Aperture fixed when 35mm lenses used on small CCD's??

I've often heard it said that a 100mm f/2.8 lens (for example) designed
for a 35mm camera if put on a smaller CCD sensor will become a 160mm
f/2.8 i.e. the 'effective' focal length gets multiplied by some factor
(1.6) in my example and the aperture remains constant. I'm not so
convinced the latter is true.

I make a few observations.

1) A 100mm f/2.8 lens put on a small digital sensor remains a 100mm
f/2.8 lens. The lens remains the same.

2) The smaller CCD sensor means the focal length (as compared to 35mm)
is longer, as everyone agrees.

3) The aperture whilst still f/2.8 is "effectively" larger, as much of
the light is thrown away, missing the sides of the sensor. So the
viewfinder will be darker than if fitted with a f/2.8 lens which filled
the sensor and no more.

At first the digital format would seem to allow long focal, fast
telephotos. i.e. my 70-200 f/2.8 would become a 112-320 f/2.8, which
would be a very nice fast lens indeed. But I'm not so sure the lens
would have the light gathering power of a real f/2.8 lens, but instead
be effectively an f3.5 (I think). I suspect if the focal length is
multipled by 1.6, the apeture will be multipled by sqrt(1.6), although I
might be wrong on the exact calculation.


PS,
Does anyone know if Nikon are developing a full frame (35mm) digital SLR
like Canon and Kodak?? It seems such a move would have a lot of
technical advantages (lower noise) and people with expensive 35mm lenses
would get the full benefit, and not throw much of the light away, which
is what I think would happen now.

 

Yes

That contradicts 1.

The smaller CCD equals cropping. Sometimes this is referred
to as a longer equivalent focal length. But it is still the same
focal length.

The aperture is still f/2.8.

But you are right with the finder being darker. Moreover,
if you want the same amount of pixels, the actual sensors
must be smaller, thus collecting less light, thus you cannot
have as high ISO - making the system less sensitive.


/Roland

 

http://www.lonestardigital.com/multipler.htm i

"Just as "zooming in" is a result of increased real focal length,
cropped area created by the multiplier factor increases the perceived
focal length of the lens."

Note "preceived" , the focal length is not changed but the image is
cropped down.

As for the smaller size of the sensor that can be a good thing in
that it would be only using that part of a 35mm with the "sweet spot"
for it's optics if a lens has problems at the edges.




********************************************************

"The fox knows many things, but
the hedgehog knows one big thing."

Archilochus
675 - 635 B.C.

 

lets put it this way. My Nikon 70-200 on my D70 remains at 2.8 all the way
through. My exposures show this to be true. If your theory is true then the
shutter has to lie or the daylight must intensify as I throw the switch!

 

So if you crop the center out of a 35mm shot, does it get darker?

Really, the F number correlates to what irradiance will hit the image
plane for a given light field entering the lens. Irradiance is the
power density in watts per square meter; varying the number of square
meters you catch doesn't change this at all.

<snip>

 

without any factual basis,
flys in the face of physics and optics...

 

What actually happens is that the *field of view* is smaller, equivalent
to the FOV you'd get with a 160 mm lens on a full-frame camera.

No. The smaller sensor means that the field of view with the same lens
is narrower. This is sometimes *described* as having the same effect as
using a longer focal length on the full-frame camera.

This is nonsense. You should try tracing some ray diagrams through a
lens. A f/2.8 lens collects a certain amount of light from *a single
point* in the subject and focuses all that light at one point in the
image, if the subject is in focus. At the same time, and entirely in
parallel, the lens is also collecting light from all of the other points
in the subject, and transferring it to corresponding points in the
image.

Now, when you reduce the sensor size, some of the points that used to be
in the image no longer are. And in a sense the light that used to go to
those points is "wasted". But for all of the points that remain in the
smaller image, they are just as bright as they always were. The
f/number of the lens does not describe total light throughput, because
that depends on angle as well as f/number. The f/number determines
brightness per point, or per unit area in the image, and that is
unchanged when some portion of the sensor area is removed.

