a FF camera. He asked me why, so I rhymed off all the standard reasons, lower noise, more shallow depth of field . . . he asked me how that worked with the shallow depth of field . . . after a few beers I think we finally figured out a way to think about it . . .

Let's say you have a full frame camera (Canon 5D) with a 50mm lens on it. You're taking a picture of someone 10 feet away at f1.8. According to this on-line depth of field calculator . . . your depth of field is about 9.4 ft - 10.7 ft (or about 1.29 ft deep).

Online Depth of Field Calculator

Ok. So a couple of nasty kids dart out from the bushes and super glue your feet and your subject's feet to the ground. Neither you or your subject can move.

Then worse, they grab the full frame camera from your hands and give you an APS-C (Canon 60D) camera with the 50mm lens. (With a plastic mount.)

You decide to take another picture of the person, but realize your field of view has changed and you're not getting the composition you really wanted. So you beg the kids to at least swap out the 50mm lens for a 35mm lens. (Equivalent field of view.) You take your picture, but its not quite the same.

So you pull out your phone and go back to the on-line depth of field calculator.

With a Canon 60D at 35mm f1.8 at 10 feet to subject your depth of field is about 9.23 ft - 10.9 ft (or about 1.68 ft).

I think I blew a piston thinking through this. I think the best way to approach this is after a few beers.

[ElJamoquio]

the best way to approach anything is after a few beers.

Fixed that for you.

[Ash]

The paradoxical feature of the Nikon lens I referenced
(FoV, focal length and true aperture of lenses)
is that the corner performance is worse on 1" than on 24x36mm!

That is key.

Let's say you have a full frame camera (Canon 5D) with a 50mm lens on it. You're taking a picture of someone 10 feet away at f1.8. According to this on-line depth of field calculator . . . your depth of field is about 9.4 ft - 10.7 ft (or about 1.29 ft deep)

With a Canon 60D at 35mm f1.8 at 10 feet to subject your depth of field is about 9.23 ft - 10.9 ft (or about 1.68 ft).

Yes, that is correct.
And it would be a similar effect if the 50mm lens were kept constant, but the photographer wasn't glued to the ground (the subject still is) and could stand back to get the same framing on the APS-C camera as he did with the FF camera. The lens is the same but the camera-to-subject distance has changed, which alters the DoF, and thus the result.

It's quite clear that the physical properties a lens do not change regardless of what format of image capturing they are subjected to. But here, we're moving on to what happens to a lens when put on different formats in terms of IQ and SNR, etc.

[tacticdesigns]

But here, we're moving on to what happens to a lens when put on different formats in terms of IQ and SNR, etc.

Ah . . . I missed that up top.

I don't think I got enough brain cells (or beer) left to understand this stuff, but I've subscribed to the thread. This is pretty cool!

Does the angle at which the light hitting the sensor play into this any?

I remember coming across some discussions about the angle at which the light hits the sensor. The shape of the photosites. Or the alignment of the AA filter to the sensor sight.

The lens is the same, but the differences in sensor design might record different light from the lens based on number of pixels, collector design and such.

?

[Ash]

I remember coming across some discussions about the angle at which the light hits the sensor. The shape of the photosites. Or the alignment of the AA filter to the sensor sight.

Not sure, but I don't think it would make a significant enough difference until you get to a fish eye lens like the 10-17, which brings in light from a 180 degree FoV at the wide end...

The lens is the same, but the differences in sensor design might record different light from the lens based on number of pixels, collector design and such

That's getting complicated. But a nice theoretical discussion point. If with the 35mm lens on an APS-C there is a larger FoV with broader range of incident light hitting the lens elements compared with a 50mm lens on a FF camera, then conceivably there are slightly different properties to the light hitting the sensor. But I'd think that's more a property of lens design rather than a direct result of differing angles of incident light. The rear element of the lens is designed to align the photons to a flat plane for a 2D rendition on the sensor, and as such the lens manages the different angles of incident light by its lens design. That's my understanding.

