nzeeman
Well-known
i dont know exactly how to formulate what i want to know-but i noticed for example 7artisans 35mm f1.2 sonnar for aps-c. does anyone know why it is easier to make fast wide sonnar for aps-c - is it only because corners would go wild if anyone tried to do it on full frame? or there are some more things to consider-maybe shorter distance to sensor allow for other tyoes of designs that werent possible before. i see that number of elements and groups are also lower in those lenses...
retinax
Well-known
False premise - a 35 for aps-c is not a wide, it's a normal.
nzeeman
Well-known
well i meant it is still wide lens-its 35mm focal length-focal length doesnt change if we change the size of sensor... same as if we use normal full frame lens -it still stays normal lens-just its image is narrow as tele..
but that wasnt a question-question is why now start to appear so many fast 35mm and wider lens with sonnar and other designs that were never before used for wider than 40 and even than they were f2.8...
but that wasnt a question-question is why now start to appear so many fast 35mm and wider lens with sonnar and other designs that were never before used for wider than 40 and even than they were f2.8...
retinax
Well-known
No, to reiterate: a 35 for aps-c is not a wide. Maybe stop thinking in terms of focal length, think in terms of angle of view. A 35 that's made for Aps-c could be the same design as a 50 for full frame, downscaled. It doesn't need to cover full frame, therefore it doesn't need to be wide-angle, unlike a lens of the same focal length that is for use on full frame or larger. And there you have the answer to why these 35 mm lenses for aps-c can be sonnar designs, very fast etc. They do the same thing a 50 for full frame does, in terms of angle of view, and projected image, and can thus be of the same design.
Jamie Pillers
Skeptic
This is an interesting question. I’m interested too, to see if one of our resident optics experts has thoughts about this.
nzeeman
Well-known
so you say that we could replicate any full frame design? and for example make super fast 35mm tessar? im not sure because we would already see all that in movie industry that already use half frame in 35mm movies, and also we would see it in many half frame cameras before...
and one edit to.make it clear- if lens is 35mm yes field of view will be normal on aps-c but all problems that wide angle have will stay-like distortions and all - you wont get portrait lens realistic perspective with 35mm if you use it on 4/3 and has FOV of 70mm you will basically have just middle of wide image kn your sensor .
and one edit to.make it clear- if lens is 35mm yes field of view will be normal on aps-c but all problems that wide angle have will stay-like distortions and all - you wont get portrait lens realistic perspective with 35mm if you use it on 4/3 and has FOV of 70mm you will basically have just middle of wide image kn your sensor .
nzeeman
Well-known
i mean i think it has more to do with distance from sensor because in p&s cameras we saw some pretty amazing low number of elements 35mm lenses and still pretty free of distortion across whole full frame...
ka7197
Established
Your ideas about the relationships of focal length, angle of view, and perspective are fundamentally wrong. That's why you fail to understand the replies to your question.
retinax
Well-known
Non-retrofocus wide angles can be quite good with few elements if slow, and yes they need to be close to the film, so don't work on SLRs, nor do they work well on any digital cameras because the light hits the sensor at too low angles (except the Leicas that are equipped with special micro-lens arrays in front of the sensor). Wide angles for mirrorless digital cameras are retrofocus, just not by as much as SLR lenses.
You mean scaled down? Why not? Of course at the extremes there will be limitations in manufacturing, and some corrections have to be re-calculated.
Why? Tessars are not very suited to high speed.
What is it exactly that you think we don't see? Fast cine lenses?
This is false, if I understand you correctly. Just try it empirically if you can't wrap your head around this in theory.
nzeeman
Well-known
i know it very well-perspective of for example 25 lens is distorted and that is why people used short tele for portraits because it keeps facial features closest to real ones.. if they could use 50mm and just crop it - they would surely do it without need to buy 70mm.. but face looks narrower on 50mm not to mention 35 and wider focal lenghts.. focal length is not same as field of view-you seem to mix those two.. even if you put 50mm lens on point and shoot camera you will keep all 50mm characteristics just will have cropped small portion of it...
retinax
Well-known
You're wrong. For instance a 35 on APS-C and a 50 on FF give the same image, if we leave depth of field out of it. Exactly because "focal length is not same as field of view". Please do try it empirically. Easy to do if you have two lenses of different FLs or a zoom, and can make a crop. Or try this simulator: https://imaging.nikon.com/lineup/lens/simulator/
nzeeman
Well-known
check this link please and scroll down where guy compare same fov of 300mm and 24mm - you will see how compressed everything is on 300mm
retinax
Well-known
Looks about right, that's what you get when you change your position between the shots so the nearest can is the same size in the pictures.
