Reading mode: Light Dark

Control of depth of field with the D60

Go to last post
Eamon Hickey said:
For any given scene, there are two ways to achieve the same
composition as you would get with a 35mm film camera when using a
smaller format D-SLR. One way is to keep the same lens on the
camera and move backwards, thus increasing your focus distance and
thereby increasing your depth of field.
Click to expand...
It would not be the same composition because when you moved the camera you will change the perspective. Perspective is controlled by the relative Camera to Subject and Background differences.
peripheralfocus said:
The other way is to stay in the same place but switch to a shorter
focal length lens, thereby increasing your depth of field.

If you stay in the same place and use the same lens, the depth of
field will remain the same, as you pointed out, but your picture
will be different -- i.e. your head and shoulders portrait will
turn into a close-up of the face.
Click to expand...
No, the DoF will also be different because of the difference in "format" size. For the same lens in a smaller film/sensor the "Circle of confusion" will be smaller by the Difference in format ratio. This is consitent with your later statement about Medium format, just going in the other direction.
peripheralfocus said:
So, again, for the same composition a D-SLR with a 1.5x FOV crop
will yield about one f-stop's worth of greater depth of field than
you'd get if you shot the scene with a 35mm camera.
Click to expand...
But this is roughly correct. If you reduce the Focal Length by the "Multiplier Factor" and make the F-number smaller by the Multiplier (1-Fstop is 1.4X or pretty close to 1.5X) you will get very close to the same composition and DoF. (But then you will either have to change the ISO or Shutter speed by one stop for the same exposure).
peripheralfocus said:
principle holds if you move up to, say, a 6x4.5cm medium format
camera where, for the same composition , you'll get about one
f-stop's worth less depth of field than you would with a 35mm
camera.
Click to expand...
 
I'm tired of correcting all the misconceptions about this floating around. Ignore what everyone else says. Read what Karl says. He's been right in all the previous incarnations of this debate. I'd also suggest to Karl that when does his tests, that he calculates approximately how much the increase in DOF with a 1.6x (and 1.5x, and 1.3x) crop corresponds to increase in DOF due to change in f/stop. That way you can answer questions like this with, "The 1.6x multiplier means for blurring backgrounds, your f/2.8 lens is only as effective as a f/X.X lens."
ed gerrard said:
I am struggling with the purchase of a new digital camera. My
concern is whether the D60 (or the D100), with their variety of
lenses, will have the control over depth of field that I would
like. With an FOV of 1.6 (1.5 for the Nikon), will I be able to
blur the background of a portrait? If so, how long a lens will I
need? Does anyone have any experience with the D60?

Thanks, Ed
Click to expand...
 
The guy who made the original post had a practical question about real world photography, and this thread has drifted off into optical arcana and, now, semantics, which doesn't help the original poster make pictures. That said, if you want to get into semantics and optical arcana, I'm fully equipped to go as far as you want.
Karlg said:
Eamon Hickey said:
For any given scene, there are two ways to achieve the same
composition as you would get with a 35mm film camera when using a
smaller format D-SLR. One way is to keep the same lens on the
camera and move backwards, thus increasing your focus distance and
thereby increasing your depth of field.
Click to expand...
It would not be the same composition because when you moved the
camera you will change the perspective. Perspective is controlled
by the relative Camera to Subject and Background differences.
Click to expand...
This is just a nitpick about the meaning of the word "composition". I used the term narrowly to mean that which is included in the field of view. I did so to keep the discussion simple, conducted in plain language, and relevant to the original poster's concerns. The original poster was concerned with DOF, so I made a conscious effort to leave out other effects like perspective, but please note the next to last sentence of my original answer to that poster, which says this:

"So choose your focal lengths based on their suitability for people pictures -- i.e. because of the subject distances and perspective effects they will yield with typical portrait compositions."
Karl Guttag said:
Karl Guttag said:
The other way is to stay in the same place but switch to a shorter
focal length lens, thereby increasing your depth of field.