No, you're wrong. It remains a f/2.8 lens. However, there's no free
lunch. To produce the same quality of image from the smaller sensor, a
given size print needs to be enlarged a factor of 1.6X more. That means
that the lens needs to deliver 1.6 times the resolution to the sensor to
get the same quality print. Thus, lenses that are marginal in sharpness
for full-frame use may look just plain unsharp with the smaller sensor,
while very sharp lenses can stand the extra magnification without
strain.

Dave

 

Hi Dave,

AFAIK "magnification factor" is a marketing term rather than an optical
term. As has been already stated, it is a cropping factor, which results in
a reduction of the FOV. If I cut off a tiny square of a 35mm film, I get a
"magnification", in terms of modern digicam marketing ;-) As far as the
f-stop is concerned, this will remain the same. Focal length and focal ratio
are characteristics of the lens - not the focal plane or whatever lies on
that. BTW the f-number is not defined through terms of power denity or
whatever interpretation anyone could come up with. In classical optics, the
f-number=focal length/diaphragm aperture, nothing more, nothing less. Try
http://en.wikipedia.org/wiki/F-stop and http://www.celestron.com/tb-trms.htm
for some terms.

I hope this helps

dimitris

P.S. I remember well that f-number used to be the inverse of F-number, i.e.
f=1/F, I couldn't find any information on this on the web...

 

Yes it is still a 100mm lens - but it's field of view becomes the same as a
160mm. What this means is that the image you get from a 100mm lens on a 1.6x
digital will be pretty much the same as a 160mm lens on a 35mm cam.

uhh... isn't this the opposite of what you said in 1? - the _equivalent_
focal length is longer. the lens is still 100mm. The behaviour of a lens in
terms of field of view is only relevant when considered in conjunction with
the sensor/film in use. a 50mm lens on a digital SLR is a mild telephoto,
on a 35mm cam it becomes a standard lens, yet on a medium format camera it
is a wide angle lens. on a 8x10 large format camera it would be a fisheye or
extreme wide angle, while on a compact digital it is an extreme telephoto.

The light is not thrown away, it is just that you are not using all of the
projected image circle. To have the equivalent of changing aperture, you
would actually have to throw away some of the light that would otherwise
have hit the sensor. Take one of your 35mm negatives or slides and cut it
down to 24x16mm. Did the negative/slide just become darker? of course it
didn't, it is still correctly exposed. This is because the light that hit
what's left of the negative was still F2.8.

Aperture is not effected at all. This is one benefit of digital - you can
have fast telephoto lenses at lower cost. Instead of having to pay for a
300mm F2, you can get a 200mm F2 for the same amount of zoom and light
gathering capability. 200mm F2 is a lot lighter and a lot cheaper to produce
than a 300mm F2. To take this to extremes, the Panasonic Lumix FZ10 has a
lens that is the equivalent of 420mm F2.8. A 420mm F2.8 lens for film would
be very heavy and would cost a fortune, however for the very small sensor in
a compact digicam, it is lightweight and compact.

The aps sized sensors are a compromise between production cost, image
quality and usability. 35mm is an arbitrary size - there is no rule that
says 35mm is "full frame". If 645 cameras were just as common as 35mm I
could just as easily ask when Nikon would bring out a camera with a 6cm x
4.5cm imaging sensor, so that I could use my old 645 lenses without having a
crop factor. It is a fact of life that every format change requires a
change in other equipment. by having APS sized sensors, people who use
their 35mm lenses benefit by having the equivalent of long focal length fast
lenses, but lose out because their 24mm extreme wide angle lens is now only
a 35mm mild wide angle. It is only at the extreme wide angles that the
smaller sensor becomes a big disadvantage, but at the extreme wide angles,
due to the different characteristics of digital and film, it is best to have
a lens designed specifically for digital anyway. Film doesn't care if light
hits it at an angle, digital does.

