PDA

View Full Version : focal length and FOV using different lenses



Amos Kim
06-25-2011, 02:16 AM
I know this question has been brought up before, ever since the universal interchangeable lens mount/ large single sensor cameras have emerged, but I want to bring it up again because I'm still unsure: Hypothetically, will a Nikon 24mm full frame f-mount lens, Nikon 24mm DX lens, Leica 24mm M lens, Sony 24mm e-mount lens, Canon 24mm EOS lens, 24mm PL mount lens, all adapted via appropriate adapter to FS100 (except native e-mount lens), give same FOV?

I believe yes, but my photographer friend says no.

Appreciate any clarification. thanks.

morgan_moore
06-25-2011, 02:24 AM
yes

same FOV

Some brands mildly misrepresent the focal lenght (esp in zooms) by .5mm or so but 24=24

S

Amos Kim
06-25-2011, 03:01 AM
Thanks for input morgan and I think you're right but why? Anybody else feel free to chime in.

Dustin Moore
06-25-2011, 05:32 AM
Thanks for input morgan and I think you're right but why? Anybody else feel free to chime in.

Ok, if you are asking why all of the lenses should have the same field of view if they
have the same focal length it is because the field of view is basically determined by
focal length and the size of the sensor.
http://mathbin.net/equations/63726_0.png

If you are down with geometry you can think of a lens as being a pinhole that is one focal
length away from the center of the film/sensor. Light coming through the pinhole and striking the edge
of the film/sensor at half the sensor width must of come through the pinhole at the angle
in the equation.

So, ideally, if the lenses are all exactly 24mm then they will have the same field of view.
In practice, though, lenses have a small amount of deviation from the perfect called
distortion. In extreme conditions you see this as barrel or pincushin effect but it
is always present.

Most normal and telephoto lenses have a distortion of less than 4% so you can expect
the field of view to agree within about 4%. The wider lenses tend to have more
distortion so you might expect a lot of variation if you were comparing 12mm lenses
rather than 24.

So that's one answer. If you were asking why do lenses in practice all have small
variations in focal length it is because lens engineering is hard. You may have a
lens design that works perfect at f=24.75mm but if you try to flex it to f=24.00mm
some nasty chromatic abberation appears, coma goes out of control or your lenses
smash into each other internally. You might be able to fix the problem with a more expensive glass or by adding a another lens element but it would be expensive. The f=24.75mm is just easier to build given the tools you have on hand for the price
you can afford.

Another answer is that mechanical tolerances of lenses are sensitive. If the elements
aren't at exactly the correct spacing, the focal length will drift. If you checkout
the lens porn here http://www.fredmiranda.com/forum/topic/953111
you can see that it takes some serious mechanics to hold a high performance lens
right at the desired spacings. Your $120 lens made in china out of molded plastic
can not maintain the same mechanical tolerances as the $30,000 made in germany.

Noel Evans
06-25-2011, 05:37 AM
Dustin, 4 posts so far at DVXUser and all quality, technical, well presented responses. Thank you! More users like you please.

morgan_moore
06-25-2011, 06:26 AM
I think one my to get straight is coverage

Say nikkor DX lens might be 14mm but only cover the S35 chip

A nikon Fullfram lens covers a FF chip

When used on fullframe 14mm is very wide, put the same lens on a S35 chip camera and you only get the FOV that you got with the DX lens which is wide, but not uber wide because the chip is smaller

S

Amos Kim
06-25-2011, 11:56 AM
Interesting morgan, but let's do that backwards, put a 14mm dx nikon lens from an aps-c nikon camera to a full frame nikon camera, same mount, but there is cropping. If all 14mm lens are really essentially the same regardless of camera system, why did the view change?

LoganMackay
06-25-2011, 02:36 PM
Because the dx lens was made to cover a much smaller sensor

Amos Kim
06-25-2011, 03:00 PM
so i'm assuming that the two 14mm lenses are internally different, although they are both "14mm". And if so, how are they physically different? And if they are different doesn't that contradict the fact that they should render the same FOV, and 14mm=14mm?

Just trying to get a clear picture. thx guys.

morgan_moore
06-25-2011, 03:42 PM
Its called image cirlcle

Think about an old 54 camera (5 inches by 4 inches negative)

No DSLR lenses would start to cover that negative they would just project a small circle of light in the middle

the image circle

Now you can get 65mm lenses that cover the 54 negative - 65 is a very wide format on a 54 negative and far less wide on DSLR size chips

If you took a 54 photo and cut the middle 35mm out of it the image would be no different from the 35mmneg/65mm lens image

same with the 14

take the FF35 image and crop the edges of and you get the S35 Field of view - less wide, same focal lenght

S

nyvz
06-25-2011, 06:39 PM
Interesting morgan, but let's do that backwards, put a 14mm dx nikon lens from an aps-c nikon camera to a full frame nikon camera, same mount, but there is cropping. If all 14mm lens are really essentially the same regardless of camera system, why did the view change?

