swing lens cameras and focussing distance

yes, the field of view of the vertical image is set by the fixed lens
(here, 28mm, or 50mm in noblex etc.), while the horizontal field of view
is determined by the angle of rotation with the slit open (110+ degrees
etc.).

The degree of enlargement, and hence the CoC used, depends on what and how
you enlarge; if you want the full panoramic you shot, then I would expect
the horizontal enlargement capability of your enlarging setup to be more
of a limitation than the vertical capability (i.e., 56mm is bigger than
24mm, and you are more likely to find the length of a 20X enlargement
(46"+) more problematic than the width (20"). Even a 10x enlargement would
be 10" x 23+"; the point being that the horiz. axis is the limiting one.

So I'd expect the CoC to be set by the horizontal enlargement factor,
based on the lens swing angle, rather than the vertical coverage of the
image (set by fixed lens used).

In practice, most of us can't print panoramics big enough so the vertical
axis becomes limiting (the axis controlled by the fixed (28mm) lens), yes?

my $.02 again ;-)

bobm

--
************************************************** *********************
* Robert Monaghan POB 752182 Southern Methodist Univ. Dallas Tx 75275 *
********************Standard Disclaimers Apply*************************

"Bob Monaghan" wrote in message
...

yes, the field of view of the vertical image is set by the fixed lens
(here, 28mm, or 50mm in noblex etc.), while the horizontal field of view
is determined by the angle of rotation with the slit open (110+ degrees
etc.).

The degree of enlargement, and hence the CoC used, depends on what and how
you enlarge; if you want the full panoramic you shot, then I would expect
the horizontal enlargement capability of your enlarging setup to be more
of a limitation than the vertical capability (i.e., 56mm is bigger than
24mm, and you are more likely to find the length of a 20X enlargement
(46"+) more problematic than the width (20"). Even a 10x enlargement would
be 10" x 23+"; the point being that the horiz. axis is the limiting one.

So I'd expect the CoC to be set by the horizontal enlargement factor,
based on the lens swing angle, rather than the vertical coverage of the
image (set by fixed lens used).

In practice, most of us can't print panoramics big enough so the vertical
axis becomes limiting (the axis controlled by the fixed (28mm) lens), yes?

No. You scan it (in sections and stitch if necessary), put roll paper in the
printer, and print as long as you want.

David J. Littleboy
Tokyo, Japan

(Bob Monaghan) wrote in message ...
yes, but once you "hacked" the camera, you could swing a 50mm lens thru
110+ degrees. It would look different from a similar 110+ degree sweep
with a 28mm lens for a variety of reasons. The image magnification would
be different, the image height on film would be different, and the amount
of enlargement needed would be different. So the CoC and the DOF would all
also be different, just as they are when you switch from a 28mm to a 50mm
lens.

Yes, they would both be 110+ degress, and yes, the magnification would
be different and the height on film different. The CoC would still be
the same though as this is to do with the film area diagonal being
divided by a standard number ( diagonal/1440 ?). The DoF would be
different -- much less for the 50mm lens due to its longer focal
length (same as lesser magnification).

However, in both 28mm and 50mm swing lens cases, you might end up with a
110+ degree angle of coverage on a 24x56mm piece of film. The horizontal
coverage is the same. The vertical coverage is obviously different. The
amount of "cigar shape" swing lens distortion is also quite different
(28mm wide is worse).

Not sure the 28mm would be worse since the cigar shape is due to the
lens swinging through an angle.

But the key point is that the you see 110+ degrees of subject horiz. on
film in each case. Therefore, the swing lens camera requires a different
analysis than simply using the fixed lens focal length in a standard DOF
calculator.