 

No, the focal length is the same, but the field of view changes, as the CCD
is a smaller format. Think of 645 vs 35mm for good film analogy.

 

correct

I don't agree; the angle of view changes, not the focal length.

You are missing the fact that film (or FF digital) NEEDS more light in
total to get the same exposure, because the same amount of light is
spread out over a much larger area. You have to look at exposure *per
square mm*, which remains the same of course, regardless of format.

They have said numerous times they aren't. It has no advantages. BTW the
bodies with the most noise are the Contax N-digital and the Kodak 14n.
Now what does that tell you?

The benefit would be they can use their old gear, which is an advantage
for users but not for Nikon. Their customers aren't the "brand-loyal"
people who have a dozen old Nikkors and a few bodies, their customers
are people who actually BUY new stuff.

Lourens

 

True, except when using a film lens on a digital camera.

It would be misleading to say that the crop factor of digital capture
devices which simulates a longer lens length doesn't come without a price.

As Dave Martindale pointed out: "However, there's no free lunch. To produce
the same quality of image from the smaller sensor, a given size print needs
to be enlarged a factor of 1.6X more. That means that the lens needs to
deliver 1.6 times the resolution to the sensor to get the same quality
print. Thus, lenses that are marginal in sharpness for full-frame use may
look just plain unsharp with the smaller sensor, while very sharp lenses can
stand the extra magnification without strain."

Film best,
me

 

In fact a small sensor increases DoF. What decreases DoF is the fact
that you need a shorter FL for the same image.

Lourens

 

addendum to my other reaction:
This exact effect in fact *decreases* DoF when a smaller sensor is used.

Lourens

 

No, you've got that reversed. A smaller sensor, while keeping the lens
focal length the same, *decreases* DoF because of the greater
magnification needed when printing.

If you want to keep the field of view the same, you need a shorter FL
lens. The shorter FL *increases* DoF by a factor of the square of the
sensor size ratio.

When both of these effects are taken into account, you get a net greater
DoF from the smaller sensor with the same field of view, by a factor
equal to the reciprocal of the sensor size ratio.

Dave

 

Nope.

This is a whole lot easier to think about if you're familiar with many
different formats of film (or sizes of image sensors). I have two 120mm
lenses, one is a wide angle (for a 4x5" camera) the other is a short
telephoto (for a 6x4.5cm camera). The 120mm f5.6 Schneider has a 179mm
image circle, so any format of film that fits inside a 179mm diameter
circle can use that lens, from APS (which only needs a 35mm image
circle) on up to 4x5" (which needs a 153mm or so minimum image circle).
It's just for really small formats, most of the image is wasted, and so
most folks use lenses that are designed and optimized specifically for
those smaller formats. My 120mm f4 Mamiya lens probably has only an 80mm
or so image circle, as it's designed for a 6x4.5cm camera (56x41.5mm
film size), but they're both the same focal length.

Lenses designed for 35mm film should have an image circle diameter of
45mm or so, they can have more, and tilt/shift lenses must have more,
but there's no real benefit to going much over 45mm. Your dSLR is just
using a smaller part of that 45mm image circle. Just calculate the
diagonal (43.3mm for 24x36mm) and round up a bit to figure the minimum
image circle required for your particular dSLR. The edges of the image
circle are generally not as sharp, so you want a little bit of excess.

This same 120mm lens has a horizontal angle of view of 53 degrees on
4x5" (95.3 x 120.7mm actual image size), 39 degrees on 6x9cm, 32 degrees
on 6x7cm, 26 degrees (horizontal, remember) for 6x6cm and 6x4.5cm, 17
degrees for 24x36mm, and a fairly narrow 14 degrees for 16.7 x 30.2, the
old APS standard which is about what many digital SLR's use.

It's always f5.6 at maximum aperture, no matter how much of that image
circle you use. Really wide angle lenses have light falloff at the
edges, which changed the amount of light that reaches the film at the
edges. This would be the case if there were a 120mm lens designed for
11x14" film, for example.

Less confused or more?

Drew