If you stay in the same place and use the same lens, the depth of
field will remain the same, as you pointed out, but your picture
will be different -- i.e. your head and shoulders portrait will
turn into a close-up of the face.
Click to expand...
No, the DoF will also be different because of the difference in
"format" size. For the same lens in a smaller film/sensor the
"Circle of confusion" will be smaller by the Difference in format
ratio. This is consitent with your later statement about Medium
format, just going in the other direction.
Click to expand...
Your first sentence above should be "Yes, and the DoF will also be different because ..." But, in any case, you're introducing additional complicating elements which I purposely left out to keep the discussion plain and narrowly focused. I could point out that, in your above statement, you are leaving out the figures for standard viewing distance for a given print size and whatever your assumption is about the visual acuity of a standard observer. Both figures are assumed in computing a permissable circle of confusion size, and neither is closed to debate. But if I said that, I would be nitpicking for the sake of nitpicking, rather than trying to have a useful discussion in good faith.
Karl Guttag said:
Karl Guttag said:
So, again, for the same composition a D-SLR with a 1.5x FOV crop
will yield about one f-stop's worth of greater depth of field than
you'd get if you shot the scene with a 35mm camera.
Click to expand...
But this is roughly correct.
Click to expand...
No, it's exactly correct. I said "about one f-stop's worth of greater depth of field." By any reasonable definition of the word "about" my statement is exactly correct. That nitpicky enough for ya'?

I try real hard not to get involved in stupid back and forth semantic arguments like this, because they don't help anyone make better pictures, but, congratulations, you got me to do what I hate doing. Now I'm done with this.
 
In the questions that follow, assume:
  • the same camera to subject distance
  • the same 35mm film camera
  • three lenses, 24mm, 50mm, 200mm
I think we can all agree that a cropped 35mm negative must be enlarged more than an uncropped negative to produce the same size print. Does more enlargement decrease DOF?

Someone posted recently that 24mm, 50mm, and 200mm lenses all have the same DOF if they are all cropped to the same FOV and then all enlarged to the same print size. In other words, he was claiming that because the 24mm cropped image must be enlarged significantly more than the 200mm uncropped image, the DOF of the 24mm image would shrink to match the 200mm image. Comments?
  • Christopher
 
When I said this:
peripheralfocus said:
Your first sentence above should be "Yes, and the DoF will also be
different because ..."
Click to expand...
in response to this:
peripheralfocus said:
peripheralfocus said:
No, the DoF will also be different because of the difference in
"format" size. For the same lens in a smaller film/sensor the
"Circle of confusion" will be smaller by the Difference in format
ratio. This is consitent with your later statement about Medium
format, just going in the other direction.
Click to expand...
Click to expand...
... I misunderstood the point you were making. I read that part of your post too fast, for which I apologize. What you are really saying is that because the smaller format image must be enlarged more to achieve any given print size, the permissable circle of confusion must be smaller and therefore DoF tables would be different in the scenario I outlined. (This isn't technically what you said -- technically, you said that the circle of confusion size will be smaller, which is false -- but I know what you meant.)

Again, I tried to leave out the complications of enlargement ratio, viewing distance, visual acuity etc. in my original statement. I was responding to the statement that lens behavior is the same regardless of format. This is true. The actual circles of confusion formed by the lens (as distinct from "permissable circle of confusion for perceived sharpness at any given print size and viewing distance") are the same size for any lens used on any format, if image magnification and effective aperture are constant (leaving out the question of lens quality.) I was talking about lens behavior because that was the specific question at issue; you are talking about enlargement behavior, which is more important in practical photographic terms (though both ways of talking about DoF have their uses.)
 
Christopher Glaeser said:
In the questions that follow, assume:
  • the same camera to subject distance
  • the same 35mm film camera
  • three lenses, 24mm, 50mm, 200mm
I think we can all agree that a cropped 35mm negative must be
enlarged more than an uncropped negative to produce the same size
print. Does more enlargement decrease DOF?
Click to expand...
Yes by the definition of DoF. DoF is defined by how big a point source of light will be in the final image. With perfectly sharp focus the point source will be point on the film/sensor. As you move away from perfect focus (forward or in back of the focus point) the point source will form a circle. That "circle" is the light "blurring." The human visual system only has so much resolution, and if that circle is small enough, it is "acceptably sharp" and considered to be withing the Depth of Field.