This is where I fundamentally disagree. If you cropped a swing lens
photo to let us say 5mm width at the centre and you took a photo with
a fixed lens camera with a 28mm lens and you cropped that to a 5mm
width at the centre then the images would be so close that you would
be hard put to tell them apart. And the depth of field that applies to
the fixed lens camera would be the same as that for the swing lens
camera just over that strip in the middle, should anyone care to
enlarge it. The only reason you have to adjust the CoC for the swing
lens camera is because of the greater film diagonal length. It has
nothing to do with the angle of coverage. If it swung back and forth
while the film got pulled through by some strange mechanism and so
only had an angle of 50 degrees on the film then the only thing that
matter for the CoC is the film diagonal.

It can't be both a 28mm and a 50mm DOF case at the same time ;-) Nor can a
28mm or a 50mm fixed lens DOF calculation take the actual coverage (110
degrees+ here) into account. Nor will a 24x36mm crop of the uniquely
cigar-shaped distorted image of a swing lens shot be the same in
appearance or coverage to a rectilinear lens shot of the same subject.

You are hung up on this idea of angle of coverage. It is nothing to do
with that. CoC is to do with film diagonal. DoF uses that CoC (which
is only to do with film diagonal) and the characteristics of a lens of
that focal length to give you the depth of field.

You could build a lens which projects 110+ degrees rectilinearly (or with
cigar shaped distortion, for extra $$ ;-) onto 24x56mm slice of film. Such
an equiv. rectilinear lens is circa 18-19mm on MF (6x6cm). There is no
rectilinear 35mm lens which can cover 24x56mm; an 18mm ultrawide on 35mm
is covering only 24x36mm, not 24x56mm.

Again you are making an invalid comparison. You are hung up on this
idea of angle of view. The image from a fixed lens camera giving a 110
degree angle of view and a swing lens camera giving the same angle of
view are very different. Just a glance will convince you they are very
different. For a start, in the centre of the image, the fixed lens
camera will show a much smaller image. the depth of field will be much
greater, therefore, for the fixed lens camera.

So I maintain that swing lens cameras are special cases, and that you
can't just take the DOF of the fixed lens (28mm in Horizon 202s, 50mm or
whatever in my hacked camera case), and say that the fixed DOF and
enlargement calculations all work the same.

I'm saying you can.

For one thing, those DOF calculators are working with the 35mm equiv.
format, 24x36mm, while I am enlarging from a 24x56mm piece of film. It is
an MF sized film crop of an MF sized image from an equiv. swing lens, not
a 24x36mm sized crop as used in the DOF calculator, right? ;-)

It just has a different CoC size due to the greater film diagonal.

my $.02

bobm

PS I suspect that you might be able to model the DOF factors, ignoring the
swing lens distortion, by creating an equiv. lens based on the film format
size (here 24x56mm) and equiv. wide angle rectilinear lens required to
capture the observed swing lens angle (here, 110+ degrees), say 18+mm for
6x6cm equiv. in this case. Now the problem is that the image height is not
that of an 18mm lens on 6x6cm, but rather, that of the underlying 28mm
lens being swung thru 110 degrees, yes?
´
It is nothing to do with angle of coverage.

So you are in effect working from a cropped and enlarged image on film. A
28mm lens for 35mm SLR covers 46 degrees vertically on the (same) 24mm
film width format in 24x36mm format. By comparison, the 18+mm on 6x6cm
covers about 114 degrees vertically. An 18mm lens on a 24x56mm format
would cover 67 degrees vertically, versus 46 degrees for the actual 28mm
lens (for a 35mm format) used. So perhaps this is equiv. to cropping the
18mm lens by 46/67 to 68%, and then enlarging back to 24x56mm size (1/.68
or 1.45X?). If you used a 50mm lens for 35mm format swung 110 degrees to
create a 24x56mm image, you would have to use an enlargement factor for
that lens, namely, 50mm covers 27 deg. vertically, so 27/67=.4 or 40% and
so 1/.4= 2.5X enlargement factor. And so on? ;-0)

In short, you can model a swing lens camera, but the horiz. coverage
(e.g., 110+ degrees) and film form (24x56mm) determines the equiv.
rectilinear lens. The actual lens focal length (28mm) and original format

You do not have to convert the swing lens coverage into an equivalent
rectilinear lens. The images are very different.