[Lowell Goudge]

In order to stimulate more technical, intellectual discussion on the optical properties of lenses on different formats, I have started this thread stemming from the initial discussions here: https://www.pentaxforums.com/forums/general-photography-industry/209882-next-...pentax-16.html

Now, aperture on a lens, let's say a FF lens like the FA 31mm f/1.8, should theoretically be constant between all the formats used to shoot with this particular lens. i.e. an f/5.6 setting on the lens should mean f/5.6 despite using the lens on an APS-C, Q or FF format

Yes

. The crop factor alters the apparent aperture according to the FoV, increasing DoF with smaller formats than FF (considering the crop factor of FF is 1.0).

not true, the depth of field for an aperture is constant (or nearly so) for the main subject, if you correct the shooting position so that the main subject is the same size in the viewfinder, for example if shooting a head and shoulder shot, with 2 different formats and changing shooting position, for each format to get the same coverage of frame. Depth of field is a function of magnification ratio and aperture nothing more.

But the actual aperture of the lens remains the same.

yes

Is there any more to the optical qualities of the lens than this, or does the different sized formats actually somehow change the lens's characteristics?

the whole issue with changing formats is simply a function of shooting distance magnification ratio and aperture, but I will point out one thing. With wider angle lenses because the magnification of background is a function of distance /focal length the background magnification is much less making things appear "sharp" it is one reason why bokeh looks better with tele lenses, the background is just. As Out of focus when enlarged , but with a tele lens the higher background magnification makes the out of focus details bigger in proportion to the subject and appears less confused

[Ikarus]

The paradoxical feature of the Nikon lens I referenced
(https://www.pentaxforums.com/forums/pentax-slr-lens-discussion/211066-field-v...ml#post2231021)
is that the corner performance is worse on 1" than on 24x36mm!

That just means that the lens resolution improves towards the center at a slower rate than the pixel density of the crop sensors increases along the corner areas. I don't see a paradox here.

The center resolution holds up on all the formats.

Unless I'm missing something, the center performance goes down accordingly.

[Laurentiu Cristofor]

Lens equivalence is explained at length (and there are summaries too) in this article.

As for aperture, it is worth noting that it is different from light transmission, for which t-stops are used.

Here's another less understood aspect. Aperture is determined based on focal length. But the focal length is determined based on focusing at infinity - at close focusing distances, the focal length actually gets longer, with the result that the effective f-stop is higher than what is selected on the lens. This is why some macro lenses also had f-stop markings next to magnification ratios on their barrels. I understand some modern macro lenses (Nikon) actually display the effective aperture as you focus closer.

[Ash]

the depth of field for an aperture is constant (or nearly so) for the main subject, if you correct the shooting position so that the main subject is the same size in the viewfinder, for example if shooting a head and shoulder shot, with 2 different formats and changing shooting position, for each format to get the same coverage of frame. Depth of field is a function of magnification ratio and aperture nothing more

Understood. The example later discussed was that of using FoV-equivalent lenses for the same subject distances and getting slightly different results. The same lens at the same subject to camera distance and then cropping the FF image to achieve the same magnification as the APS-C image is essentially the same.

This might be helpful:

[Fogel70]

The FA 77/1.8 lens will still let in the same amount of light at ISO 100 and 1/250. This size of the sensor/film does not change that fact.

It's kind of obvious that the lens do not physically change depending camera it is used on, and let the same amount in on all cameras, but the sensor size change how much of the light is used for exposure.

Or do you think that if you put one small and one large bucket out in the rain, that both of them will collect the same amount of water for a given time?

[lytrytyr]

I don't see a paradox here.

The behavior is the opposite of what you suggested in your earlier post:

If the lens still outresolves the sensor in this area, its performance will tend to be better on the cropped format than on FF, because the weaker corners will be cropped.

Unless I'm missing something, the center performance goes down accordingly.

It reaches "Excellent" in each case,
using Photozone's criteria for comparison.

[Ash]

Or do you think that if you put one small and one large bucket out in the rain, that both of them will collect the same amount of water for a given time?

Good analogy, but I don't think anyone's disputing this. It's recognised that light intensity hitting the sensor is the same regardless of format, otherwise exposure values would be different, but there are other discussions occuring about f/2.8 not being f/2.8 and lens qualities (like IQ) being different between formats.