If the camera was in the same position for the 24mm shot, you could crop a tiny frame from the 24mm shot and it would look like the 300mm shot, if resolution permitted.
Edit: It's of course not the same field of view. He compares the different fields of view, yes?
Pherdinand
the snow must go on
Sorry nzeeman but you are still totally wrong, it is as retinax explained and nothing more.
As to the original question- There is more to design besides distortion and resolution.
Light fallff and color aberrations also increase dramaticallt towards the image corners. If you need a lens to cover "full frame" you have more limited design choice as you suspect "corners would go wild" but in many different ways. A lot of lens elements are added in fast lens designs just to maake sure that corner performance is acceptabe too.
As to the original question- There is more to design besides distortion and resolution.
Light fallff and color aberrations also increase dramaticallt towards the image corners. If you need a lens to cover "full frame" you have more limited design choice as you suspect "corners would go wild" but in many different ways. A lot of lens elements are added in fast lens designs just to maake sure that corner performance is acceptabe too.
Larry H-L
Well-known
The somewhat compressed look of “good” portraits has to do with the distance from camera to subject, not the focal length of the lens.
To fill the frame, a short lens must be very close to the subject, and that makes the subject's ears proportionately much further away than the nose. Results in a weird portrait.
With a longer lens, to fit the face in the frame, the distance to the subject must be considerably greater, and proportionately the ears and nose are about the same distance from the camera, and that yields a “better” portrait.
To fill the frame, a short lens must be very close to the subject, and that makes the subject's ears proportionately much further away than the nose. Results in a weird portrait.
With a longer lens, to fit the face in the frame, the distance to the subject must be considerably greater, and proportionately the ears and nose are about the same distance from the camera, and that yields a “better” portrait.
d__b
Well-known
Exactly. Distance is the only factor.
If you photograph two trees which are 100m apart and your camera is 100m away from the first tree, the first tree will be twice the size of the second (100m away vs 200m).
If you move much closer to the first tree to let's say 10m then the first tree will be more than a 11 times larger than the second tree (10m vs 110m).
And that's completely independent from what camera/lens combo you're using. Lens and camera just determine the framing, not the relationship between foreground and background.
Also if the focal length would make such a difference than pictures form smart phone cameras (with their 3mm or thereabout lens) would look extremely distorted and those from 8x10 large format cameras (with a 300mm normal lens) would look super flat. But they don't. They both look normal.
If you photograph two trees which are 100m apart and your camera is 100m away from the first tree, the first tree will be twice the size of the second (100m away vs 200m).
If you move much closer to the first tree to let's say 10m then the first tree will be more than a 11 times larger than the second tree (10m vs 110m).
And that's completely independent from what camera/lens combo you're using. Lens and camera just determine the framing, not the relationship between foreground and background.
Also if the focal length would make such a difference than pictures form smart phone cameras (with their 3mm or thereabout lens) would look extremely distorted and those from 8x10 large format cameras (with a 300mm normal lens) would look super flat. But they don't. They both look normal.
Last edited:
drew.saunders
Well-known
Every lens design has a maximum aperture, and an angle of coverage. It's rare to get the angle of coverage information from manufacturers except for large format lenses. In LF, the manufacturer doesn't know what film size you're using, so needs to provide the angle of coverage, which will determine the image circle, so you'll know if a lens covers just 4x5" or 5x7" or 8x10" or larger.
The Tessar first came out with a maximum aperture of f/6.3 and an angle of coverage around 50 degrees (I don't have the exact information) and, over time, was re-worked to get f/2.7 and 60-64 degrees. There are no f/2.0 Tessar's, it just can't be done, you'll need a different design (such as the Sonnar) to get there.
Given an angle of coverage and focal length, you can calculate the image circle. See post #4 here: https://www.largeformatphotography....-to-calculate&highlight=biogon+angle+coverage
For a LOT of good information.
For 24x36mm film or digital, you need an image circle that's at least equal to the diagonal (43.3mm) and preferably a bit more, so 44mm or larger is best. For APS-C, which is not quite 16mm by not quite 24mm, you need an image circle of about 29mm.
I don't know the angle of coverage of the Sonnar and couldn't find that information, so let's take a theoretical lens design with a 50 degree angle of coverage. With a 50mm focal length, that lens would have an image circle of 46.6mm, and would cover 24x36 and APS-C. With a 35mm focal length, that lens design would have an image circle of 32.6mm, and would not work on 24x36mm, but would work on APS-C.
