Do different format sizes produce different images?

[tomfrh]

One way of thinking about it is that there is only a certain amount of detail resolution that a lens can provide, and if you spread it out over a larger image circle it is hard to provide the same line-pairs per millimeter at the film plane (a conceptual description, which may not be acceptably accurate description to an optical physicist/engineer). So companies like Rodenstock freely admit and publish MTF charts for large format lenses which shows exactly that!...fewer line-pairs per millimeter from LF lenses than the MTF charts for 135 format lenses.

Yes, I agree with all that. What I was querying was that thuggins said it all evens out and different formats lenses project the same number of line pairs. However you said, no, the LF comes out with twice as many line pairs as the 35mm.

After you take the 44 line-pairs per millimeter on 4x5 film, and enlarge it by 4X, you end up with 11 line-pairs per millimeter of 8x10" paper print; similarly 80 line-pairs per millimeter at the 135 film, after enlarged by 8.5X, results in 80 / 8.5 line-pairs or 9.4 line-pairs on 8x10" paper print.

You enlarge the 4x5 by a factor of 2x get to 8x10, not 4x.

 

[wiltw]

Yes, I agree with all that. What I was querying was that thuggins said it all evens out and different formats lenses project the same number of line pairs. However you said, no, the LF comes out with twice as many line pairs as the 35mm.

My explanation is also what Rodenstock says! do you doubt their veracity? So the other explanation must have a hole in it, or some other reason not considered by Rodenstock.

You enlarge the 4x5 by a factor of 2x get to 8x10, not 4x.

BrainfArt! I often describe concepts in terms of the 16x20" print, to better show the advantage of FF over APS-C in digital cameras -- especially since so many digital photograpers think only of the sensor and do not factor in optical performance of the lens being affected by enlargement (as if 'enlargement' did not occur in the print making!), so the brain got mixed up talking about 8x10 prints

 

[tomfrh]

My explanation is also what Rodenstock says! do you doubt their veracity?

I don't know the answer. My queries to you were not to say you are wrong - I was simply wondering about the discrepancy between your and Rodenstocks claim that LF wins (which is supported in detail here http://www.normankoren.com/Tutorials/MTF1A.html), and thuggins claim that they tie (which is supported in detail here http://www.photodo.com/topic_138.html),

BrainfArt! I often describe concepts in terms of the 16x20" print

Ok. With those corrected number it comes out the same as what you said before, with the LF image being about twice as detailed as 35mm when scaled to 8x10. As I noted, thuggins (and that link) say they are same level of detail when both scaled to 8x10



None of this resolution stuff is what I was originally asking about - but it's interesting nonetheless!

 

[klownshed]

This is too much like the 'equivalence' arguments on the 'd' forums.

Imagine a large format camera existed with a zoom lens that can take different format film backs and you have the same film type for 35mm and 120.

You can then take two photos, one on each format, with the resulting image as close as possible with respect to Field of view and depth of field by adjusting the foca length and aperture accordingly.

You will then get two negs that are as similar as possible reducing as many variables as you can.

When you enlarge the two negs to the same size print there will be differences due to the grain and the effects of using different focal lengths.

It would be a similar difference you would get if you just used one camera and one film size and did the same test with a wide angle and telephoto, cropping the wide angle photo to match the field of view of the other photo.

The two photos will look different.

The difference in perspective distortion will be less pronounced the further away the subject is.

 

[Steve Smith]

The larger the format, the more shallow the depth of field, assuming the same angle of view and aperture setting.


Steve.

 

[tomfrh]

differences due to ... the effects of using different focal lengths

The two photos will look different.

What differences?

 

[tomfrh]

The larger the format, the more shallow the depth of field, assuming the same angle of view and aperture setting.

Assume equivalent focal length and aperture.

 

[klownshed]

They will look different because different focal lengths have different perspectives. As I said above, this difference diminishes the further the subject is away.

Imagine a wide angle lens and very close subject. You will get a large degree of perspective distortion. If the subject is a person you will get a very distorted face. If you crop that image to match the field of view of a portrait taken with a short telephoto the perspective effect will be very different. The portrait taken with a loner lens will be (far) less distorted. Which is why people tend to use a short tele for portraits and BT ultra wides (unless you want the effect of an ultra wide!).

So even if the depth of field and field of view match you will get a different perspective, which is more pronounced the closer the subject is.

 

[tomfrh]

They will look different because different focal lengths have different perspectives. .

Focal length doesn't affect perspective. The only thing which affects perspective is camera position.

Perspective is irrelevant to my original question. The cameras are taking the same photo from the same position. The perspective is the same.