It didnt. There is no cropping, there is vignetting, which is in a sense the opposite of cropping. If you put a 14mm lens on FF35 that covers the full frame and a 14mm lens that is only designed to cover APSC, they will both render the scene as the same size and perspective, its just that one lens will show vignetting so parts of the scene will be obscured by the black shadow cast by internal lens components. This is an effect of the design of the lens more than the focal length. You will have to crop your image to get an image out of the lens that does not have vignetting, but cropping is effectively decreasing your sensor size. Focal length and field of view will be constant. It may be confusing to some because we talk about field of view as being the angle of view seen from one side of the frame to the other, but another way to think about it is the ratio of the angle of view to the image width or height. Even if the lens is not designed to cast an image on an entire FF35 format sensor, if it has the same focal length as another lens that does cover the whole sensor, the two lenses will still have that same degrees of angle of view per mm of project image width or height for the portion of the image that is visible to the sensor from the lens.

Amos Kim
06-25-2011, 07:05 PM
yes, you're right, it's not cropping, it's vignetting. But is this vignetting due to anything inside the lens, thus being "designed for a certain sensor size" as you say, or is it a function of flange back distance or something outside the physical lens? The latter explanation still supporting the 14mm lens=14mm lens=14mm lens theory regardless of camera system or sensor size etc. etc.

Mitch Gross
06-25-2011, 09:21 PM
Perhaps this video I did would be helpful:

http://blog.abelcine.com/2010/05/14/a-lens-is-a-lens-is-a-lens/

Dustin Moore
06-26-2011, 05:52 AM
so i'm assuming that the two 14mm lenses are internally different, although they are both "14mm". And if so, how are they physically different? And if they are different doesn't that contradict the fact that they should render the same FOV, and 14mm=14mm?

Just trying to get a clear picture. thx guys.


The focal length says how much "magnification" the lens provides between the reality and
the sensor (a "zoom" factor maybe). It does not guarantee that light will actually hit the sensor.

If you take a 14mm lens designed for a DX frame camera, it will let enough light through to cover
a sensor that is about 22mm wide. Outside of that width, the lens is going to start to vignette horribly.

If you take a 14mm lens designed for full frame, it will let enough light through to cover a width of 36mm
wide. Outside of that it will vignette horribly. It will also weigh about 30% more, cost 20 to 50% more and have
more parts inside of it.

If you put the DX or APS-C lens on a full frame camera and snap a picture, you will basically see an image
that covers the middle part of the frame but part way out the vignetting roll in quickly until it cuts all of the light
out. You still have the same mathematical field of view but vignetting has gone to 100% well before the edge
of the image.

If you put a full frame lens on a DX body and snap a picture, life is good. Not only do you have the same
field of view as the DX lens but you probably have very little vignetting. The FF lens is just starting to
vingette a tiny bit at the absolute edge of your smaller APS sized sensor. On the downside, you are using
a heavier and more expensive lens than you really need to.

nyvz
06-26-2011, 12:48 PM
so i'm assuming that the two 14mm lenses are internally different, although they are both "14mm". And if so, how are they physically different? And if they are different doesn't that contradict the fact that they should render the same FOV, and 14mm=14mm?

Just trying to get a clear picture. thx guys.

Why would physical differences in similar focal length lenses that don't significantly affect focal length or FOV contradict the focal length FOV correlation? There are infinite ways to build a lens using any number of elements, and even lenses built to be the same have components that are built to certain finite tolerances so there will always be variations in the physical attributes of even lenses that are designed to be the same model from the same factory, but that does not really have any bearing on the simple concept that lenses are designed with a focal length and image circle size in mind among many other characteristics...

Amos Kim
06-26-2011, 01:19 PM
I just don't understand what it physically is inside the lens that makes a dx lens vignette on a full frame camera, while the full frame lens equivalent (for example two 14mm lenses, dx and fx) covers the whole full frame sensor, thus, these lenses being "designed for" certain sensor sizes. That's the physical difference I'm trying to figure out. But I guess that's a question for the optix engineers and really don't have any practical relevance. We just know that a 50mm dx lens is different from 50mm fx lens, but they do render same fov because they are same focal length.

btw, that was a great explanation dustin, thanks.

morgan_moore
06-26-2011, 01:51 PM
Have a think about creating a 14mm lens that would cover a sensor the size of your bedsheet (I joke of course)

It would be a big lens, very big

A ff35 14 covers a FF35 sensor - it s a lot bigger than S35 sensor

So while the angle of light coming out of is is the same as the S35 sensor - its just bigger !

if a lens projects onto your bedsheet you could lay your S35 chip onto that and it would pick up the light too

S

Dustin Moore
06-29-2011, 05:04 AM
I just don't understand what it physically is inside the lens that makes a dx lens vignette on a full frame camera

It is what is not inside the DX lens that makes it vignette. A lens is basically and iris/aperture stop with glass wrapped around it.

If the glass isn't wide enough to allow light to come in at a steep enough angle to hit the sensor, that part of the sensor is vignetted. It's not quite
easy to imagine this without a diagram. Checkout the lower right sketch in this plot:
http://www.sinopt.com/software1/usrguide54/examples/IMAGES/DEMOTRIP.GIF

You can see that the width of the beam that is going at the steep up angle is far less than the width of the on axis beam going
straight right to the image. The off-axis beam is being clipped by the first piece of glass and the last piece of glass. Since the off
axis beam is less wide, it has a higher f number and is dimmer. At a high enough angle, no light gets through at all and it is totally
vignetted.

In order to make this lens cover a large image circle, the first and last elements would need to be wider/taller. This usually involves
making the glass elements larger and also adding new elements to compensate for the additional abberations that creep in
as the image circle gets wider and wider.

Amos Kim
06-29-2011, 11:43 AM
Another great explanation, thanks dustin, it's a little clearer now.
Mitch, thanks for the video post.