When you enlarge something that has be "captured" on a 2D surface (the sensor/film), not only are you enlarging the image, you are enlarging the blur (due to DoF, Lens quality, camera motion, and subject motion). If you enlarge it large enough, you will be able to see blur you could not see on a smaller image.
Christopher Glaeser said:
Someone posted recently that 24mm, 50mm, and 200mm lenses all have
the same DOF if they are all cropped to the same FOV and then all
enlarged to the same print size. In other words, he was claiming
that because the 24mm cropped image must be enlarged significantly
more than the 200mm uncropped image, the DOF of the 24mm image
would shrink to match the 200mm image. Comments?
Click to expand...
On first thought I would say that this is wrong, but it may be that you got the "conditions" mixed up a bit. DoF tends to go down by the Focal Length squared (except when very close or very far away). When enlarging the DoF goes down Linearly with the enlargement for the same viewing distance. Thus the net would be that the longer focal lengh images would in this case have a shorter DoF.

A similar question is what happens if one frames the Subject the same with different focal length lenses? I know that it is approximately true (is falls apart in macro or very distant shots) that if the subject at the focus point fills the frame the same, DoF will be approximately the same assuming the same aperture.

Karl
Christopher Glaeser said:
  • Christopher
Click to expand...
 
Eamon, I appoligies if you feel I nit picked you.

The simple and short answer is that as you said, to stand in the same place and use a 1.6X shorter focal length, then open up the Aperture 1-Stop (1.4X smaller F-number) if the lens will do it. This will be "close enough" for all but the most critical of people.

The problem is too many people do not believe this answer. A large number of people including Phil Askey (see: http://www.dpreview.com/learn/Glossary/Optical/Depth_of_field_01.htm

under the DoF calculator) apparently, seem to think that the DoF will be that of the shorter focal length lens, ignoring the fact that the image will have to be enlarged more for the same size output and this enlargement requires a smaller CoC (by the enlargement amount). It is like the DoF equation did not have a CoC.

As you seem aware, the classic DoF equation (for a "simple lens") has 4 variables. Focal Length, Aperture Ratio = F-number, Subject/focus distance, and Circle of Confusion (simplist to think of as a sharpness metric). The First three terms are all Physics and the laws of physics are not changed by using a smaller sensor. But then we have that peskey CoC term. People that don't understand it seem to want to just ignore it. Everything I have read on the subject says that the CoC for the same size output should be reduced by the amount that the sensor/film format is reduced.

Karl
Eamon Hickey said:
peripheralfocus said:
Your first sentence above should be "Yes, and the DoF will also be
different because ..."
Click to expand...
in response to this:
peripheralfocus said:
peripheralfocus said:
No, the DoF will also be different because of the difference in
"format" size. For the same lens in a smaller film/sensor the
"Circle of confusion" will be smaller by the Difference in format
ratio. This is consitent with your later statement about Medium
format, just going in the other direction.
Click to expand...
Click to expand...
... I misunderstood the point you were making. I read that part of
your post too fast, for which I apologize. What you are really
saying is that because the smaller format image must be enlarged
more to achieve any given print size, the permissable circle of
confusion must be smaller and therefore DoF tables would be
different in the scenario I outlined. (This isn't technically what
you said -- technically, you said that the circle of confusion size
will be smaller, which is false -- but I know what you meant.)

Again, I tried to leave out the complications of enlargement ratio,
viewing distance, visual acuity etc. in my original statement. I
was responding to the statement that lens behavior is the same
regardless of format. This is true. The actual circles of
confusion formed by the lens (as distinct from "permissable circle
of confusion for perceived sharpness at any given print size and
viewing distance") are the same size for any lens used on any
format, if image magnification and effective aperture are constant
(leaving out the question of lens quality.) I was talking about
lens behavior because that was the specific question at issue; you
are talking about enlargement behavior, which is more important in
practical photographic terms (though both ways of talking about DoF
have their uses.)
Click to expand...
 
Karl Guttag said:
On first thought I would say that this is wrong, but it may be that
you got the "conditions" mixed up a bit.
Click to expand...
The original thread is at http://www.dpreview.com/forums/read.asp?forum=1019&message=2267618 and the relevant sentence is "a 20mm lens at f2.8 focused at 6 feet has the same DOF as a 300mm lens at f2.8 focused at 6 feet (if such a lens could focus that close)." Of course, he is making this assertion within the context of cropping to the same FOV and then enlarging to same print size.