 

[Punker]

To answer OP's original question. To me, yes. I noticed the difference in tonality in the move to 4x5. I assume it's because you have so many more silver grains that make up, for instance, the branch of a tree. Perhaps on 135 it would be 1 grain wide, but on 4x5 it would be at least 6 or more. That allows for not only greater detail, but a more smooth gradation of tones across those extra grains. Honestly I don't notice the difference in tonality between 135 and 120, nor in the depth of field. I shoot 120 when I want more detail or a larger print but don't want or need to lug around the 4x5.

 

[removed account4]

the tonality differences all depend on how one processes and exposes the film
it isn't hard to have smooth toned 35mm or 120 format negatives,
just like it isn't hard to get rough 4x5 or larger negatives.
the person using the camera and chemistry just needs to know what they are doing.
there is also similar commentary that 35mm can't be enlarged past 11x14 or 8x10
which couldn't be more nonsensical ( i hope that is a word, and if it isn't it should be )
photographers seem to be the only people on the planet who put their nose to every print they see.

 

[AgX]

They will look different because different focal lengths have different perspectives.

No.
It is the photographer changing his position to and fro the subject that makes the perspective change.

 

[Huub]

Apart from the discussing the technical aspects of the different film formats, for me the differences inthe various formats has more to do in the way you use a camera. I tend to use my 35mm camera's for more spontanious photography, like snap shots and street photography. Mainly subjects where i can move pretty fast and only have a moment to take the shot. I use these camera's always hand held at eye level, with a 20mm and 28mm as favourite lenses.
Medium format i use for more deliberate photography, often portraits of people in situations where there is more time to build up contact with the model. The square 6x6 screen also forces me into a different way of composing the image, as is using the chimney viewer, which makes me bend into a lower viewpoint.
Large format is mostly pretty static photography for me, always working form a tripod. It forces me to think twice of thrice before i set up te camera, compose the image carefully and work in a very thought of way. It makes the pictures less spontanious but comes with a higher technical quality.

So for me it is not only in the technical aspects of the different camera systems, but more in the way i use the different systems.

Then a question: what is the influence of hand holding a camera on the resolution of the resolving power?

 

[wiltw]

I don't know the answer. My queries to you were not to say you are wrong - I was simply wondering about the discrepancy between your and Rodenstocks claim that LF wins (which is supported in detail here http://www.normankoren.com/Tutorials/MTF1A.html), and thuggins claim that they tie (which is supported in detail here http://www.photodo.com/topic_138.html),
Ok. With those corrected number it comes out the same as what you said before, with the LF image being about twice as detailed as 35mm when scaled to 8x10. As I noted, thuggins (and that link) say they are same level of detail when both scaled to 8x10

The photodo article makes a statement that is flawed.

Statement 1
"The normal focal length corresponds to the diagonal of the negative size. The perspective we acquire when using a normal lens is the same perspective that we see when we look at the object with the naked eye. The relationship in size between close and distant objects is the same in reality as it is in the picture. The normal focal length for 35 mm film is 50 mm, for medium format 80 mm, and for 9x12 cm approximately 150 mm."

• I have already dealt with the fact that 50mm is NOT the 'diagonal' but 43mm is, so it by default would NEVER be 'same FOV' as 150mm on 4x5 (or 79mm on 6x6, which are both very close to the true diagonal).
• I have already provided the alternative means of getting identical FOV vertically, when based upon a multiple of the vertical frame dimension and cited concrete examples, rather than a 'convention' such as 'diagonal' which fails to address the issue that one format is square (1:1), one is 1.25:1 (4x5) and one is 1.5:1 (135); when one uses the vertical * multiple, you CAN get truly identical vertical FOV from one camera location for all three cameras. As an aside, if one added 6x9 to the same test (of four cameras) it would still use the identical FL as 6x6 because the verticals do not differ although the 6x9 format is 'overlong' like 135 format.
• Unless the item being photographed occupies the same percentage of the frame (vertically) one is NOT analyzing the true differences in the advantage/disadvatage to using different format sizes; using the vertical * multiple allows one to truly equal the percentage of the frame that the identical subject fills, for all formats. So comparing 6x6 vs. 6x9 formats, 'the convention' of the 'diagonal' would result in putting a 'normal' which is 1.5X longer on the 6x9 camera, and suddenly the subject being photographed is 1.5X taller on the 56mm x 69.5mm 6x9 format frame size, yet 6x9 has NO DIFFERENCE in reality from 6x6 on the basis of tonality or gradation or resolution!