A lens design with a 64 degree angle of coverage and a focal length of 35mm would produce a 43.7mm image circle, just enough for 24x36 (and more than enough for APS-C). Later Tessar lenses (such as the 35mm f/2.8) just barely covered 24x36mm, so likely had an angle of coverage around 64-65 degrees, but that information is hard to find. I have several large format Tessar design lenses (one Zeiss Tessar, two Fujinon Fujinars, one Schneider Xenar and one Nikkor-M), and most are around 60 degree in coverage, or a bit less.
That's a rather long winded way of saying that a 35mm focal length Sonnar almost certainly doesn't cover the 24x36mm format.
Does that make everything clear as mud?
The Tessar first came out with a maximum aperture of f/6.3 and an angle of coverage around 50 degrees (I don't have the exact information) and, over time, was re-worked to get f/2.7 and 60-64 degrees. There are no f/2.0 Tessar's, it just can't be done, you'll need a different design (such as the Sonnar) to get there.
Given an angle of coverage and focal length, you can calculate the image circle. See post #4 here: https://www.largeformatphotography....-to-calculate&highlight=biogon+angle+coverage
For a LOT of good information.
For 24x36mm film or digital, you need an image circle that's at least equal to the diagonal (43.3mm) and preferably a bit more, so 44mm or larger is best. For APS-C, which is not quite 16mm by not quite 24mm, you need an image circle of about 29mm.
I don't know the angle of coverage of the Sonnar and couldn't find that information, so let's take a theoretical lens design with a 50 degree angle of coverage. With a 50mm focal length, that lens would have an image circle of 46.6mm, and would cover 24x36 and APS-C. With a 35mm focal length, that lens design would have an image circle of 32.6mm, and would not work on 24x36mm, but would work on APS-C.
A lens design with a 64 degree angle of coverage and a focal length of 35mm would produce a 43.7mm image circle, just enough for 24x36 (and more than enough for APS-C). Later Tessar lenses (such as the 35mm f/2.8) just barely covered 24x36mm, so likely had an angle of coverage around 64-65 degrees, but that information is hard to find. I have several large format Tessar design lenses (one Zeiss Tessar, two Fujinon Fujinars, one Schneider Xenar and one Nikkor-M), and most are around 60 degree in coverage, or a bit less.
That's a rather long winded way of saying that a 35mm focal length Sonnar almost certainly doesn't cover the 24x36mm format.
Does that make everything clear as mud?
nzeeman
Well-known
thanks!!! that's actually very useful!! cleared many things that confused me! and thanks to everyone else for helping me understand stuff here... sorry i answer a bit later but was trapped with a baby and baby stuff-you know how kids eat time...
willie_901
Mentor
Perspective depends on one thing - the subject to lens distance.
"why it is easier to make fast wide sonnar for aps-c" ?
The surface area is lower for APS-C. This reduces the manufacturing costs for materials, lens element size and quality assurance.
However, a 62-62 degree angle of view with APS-C requires a 24 mm focal length. Now some (or most) of the manufacturing advantages are reduced. This accounts for the size and cost of FUJIFILM's best XF lenses The Ultra-Fast line) in the range of 14-23 mm.
There is no free lunch.
The f2.8 FUJIFILM XF 16 and 23 mm (and longer focal length) Compact Line lenses require on-board distortion correction correction parameters. This reduces lens size and manufacturing costs.
APS-C offers no inherent advantage in terms of lens manufacturing costs.
"why it is easier to make fast wide sonnar for aps-c" ?
The surface area is lower for APS-C. This reduces the manufacturing costs for materials, lens element size and quality assurance.
However, a 62-62 degree angle of view with APS-C requires a 24 mm focal length. Now some (or most) of the manufacturing advantages are reduced. This accounts for the size and cost of FUJIFILM's best XF lenses The Ultra-Fast line) in the range of 14-23 mm.
There is no free lunch.
The f2.8 FUJIFILM XF 16 and 23 mm (and longer focal length) Compact Line lenses require on-board distortion correction correction parameters. This reduces lens size and manufacturing costs.
APS-C offers no inherent advantage in terms of lens manufacturing costs.
phofseth
Established
Lots of additional confusion might stem from naming conventions. Zeiss has several well regarded and established trademarks. Example: Tessar and Sonnar, so they keep these names for different constructions even if the innards are totally different from the original patents.
p,
p,