If I understand the two points of view, he is saying that the change in enlargement is inversely proportional to the change in focal length (and therefore it all works out in the wash), and you are saying they are not inversely proportional because enlargement affects DOF linearly while focal length affects DOF non-linearly (squared).
  • Christopher
 
Christopher Glaeser said:
Christopher Glaeser said:
On first thought I would say that this is wrong, but it may be that
you got the "conditions" mixed up a bit.
Click to expand...
The original thread is at
http://www.dpreview.com/forums/read.asp?forum=1019&message=2267618
and the relevant sentence is "a 20mm lens at f2.8 focused at 6 feet
has the same DOF as a 300mm lens at f2.8 focused at 6 feet (if such
a lens could focus that close)." Of course, he is making this
assertion within the context of cropping to the same FOV and then
enlarging to same print size.

If I understand the two points of view, he is saying that the
change in enlargement is inversely proportional to the change in
focal length (and therefore it all works out in the wash), and you
are saying they are not inversely proportional because enlargement
affects DOF linearly while focal length affects DOF non-linearly
(squared).
  • Christopher
Click to expand...
http://www.photo.net/learn/optics/lensFAQ

f focal length
So distance from front principal point to object (subject)
Sfar distance from front principal point to farthest point in focus
Sclose distance from front principal point to closest point in focus
Si distance from rear principal point to film (image) plane
M magnification
N f-number or f-stop
Ne effective f-number (corrected for bellows factor)
c diameter of largest acceptable circle of confusion,
or the diameter of the circle of confusion
h hyperfocal distance

Ne = N*(1+M)
h = f^2/(N*c)
hr = h/(So - f)
M = f/(So-f)
c = 0.025mm for 35mm film

Q11. What is depth of field?

A. It is convenient to think of a rear depth of field and a front
depth of field. The rear depth of field is the distance from the
object to the farthest point that is sharp and the front depth of
field is the distance from the closest point that is sharp to the
object. Sometimes the term depth of field is used for the combination
of these two, i.e. the distance from the closest point that is sharp
to the farthest point that is sharp.

frontdepth = So - Sclose
frontdepth = Ne*c/(M^2 * (1 + (So-f) h))
frontdepth = Ne*c (M^2 * (1 + (N*c) (f*M)))
frontdepth = So (hr + 1)

reardepth = Sfar - So
reardepth = Ne*c/(M^2 * (1 - (So-f) h))
reardepth = Ne*c (M^2 * (1 - (N*c) (f*M)))
reardepth = So (hr - 1)

----------------------------------------------------

Ok, now that cropping (and film sensor size) is irrelevant, let's compare three cases that would yield identically framed final pictures:

A. 320mm 2.8 at 6m
B. 32mm 2.8 at 6m with 10x cropping
C. 200mm 2.8 at 6m with 1.6x cropping (D60 equivalent)

For all three:

So = 6m
N = 2.8

A.
f = 320mm
c = 0.025mm
h = f^2/(N*c) = 320^2/(2.8*0.025) = 1462856mm
hr = h/(So-f) = 1462856 / (6000 - 320) = 257.5

frontdepth = So/(hr+1) = 23.2mm
reardepth = So/(hr-1) = 23.4mm

B.
f = 32mm
c = 0.0025mm
h = f^2/(N*c) = 32^2/(2.8*0.0025) = 146285mm
hr = h/(So-f) = 146285 / (6000 - 32) = 24.5

frontdepth = So/(hr+1) = 235.3mm
reardepth = So/(hr-1) = 255.3mm

A 10x crop taken with a 32mm lens will have about 10x more DOF than an uncropped pic taken with a 320mm lens.

C.
f = 200mm
c = 0.015625mm
h = f^2/(N*c) = 200^2/(2.8*0.015625) = 914286mm
hr = h/(So-f) = 914286 / (6000 - 200) = 157.6
frontdepth = So/(hr+1) = 37.8mm
reardepth = So/(hr-1) = 38.3mm

About 1.6x more DOF

Trust me when I say Karl is right. =)

----------------------

And for completeness:

hr = f^2/[(So-f) (N*c)]

With camera (D60) with f focal length lens set to aperture N, equivalent DOF could be achieved on camera' (35mm) with f' focal length set to aperture N' = ?

c' = c*1.6
f' = f*1.6
hr' = hr
So' = So
N = 2.8
solve for N'

f^2/[(So-f) (N*c)] = f'^2/[(So-f') (N' c')]
(f^2) (So-f') N' c' = (f'^2) (So-f) N*c
(f^2) (So-1.6*f) N' 1.6*c = (1.6^2)(f^2) (So-f) N*c
(So-1.6*f) N' = 1.6*(So-f) N
N' = 1.6N * (So-f) / (So-1.6*f)

If you're using your D60 with 100mm lens set at f/2.8 to shoot a portrait 2m away, the resulting photo will have the same DOF as a 35mm shot with a 160mm lens set at f/4.6.
 