I will readily admit that for a real test, unfortunately it is very difficult to find FL lenses suited for each format which are exact mulltiples, so photodo test use 50mm and 150mm lenses.
If we evaluate the FL against the multiple technique (above) to see the mismatch of that comparison...
50mm = 2.08 x FL, and 150mm = 1.61x * 93mm (the measured short length exposed by a film 4x5 film holder) so the subject's image on 135 will occupy about 1.3x as much of the frame height on the 135 neg.

• If we used a 100mm (2x 'normal') lens on 135, at subject distance of 100' the frame captures a 24' x 36' area, and a 20' tall object fills 83% of the frame height;
• If we used 300mm lens on 4x5, at subject distance of 100' the frame captures a 33.5' x 42' area, and a 20' tall object fills only 60% of the frame height

...20' subject 83% of 135 frame vs. 60% of 4x5 frame, 83% / 60% = 1.38x (close to the 1.3x predicted above; difference is due to the FOV program assumption about 4x6 frame being 90mm rather than the measured 93mm!) While photodo test could have used 24mm FL on 135 test shoot, and 90mm FL on 4x5 test shoot so that comparing those two MTF with the subject being much, much closer to occupying the same percentage of the frame, that test would have been complicated by the fact that the 135 lens test uses retrofocus optical design (to allow sufficient space for the SLR reflex mirror) and I do not know what effect that would have had on ultimate lens performance (if any).​

I will readily admit that the photodo testing does something which is totally UNFACTORED in my analyses so far in this thread...my analyses fail to consider actual lens MTF. However, a typical 135 format normal lens at f/4 would be close to its peak MTF performance as would a typical 4x5 lens at f/16.
Yet there is also the reality that real vs. theoretical differences in on-film performance would be manifest in actual testing using real lenses + film. That likely to be a key reason for any apparent difference between Rodenstock theory and photodo reality! As the photodo article states so aptly,

"Theory and data sheets are one thing, practical photography another, and of course we wanted to see what things looked like 'for real'."​

I would like to see the photodo testing repeated with more effort to make the test target shapes scale proportionally in the formats, and NOT use the simplistic 50mm and 150mm lens, but that might be very difficult to achieve, given the very limted FL available for large format lenses.

 

[Fixcinater]

This is the sort of comment I was originally asking about in the original post. Good to have someone here who experiences it.

How would you describe the differences between the 105/2.4 and equivalent 35mm lenses?

Weird. So, aside from image circle, what's the difference between the 6x7 and 35mm 105s then? Can anyone explain it?

Well, on 35mm film, it has the same sort of difference as between the 35mm Tak and say, a Nikkor 105mm. Comparing the P67 on 6x7 to a 50/1.4 on 35mm, there's a larger visible difference between the characters of the overall print as you are using the entire 67 image circle. That's where my knowledge stops as I'm not a lens designer but it's pretty darn apparent. Looking at the P67 Flickr group, even with all the other variables, it's a distinct look that the 105/2.4 provides.

AFAIK, the P67 105mm is a "normal" design, upsized but similar to the 35mm 50/1.4. The M42 105mm is a telephoto design.

 

[tomfrh]

I will readily admit that the photodo testing does something which is totally UNFACTORED in my analyses so far in this thread...my analyses fail to consider actual lens MTF

Oh! Are you (and Rodenstock) talking about resultant lens+film MTF?, not lens MTF? That changes things a bit. I thought you were talking about lens MTF.

As for your comments about aspect ratio and focal length multiples, I tend to agree with you. I much prefer your idea of using a multiple of height to compare apples to apples, rather than photodods method of diagonals and "normal lenses". I think their method needlessly muddies the water with scaling effects. I don't think it would significantly alter their conclusion that large format and 35mm lens MTF is practically the same (assuming that conclusion is true), but nonetheless it's annoying. Clearly it's leading you to question their conclusions.

 

[wiltw]

Oh! Are you (and Rodenstock) talking about resultant lens+film MTF?, not lens MTF? That changes things a bit. I thought you were talking about lens MTF.

As for your comments about aspect ratio and focal length multiples, I tend to agree with you. I much prefer your idea of using a multiple of height to compare apples to apples, rather than photodods method of diagonals and "normal lenses". I think their method needlessly muddies the water with scaling effects. I don't think it would significantly alter their conclusion that large format and 35mm lens MTF is practically the same (assuming that conclusion is true), but nonetheless it's annoying. Clearly it's leading you to question their conclusions.

Rodenstock mentions nothing about MTF measurement methodology on their lens brochures, whereas photodo (to their credit) took great pains to point out their tests were the practical combination of lens+film. There is a question, with any MTF characterization of 'with which film?', when it is not explicitly stated.

