Disclaimer: I’m not very familiar with optics I might well be wrong about this.
[quote=“CarVac, post:11, topic:422”]
The wavelength of light being fixed is compensated for by the fixed entrance pupil size due to using the 135mm at f/4 and the 85mm at f/2.8.
[/quote]As far as I know the fixed entrance pupil size will mean that they have the same amount of diffraction, due to that pupil. What will still be the case is that the lens needs to bend the light more (the light will come in at more oblique angles), which I would expect to cause more issues with CA and Vignetting. It’s of course much less of an issue when you work at the focal length’s and apertures you stated. It becomes quite relevant. So I would totally question if a Voigtlander Nokton 17.5mm f/0.95 Lens for four thirds performs similar to a 35/f2 on a full frame.
[quote=“patdavid, post:4, topic:422”]
I feel like his main point is simply that the technical arguments around gear do little in the way of advancing a persons eye for composing images. The limiting factor for the majority of folks is not what the technical aspects of the camera are, but rather the 12 inches behind it…
[/quote]That is certainly true.
Oh, definitely true. That’s the real advantage of larger formats: it’s easier to make a perfect lens with a given entrance pupil size if you have a longer focal length because the f-number is larger. It’s probably very easy to make a large format lens crazy sharp at f/8.
What do you mean that “optics don’t work like that”?
This is from personal experience. They look identical, and they should be. The biggest telltale difference is the aperture being hexagonal in my stopped down 135.
Well, there are many things that characterize a specific focal length. The field of view is one, but there are also things like distortion or compression of the background. By multiplying with the crop factor you can get an equivalent field of view, however everything else stays the same. A 50mm on crop will have a field of view like a full frame 80mm, but it stays a 50mm, so you still get big noses when shooting portraits, … Granted, the longer the focal length gets the less that matters as the typical distortion of long lenses is not as pronounced as that of wide lenses.
Big noses in portraits only depends on lens distortion and subject distance. Assuming good lenses, distortion is the same (zero).
Subject distance should depend only on field of view, which is equivalent.
Background compression is also only affected by subject distance and angle of view, thus only by angle of view, which is equal for equivalent lens/sensor combinations.
I was thinking about this some this morning, and the best way to approach the topic in an approachable way.
What I thought about was this:
The focal length of a lens is a function of the lens only (not the camera).
The aperture of a lens is a function of the lens only.
If I take an 85mm lens on a full frame camera, and then take the same shot, same location, focus, etc., but use a crop camera, I should end up with an identical image in every way, but cropped to a smaller portion of the full-frame image. This should be the case, right? (I don’t have the hardware to test this idea at the moment, but I think it should be true - the resulting image being created from a given lens doesn’t care what type of capturing medium is on the other end).
So, same exact lens, crop vs. full frame camera = same image, just cropped to a smaller portion of the resulting image. The optical characteristics of the 85mm are not changed in any way with crop vs. full frame except for a smaller area of capture.
This thought experiment works in the opposite direction as well. Use a lens designed for a crop camera, and shoot it adapted to a FF body - you should get an identical resulting image for the most part.
Does this make sense?
TL;DR: a 50mm is a 50mm regardless of what’s actually capturing the image it creates. You just might end up with a smaller FoV on a crop.
The point is not what gets an identical image, but rather an equivalent one when using the full potential of your equipment.
If you wanted the same field of view as 85 on crop, you wouldn’t crop from an 85mm on the full frame camera. You’d use a 135mm, one stop down, because it’s equivalent but better.
135/4 on the whole 36x24 frame is pretty darn close to equivalent to 85/2.8 on Canon’s APS-C sensor. Same angle of view, same depth of field given a given subject distance, same amount of background blur, same amount of diffraction-“limited” resolution.
However, since usually you’re sensor performance limited, on full-frame you’ll generally have more resolution and dynamic range and so on. And so we talk about using equivalent lenses: 25/1.4 on M43 is equivalent to 50/2.8 on full frame.
In my mind, that’s completely wrong. 50mm is just a number if you don’t know the sensor size; field of view is what really matters. 50mm on a Hasselblad is not 50mm on m43. It’s a completely different use case, and unless desperate, nobody would ever use the wideangle 50mm on a Hasselblad for the purposes that a short telephoto 50mm on M43 is suited.