Many thanks for the detailed reponse. The math is the easy part for me, and this is very helpful. There is one more piece of the puzzle I would like to understand, and it deals with the equation c'=c*1.6.

For a given focal length and aperature, does the sensor size dictate the size of the glass in the lens? In practice, we know that smaller sensors generally means smaller glass (e.g. D30 vs G2), but is that an optics requirement? Before answering, please allow me to ellaborate on my confusion. The D30 uses the same glass as film. Of course, the D30 could use less glass, but when using 35mm glass, does the sensor use all of the glass? In other words, if I put my finger near the edge of the glass, will it affect the image? And if it does affect the image, doesn't that mean that sensor size does not dictate size of glass. And if sensor size does not dictate size of glass, then why does c'=c*1.6 without considering the glass.

Does my question make sense? Can you see what is puzzling me? Thanks in advance for an explanation that helps me through this.
  • Christopher
 
John Kim said:
I'm tired of correcting all the misconceptions about this floating
around. Ignore what everyone else says. Read what Karl says.
He's been right in all the previous incarnations of this debate.
I'd also suggest to Karl that when does his tests, that he
calculates approximately how much the increase in DOF with a 1.6x
(and 1.5x, and 1.3x) crop corresponds to increase in DOF due to
change in f/stop. That way you can answer questions like this
with, "The 1.6x multiplier means for blurring backgrounds, your
f/2.8 lens is only as effective as a f/X.X lens."
Click to expand...
I'll go with that. Karl/John - I gave up answering threads like this, and just go with backing up whoever is giving the correct answer - usually Karl who has much more stamina than me for these things.

Tony C
 
This seems to trip so many people up.

The larger the film....the shallower the depth of field. This is assuming you are framing the picture the same and everything else is equal.

Current SLR digicam sensors are not as tiny as consumer P&S digicam sensors and therefore have less depth of field.

Danny
Eamon Hickey said:
Steve said:
The D30, D60, etc. etc. have DOF behavior the same as their film
counterparts. The FOV crop doesn't have anything to do with this.
There are many examples of photos that demonstrate this in this
forum (several of them from me).
Steve
Click to expand...
You and I are saying the same thing. Please re-read my post.

I said that for the same composition the smaller format camera
(in this case, a DSLR) will have greater depth of field. That is
completely, unambiguously, 100% true. It is also not in
contradiction with your correct statement that the D30/D60 have the
same DOF behavior as their film counterparts.

For any given scene, there are two ways to achieve the same
composition as you would get with a 35mm film camera when using a
smaller format D-SLR. One way is to keep the same lens on the
camera and move backwards, thus increasing your focus distance and
thereby increasing your depth of field.

The other way is to stay in the same place but switch to a shorter
focal length lens, thereby increasing your depth of field.

If you stay in the same place and use the same lens, the depth of
field will remain the same, as you pointed out, but your picture
will be different -- i.e. your head and shoulders portrait will
turn into a close-up of the face.

So, again, for the same composition a D-SLR with a 1.5x FOV crop
will yield about one f-stop's worth of greater depth of field than
you'd get if you shot the scene with a 35mm camera. The same
principle holds if you move up to, say, a 6x4.5cm medium format
camera where, for the same composition , you'll get about one
f-stop's worth less depth of field than you would with a 35mm
camera.

I chose to talk about it this way to the original poster because
scene composition counts more than any of the other factors we're
discussing here. If you want a head and shoulders portrait, you
aren't gonna' change your mind and take a full-face portrait in
order to prove that the DOF behavior of all lenses and camera
formats is actually the same. Instead, you will back up or change
to a shorter lens so you can take the picture you want to take.
Click to expand...
 
The larger the film size, the shallower the depth of field. This was known by photographers of large format vs. 35mm for years. The same holds true for prosumer digital sensors vs. 35mm. The same holds true for proconsumer digital SLR sensors vs. consumer P&S digital sensors.