'If the test result by photodo is a 'practically the same' final assessment, it does make you wonder...
If they had neutralized the question of scaling differences so that subjects occupied the same frame height fraction on both formats, would the results have been more clearly in one direction?

 

[tomfrh]

'If the test result by photodo is a 'practically the same' final assessment, it does make you wonder...
If they had neutralized the question of scaling differences so that subjects occupied the same frame height fraction on both formats, would the results have been more clearly in one direction?

I can't see how it would. The differences in scaling methodology are fairly minor compared to the large differences of opinion we are talking about.

Using a 45mm instead of a 50mm wont halve the detail.

 

[Dan Fromm]

Hmm. I used to shoot flowers using flash illumination, ISO 25 Kodachrome, any of several MicroNikkors and, yes, a Nikon SLR. I wasn't completely satisfied with the results. If I framed the bloom(s) tightly to get good detail I lost the setting. If I framed more loosely to show the bloom(s) in setting I lost good detail in the blooms. Can't have everything with 35 mm, its just too small.

So I moved to 2x3, ISO 100E6, still used flash and any of several decent LF macro lenses. My little 2x3 Graphics are much slower working than the Nikon and for many reasons getting good results was harder.

But and however when I did the same shot (same subject, same framing) with 35 mm and 2x3 magnification was higher with 2x3. Who'd have thought it? The result was better detail with 2x3 even though my ISO 100 E6 film had lower resolution than ISO 25 Kodachrome.

When I did the same shot (same subject, same framing, same aperture set) I got less depth-of-field with 2x3. Who'd have thought it? Of course, since DoF is uniquely determined by magnification and aperture set this won't surprise any of you, will it.

Having done all this, I think the OP's question is, if not silly, badly posed.

 

[tomfrh]

Having done all this, I think the OP's question is, if not silly, badly posed.

I suspect most of us have "done all that" and seen that the same aperture on a bigger format gives less depth of field. Shallow depth of field is of course one of the reasons people go for medium format. F2 on medium format is real shallow.

My original question concerns EQUIVALENT aperture and focal length. Ie when depth of field is the same, not different.

 

[tomfrh]

That's where my knowledge stops as I'm not a lens designer but it's pretty darn apparent. Looking at the P67 Flickr group, even with all the other variables, it's a distinct look that the 105/2.4 provides.

Yes I hear a lot about the 105 2.4 and it does have a certain look.

I've often wondered if you could make a 50mm f1.2 which looked the same, or is there genuinely a size effect at work...

 

[Dan Fromm]

I suspect most of us have "done all that" and seen that the same aperture on a bigger format gives less depth of field. Shallow depth of field is of course one of the reasons people go for medium format. F2 on medium format is real shallow.

My original question concerns EQUIVALENT aperture and focal length. Ie when depth of field is the same, not different.

In that case, the larger negative contains more information than the smaller. They're not the same image.

I didn't move up in format to make depth of field narrower. At the magnifications I routinely shot depth of field was smaller than I wanted and there was no way to increase it. Unfortunate fact of life.

I moved up to 2x3 to capture a larger area (subject side, not film side) without losing fine detail.

I didn't go up to 2x3 to shoot at ~ 2.28:1 what I'd shot at 1:1 on 35 mm. I went to 2x3 to get more in the frame at 1:1. Clear? I didn't give up DoF when I shot at the same aperture (flash rig, y'know) and magification with 2x3 as with 35 mm.

The law is: depth-of-field is uniquely determined by magnification and aperture set. Ignorance of the law is no excuse.

 

[tomfrh]

n that case, the larger negative contains more information than the smaller. They're not the same image

Obviously a large format neg contains more information.

My question is what difference is there *aside* from more information.

People claim to be able to see a difference between large format and 35mm even at small enlargements and web images, ie well prior to exhausting the information in a 35mm image. What is it they're seeing?

The law is: depth-of-field is uniquely determined by magnification and aperture

I know what depth of field is. Im talking about different formats with EQUIVALENT depth of field, not different formats using the same aperture, which will of course produce a different depth of field.

 

[Dan Fromm]

For the third time, depth-of-field is uniquely determined by magnification and aperture. Focal length and format have nothing to do with it.

 

[tomfrh]

For the third time, depth-of-field is uniquely determined by magnification and aperture. Focal length and format have nothing to do with it.

What specific statement of mine are you objecting to?

I know aperture and magnification (and hence format size) determine depth of field and am not sure why you think I'm disgreeing with you?