I’m an advocate for completely ignoring the physical focal length except when identifying a lens model.
My point here was only to demonstrate that using an 85mm lens on a crop sensor does not magically imbue it with any characteristics of a ~130mm lens. It only produces the same field of view you may get. The rendering characteristics of the lens remain the same, and more importantly, this includes any distortions.
Except that I’m reasonably confident that it is not wrong, in the point that the only difference in what you will get with the lens when using it with different sensors is only a crop/fov difference. Everything else will be the same.
Yes, with the sole exception of field-of-view, a 50mm on a hassy is the exact same thing in terms of rendering and distortion as the same lens on a m4/3.*
*It is also very likely that I am wrong, I am only relying on some simple engineering logic to come to these conclusions. I am thinking that if we set up a camera obscura type device, with only a lens and a viewing screen, the size of the viewing crop on the screen will have no impact on the lens characteristic other than to show less (or more) of the resulting image (crop). Right? Distortion and rendering will remain unchanged.
I agree as well! I am only pursuing this topic in hopes of finding a nice way to clarify some of these things for others to better understand.
[edit]
What we need is a way to purchase an FF and crop in a system with the same lens for all of us to loan out as needed so we can test these things. pixls.us hardware lending library! (Oh, and I guess we could shoot some photos with it as well occastionally… ).
There’s always a difference in terminology behind disagreements on the internet.
For me, ‘equivalence’ only refers to the depth of field/background blur/entrance pupil size and the angle of view. Nothing else.
You can’t “set the lens to Otus” or “Dial it out to Lensbaby”, (unless you have one of those old soft focus lenses), but you can zoom and set aperture freely; furthermore most people are always in search of more detail and sharpness so having a ‘lensbaby’ mode isn’t even desirable.
##Distortion
What distortion do you refer to? Geometric distortion (bending of straight lines in imperfect lenses) or perspective distortion (based on how far from the subject you are and where it is in the frame, with either perfect or imperfect lenses)?
Geometric distortion does change with crop sensors. You chop off the corners, and there will be less distortion, almost always (except for weird moustache distortion like on the Zeiss 21/2.8 Distagon).
Perspective distortion: With constant framing, perspective distortion also changes because your angle of view changes and so you have to be at a different distance. Or if you keep constant distance, using the 50mm lens on M43 will crop out any weird perspective distortion effects visible on the sides of the Hasselblad shot.
If there’s any other distortion you are thinking of, please enlighten me.
##Rendering
Not relevant, in my mind, for equivalence, since it varies even between lenses with exactly equal specs (see 35L versus Zeiss 35/1.4 versus Sigma 35 Art). This isn’t a parameter you can just ‘set’ on most lenses.
Furthermore, rendering does change with magnification (sensor size). The more closely you inspect a lens, the more its flaws show up. On crop sensor, my 135 is noticeably soft wide open but if you view its full image circle you magnify the output less and so it looks much sharper.
If you want me to, I wouldn’t mind renting a full-frame camera again and doing experiments with equivalence pairs.
My 85 and 135 (85*1.6 = 136), which render nearly identically.
My 28 and 45 (28*1.6 = 44.8), which could hardly render more differently.
I would test constant distance and constant framing, constant f-number and constant entrance pupil size, and constant focal length and constant angle-of-view, and then I could put the whole thing to rest.
I spent some time thinking about this today and I think we’re close to a middle ground of understanding.
In short, for any given particular lens, it will always produce the same image in it’s imaging circle regardless of what sensor you may put behind it. I don’t think any of us are questioning this.
As such, a 50mm lens from a FF camera will produce the same exact image in it’s imaging circle. Putting a crop sensor there instead of FF will produce an identical image, only cropped to the smaller sensor size.
If your 85 and 135 are both FF lenses, then the 85 on a crop sensor will not produce an identical image to the 135 on a FF. It may produce a similar angle-of-view, but the distortions (both geometric and perspective) will still be there (though they may be cropped in the corners as you mentioned).
What I’m not sure of is if an 85mm lens for a crop body will produce “identical” results to a 135 on a FF. For that I simply lack the optical physics background to understand it better. Is your 85mm an 85mm for a crop sensor, or full frame?