Danny
Karlg said:
The size of the sensor does have an affect on DoF. The Consumer
P&S sensors are so small that there is very limited DoF control for
typical scenes . The D-SLR sensors are MUCH bigger (about 2.7X in
each direction) that most P&S sensor.

The simple answer is that if you stand in the same place with a
Digital SLR as a film camera with a given Focal Length and use a
lens that is 1.6X less than "A"(in the case of the D30/60) shorter
focal length you will get the same framing as with a film camera
and 1.6X more DoF with the D-SLR. There are TWO factors that
affect the DoF:

1. You are using a 1.6X shorter focal length to frame the subject
the same in the same place. Over a typical protrait range this
increases the DoF by 1.6 SQUARED.

2. You are cropping the image and this affects what is known as the
Circle of confusion (CoC) in the DoF equation. Many people ignore
the FACT that the more you have to enlarge the image presented to
the Film/Sensor the smaller the CoC has to be. The CoC gets
reduced by the Cropping factor (in this case 1.6X).

Factoring 1) and 2) together you get in increase by 1.6X squared
and reduced by 1.6X or a net of 1.6X MORE DoF if you use a 1.6X
cropped image and a 1.6X shorter Focal length with the same
aperture.

DoF over the normal portrait range goes up or down proportional to
the F-number. As somebody said, you can get close to the same DoF
with a D30/D60 by opening up the aperture by 1 F-stop (1.41X).

You can use any DoF Calculator out there that supports changing the
CoC (more of them do now with digital, but you will find a lot of
them out there that have assumed 35mm film due to its history).
You plug in the REAL Focal length (not equivalent), the REAL
F-number, the Real subject distance and adjust the CoC by 1.6X
smaller -- it is that simple.

All the above assumes "optical equivalent scaling." In a "digital
darkroom" you can do scaling that preserves edges and may give a
high apparent DoF.

Karl
ed gerrard said:
I am struggling with the purchase of a new digital camera. My
concern is whether the D60 (or the D100), with their variety of
lenses, will have the control over depth of field that I would
like. With an FOV of 1.6 (1.5 for the Nikon), will I be able to
blur the background of a portrait? If so, how long a lens will I
need? Does anyone have any experience with the D60?

Thanks, Ed
Click to expand...
Click to expand...
 
Christopher Glaeser said:
Many thanks for the detailed reponse. The math is the easy part
for me, and this is very helpful. There is one more piece of the
puzzle I would like to understand, and it deals with the equation
c'=c*1.6.
Click to expand...
c' = c*1.6 because you're taking a smaller captured image (the "crop") and enlarging it to the same size for final viewing. To do that requires the cropped image be enlarged further than the uncropped image. When you do that, the blur circles get enlarged, and the apparent DOF decreases. For the D60 with a 1.6x multiplier (62.5% crop), the circle of confusion c thus needs to be 1.6x smaller.
Christopher Glaeser said:
For a given focal length and aperature, does the sensor size
dictate the size of the glass in the lens?
Click to expand...
Only to the extent that the lens has to be designed so it's image circle is larger than the image sensor.
Christopher Glaeser said:
In practice, we know
that smaller sensors generally means smaller glass (e.g. D30 vs
G2), but is that an optics requirement?
Click to expand...
When you have a smaller sensor, two things happen:

1. If you're trying to produce an image of X with Y field of view at Z distance, the (true) focal length of the lens needed to produce this image is smaller if your image sensor is smaller. In itself, this doesn't gain you anything (except in some cases a shorter lens). The f/ stop is a ratio of the focal length to lens aperture (lens diameter). Since a smaller sensor needs a smaller focal length to capture the same field of view, the required aperture (lens diameter) is smaller for a given f/ stop.

So a 100mm f/2 lens needs to collect light from an aperture 50mm in diameter. If your image sensor is half the size, then you need a 50mm f/2 lens to produce an identically framed image. A 50mm f/2 lens only needs to collect light from an aperture 25mm in diameter. So the required lens to shoot a picture in a given lighting condition is smaller, not only in length but also in diameter.

2. The lens has to project light onto an image circle at least as big as the image sensor. Not a problem with telephotos, but a huge problem with wide angles. If you hold focal length constant and increase sensor size (opposite of cropping), you increase field of view. Increase the sensor size enough and eventually you'll surpass 180 degrees, which would require a lens capable of taking in light that's coming from behind it. So you can see why a large field of view can be challenging for lens design.

Given this constraint and the aperture constraint above, for 35mm the "sweet spot" for builidng a lens is around 50mm. That's why 50mm lenses are so small and cheap. With longer focal lengths, the main constraint is the aperture (a 200mm f/1.8 lens needs to be at least 111mm in diameter to collect enough light to be an f/1.8). With shorter focal lengths, the main constraint is getting a large enough image circle (which is also tied to collecting enough light from the sides to reproduce the entire field of view). That's why lenses both smaller and larger than 50mm are bigger and more expensive.

When you make the image sensor smaller, you reduce the required image circle and the "sweet spot" shrinks as well. The cheapest/smallest lens you can economically produce for your sensor thus becomes smaller. (Don't ask me if this sweet spot scales linearly or otherwise with imager size. I don't know.)
Christopher Glaeser said:
Before answering, please
allow me to ellaborate on my confusion. The D30 uses the same
glass as film. Of course, the D30 could use less glass, but when
using 35mm glass, does the sensor use all of the glass? In other
words, if I put my finger near the edge of the glass, will it
affect the image?
Click to expand...
Depends how near, but for the most part, yes it will affect the image. An object reflects light in all directions. A lens collects the portion of that reflected light that falls on its surface, and focuses it onto the image sensor. If you put your finger near the edge of the glass, you will block part of that light, and the resulting image will be dimmer for it.

The only time it wouldn't matter is if your finger were outside the field of view of the image the sensor is capturing. In that case, we're talking about your finger not being on (or in) the edge of the glass, but slightly in front of the edge of the glass - enough to show up in the picture until you switch to a smaller sensor (and smaller FOV).

Also keep in mind the two constraints I listed above. For telephoto lenses, a D30 most definitely could not use less glass. The main constraint there is aperture, so you need all that glass to produce the desired f/ ratio. For wide angle lenses, the edges of the glass are primarily there to capture light from a greater field of view. With those lenses, a D30 could get away with less glass.
Christopher Glaeser said:
And if it does affect the image, doesn't that
mean that sensor size does not dictate size of glass.
Click to expand...
It does affect the image, and the sensor size does dictate the size of the glass. =)

Where you're getting confused is you're mixing up two different comparisons. When you have a 50mm f/1.8 on a D60, it's not like having the middle portion of a 80mm f/1.8.
  • A 50mm f/1.8 on a D60 is just a crop (smaller field of view) of a 50mm f/1.8 on a EOS3.
  • The middle portion of a 80mm f/1.8 on either camera would be a 80mm f/2.5 or some other f/ ratio.
So you can compare the same 50mm lens on a D30 vs. on a EOS3 taking pictures of different things (since their FOV is different). Or you can compare the same pictures (FOV) on the D30 vs. EOS3 using different lenses. But you cannot compare the same pictures (FOV) on the D30 vs. EOS3 using the same lens, since it's not possible to do that unless you crop the EOS3 picture.
Christopher Glaeser said:
Does my question make sense? Can you see what is puzzling me?
Thanks in advance for an explanation that helps me through this.
Click to expand...
Hope that explained it.
 
Karlg said:
2. You are cropping the image and this affects what is known as the
Circle of confusion (CoC) in the DoF equation. Many people ignore
the FACT that the more you have to enlarge the image presented to
the Film/Sensor the smaller the CoC has to be. The CoC gets
reduced by the Cropping factor (in this case 1.6X).
Click to expand...
Karl's right, but I'll add a few clarifications (on a bit of a tangent):

1) The finite lower limit of the CoC is the physical pixel size on the imager.

2) The quality of the lens (among other parts of the overall system) affects the smallest "achieveable" CoC even in the sharpest parts of the image.

3) CoC is based on viewing a print from a finite distance (so the more you enlarge it the tighter "tolerance" you need on the CoC). However, if you move back from the print you can reduce your CoC tolerances.