When you magnify the 85mm image by 1.6x, then the image will be identical. The images cast on the sensor are exactly the same except for 1.6x scaling, which happens naturally when you view the image. THAT is what is meant by equivalence between sensor sizes.
And I still don’t understand what you mean by distortions; let’s just assume there is none.
85mm is 85mm whether for crop or for full frame. The image they cast on the sensor is the same scale. 85mm crop sensor lenses just cannot illuminate the whole frame, so you MUST magnify them.
(All of my manual lenses are full frame Contax/Yashica mount)
When I was thinking about this, my equivalence was to keep all the physical parameters of taking the photo exactly the same, only switching out a FF body for a cropped one (same locations, subject distances, etc…). If that’s the case, then using the example image below:
Simply replacing any given image with a crop sensor will not magically make it look like the other (replacing the 35mm image with a crop sensor, and keeping everything else the same - physical locations, will not magically make it look like the 70mm image on a 2X crop body).
Here’s the crux, unless you maintain a constant angle-of-view, as you actually mentioned above (and assuming the simple pinhole camera approximation). That is, swapping out the 35mm FF, with a crop sensor and move the camera back to maintain constant angle-of-view (this is the part where we were crossed up, I think). Then it should be equivalent resulting images.
I think we’re both saying the same thing, I just should have paid closer attention to exactly what you said earlier.
The question now is, can/should we write something up that can explain and illustrate these points in a concise way that is easy for most users to digest?
Are you saying that all of those images are equivalent? If so, I disagree.
You cannot have equivalent images with different angles of view. (Technicality: it might look the same if you are taking a picture of a piece of paper.)
If you move, the relative sizes of foreground and background will change.
If you don’t move, then the subject will change size in the frame.
Thus, the images can’t be equivalent. You must (necessary but not sufficient) match the field of view.
Here’s the crux of equivalence:
The only things that matter for system equivalence are field of view and entrance pupil size. If the field of view is the same and the entrance pupil is the same size, then the images will be equivalent.
85 on a crop camera has the same field of view as my 135 on a full-frame camera.
My 85 at f/2.8 has the same entrance pupil size as my 135 at f/4.
Thus, 85 on a crop camera at f/2.8 is equivalent to 135 on full frame at f/4. They will produce images with the same field of view, and when focused at the same distance they will have the same amount of background blur.
(I guess aspect ratio matters too, but 4:3 to 3:2 is close enough for the purposes of equivalence)
I’m not sure if you’re being argumentative now for it’s own sake…
I am saying that in those examples you see a good example of geometric distortion due to proximity of the lens to subject. In the 35mm example, if you don’t move anything at all, and simply swap out the sensor for a crop sensor (2x), then you don’t magically get a 70mm result. You simply get a cropped view of the same image. That’s all.
Perhaps you’d like to start on an article going over these thoughts?
Why wouldn’t it be the same as instead mounting a 70mm lens in the same location? With the same entrance pupil size, a 2x crop from 35mm or switching to a 70mm lens should be identical.
I would write an article but I don’t understand how you’re approaching it.
Because that’s not what I’m saying. I am saying that keeping all else the same for setup of a shot, simply swapping the FF for a crop sensor will not magically get you the 70mm result (assuming 2x crop). Do you not agree?
Yes, I think this is true and agree with you.
How I approach it shouldn’t matter to an article being written. However, I’ve tried to explain my thought process in an attempt to describe how some folks may view the problem, and the types of misunderstandings that might arise from it (our exchange is a good example of this).
My approach would be to consider the audience, assume a relative level of competency and direct your writing to that. Try to anticipate where problems might occur in understanding the material and have clear information available to address it. Large, visual examples to help illustrate concepts will helps as well.
Why don’t you start a thread and take a first stab at it? We can all pitch in.
I think I see where we are misunderstanding each other.
I think that the “cropping the 35mm result” is not relevant to a discussion of equivalence, because in the first place if you wanted the 70mm result you wouldn’t be standing where you took the 35mm shot.
What would be relevant would be swapping between full frame+70mm and 2x crop+35mm, not just the lens or just the crop. Otherwise you’re comparing systems that aren’t equivalent.
What I’m trying to figure out is why you brought it up. This is because if you would bring it up then someone else may and I want to understand your point of view before embarking on the article.
).