These together, I think, explain the bulk of the old "D30 better than film" vs. "D30 worse than film" disagreements. When prints are viewed from a fixed (close by) distance, there is a range of enlargements where film looks better than D30. However, if you view a D30 print from far back (thus reducing the necessary CoC tolerances), it will typically look better than that of film (as in Stephen's huge enlargements). A chain is only as good as it's weakest link: a D30 is CoC (and thus sharpness) limited typically by the pixel size first (with a reasonable quality lens) whereas film will first be limited by the lens or grain-induced sharpness "artifacts" (usually at a smaller CoC size).

The interesting question here is whether, with the D60, we've reached the point where the lens now becomes the "weakest link" for enlargments viewed from a close distance -- is the pixel size "small enough" that we need "better than L" lenses or larger (physical size) imagers before it makes sense to increase resolution further?
 
John Kim seems to have given you good information. Maybe I can make it a bit simplier.

F-number is the Focal length divide by the diameter of the effective aperture. This sets the MINIMUM diameter of the front element (wide angle lenses have to be wider still). Basically this is the light gathering cabability of the lens.

With a smaller sensor you can use a shorter focal length for the same FoV and thus things get smaller roughly Cubicly in terms of volume and weight.

The other aspect is the image circle. While the front element is set by physics (if you want a 200F2.8, it is going to be about 3 inches in diameter or 200/2.8 mm), a lot of the exotic/expensive materials, the number of elements, and other design issues are to generate a larger image circle that is in effect being thrown away. You could thus "cheapen out" the center groups of the lens if you were designing the lens for a smaller image circle.

The next issue is Wide angle. The big problem here is that the back of the lens is too far away from the film/sensor plane. This creates a whole bunch of problems for making a wide angle lens. A wide angle lens could be made much cheaper if the lens was closer to the image plane. That is why the Point and shoots can do a 7mm lens so cheaply. In theory one could (and I think somebody did with an APS body) redo the whole mirror size and location base on the smaller sensor for a newly designed wide angle and then have effectively an extension tube on the front (that the wide angle would reach through) to accept the existing longer lenses so that they would still be able to focus at infinity.

In the end a lot of this comes down to business. Compared to many other industries, the SLR and D-SLR business are pretty small in terms of volume. I think they are playing it a bit conservative so as to not get hung out with a lenses system that only works for 1 generation of cameras.

Karl
Christopher Glaeser said:
Many thanks for the detailed reponse. The math is the easy part
for me, and this is very helpful. There is one more piece of the
puzzle I would like to understand, and it deals with the equation
c'=c*1.6.

For a given focal length and aperature, does the sensor size
dictate the size of the glass in the lens? In practice, we know
that smaller sensors generally means smaller glass (e.g. D30 vs
G2), but is that an optics requirement? Before answering, please
allow me to ellaborate on my confusion. The D30 uses the same
glass as film. Of course, the D30 could use less glass, but when
using 35mm glass, does the sensor use all of the glass? In other
words, if I put my finger near the edge of the glass, will it
affect the image? And if it does affect the image, doesn't that
mean that sensor size does not dictate size of glass. And if
sensor size does not dictate size of glass, then why does c'=c*1.6
without considering the glass.

Does my question make sense? Can you see what is puzzling me?
Thanks in advance for an explanation that helps me through this.
  • Christopher
Click to expand...
 
Dan Foody said:
1) The finite lower limit of the CoC is the physical pixel size on
the imager.
Click to expand...
And the D30 pixel size is almost exactly CoCmax for a 1.6x sensor..
Member said:
3) CoC is based on viewing a print from a finite distance (so the
more you enlarge it the tighter "tolerance" you need on the CoC).
However, if you move back from the print you can reduce your CoC
tolerances.
Click to expand...
Absolutely agree,, but what tends to be the case in today’s world is viewing an image, maybe even parts thereof, on a 19" monitor, at arms length..

Cheers,
Phred
 
You must log in or register to reply here.

Similar threads

N
Depth of Field with Micro 4/3
6 7 8
Replies
149
Views
7K
Ysarex
Ysarex
I
How long to learn camera controls?
2
Replies
24
Views
5K
5354photos
5354photos
J
Has Auto ISO in Manual mode made Shutter Priority AE obsolete?
Replies
7
Views
1K
sherman_levine
sherman_levine
Z
A7Cii 4k video quality full frame vs APSC crop
Replies
17
Views
6K
Zack23
Z
A
Looking for thoughts and advice on some image problems
2
Replies
28
Views
1K
Victor Engel
V

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top