(24x36mm) is used to determine the vertical coverage only. In most cases,
this would be in effect a center crop and enlargement by some factor onto
the swing lens film format. Further enlargements from the 24x56mm film
slice would represent additional equiv. enlargements, with impacts on DOF
based on the original swing lens equivalent (here, 18mm), taking into
account the degree of vertical cropping and enlargement from the smaller
format lens (i.e., 28mm here). Phew! ;-0)

This is all different from the DOF case of a fixed 28mm lens on a 24x36mm
film slice ;-)

grins bobm

Angle of view -- nothing to do with -- simple as that.

Why do you think the film diagonal is independent of field of coverage in
determining the enlargeability? If this were the case, then all photos
with the same diagonals would have the same enlargeability, and they
don't! ;-) The field of view is critical because it determines how big the
images are on film, and hence, how much enlargement they can sustain.

Enlargeability is magnification factor(s); the image height is determined
by scale or size of the image; the limiting axis is usually the long axis
of the film where a large angle is compressed onto the film's horizontal
axis (e.g., 110+ degrees with horizon 202 on 24x56mm), i.e., you have an
18+ mm equiv. (on 6x6cm!) lens horizontal coverage being the limiting
factor for enlargement (not the vertical 28mm lens coverage of the fixed
lens, or 50mm lens if one is hacked and used etc.

quoting Roland:
view are very different. Just a glance will convince you they are very
different. For a start, in the centre of the image, the fixed lens
camera will show a much smaller image. the depth of field will be much
greater, therefore, for the fixed lens camera.
end-quote

How can the fixed 28mm lens show a smaller image in the film/image center
than the same 28mm lens rotated in a swing lens camera? Does the lens know
it is being rotated and changes magnification modes? Of course not ;-) The
"cigar" distortion of a swing lens camera is due to the ends of nearby
objects (such as a bench) which are farther away than the center (with the
camera focused at the center of a bench in this case obviously). With wide
angle lenses (such as 28mm), there is a rapid falloff in image height with
distance. So the farther away edges appear rather smaller than the nearer
center of the object (e.g., a park bench here). The result is the familiar
cigar shaped distortion of the swing lens camera.

But the image height of the various scenic elements in a swing lens camera
are the same as that projected by the center of the fixed lens used in the
swing lens design. Some types of lens corrections may not be available in
the swing lens design because only the center portion of the fixed lens is
used.

But the attraction of the swing lens design is not that it is a 28mm fixed
lens camera; rather, that it provides an angle of coverage (e.g., 110 to
120+ degrees) which is otherwise not available with rectilinear lenses (at
that price, anyway ;-). The image is one which simulates what we would see
with a much wider lens, and the enlargeability is determined by that
resulting image's qualities. Usually, as I noted, it is the longer axis
which limits our enlargeability capabilities first, so the CoC of the
short axis is not what is limiting us...

again, my $.02 ;-)

bobm

--
************************************************** *********************
* Robert Monaghan POB 752182 Southern Methodist Univ. Dallas Tx 75275 *
********************Standard Disclaimers Apply*************************

"Bob Monaghan" wrote in message
...

Why do you think the film diagonal is independent of field of coverage in
determining the enlargeability? If this were the case, then all photos
with the same diagonals would have the same enlargeability, and they
don't! ;-) The field of view is critical because it determines how big the
images are on film, and hence, how much enlargement they can sustain.

This is dead wrong. When you take a photograph, you take a photograph with a
composition. How much enlargement that composition can sustain has nothing
to do with the size of the images of the individual objects on the film.

Enlargeability is magnification factor(s); the image height is determined
by scale or size of the image; the limiting axis is usually the long axis
of the film

Wrong again. People who print panoramas figure out a way to print them at a
sensible size, for example by putting roll paper in their inkjet, or by
glueing the final print together from sections.

where a large angle is compressed onto the film's horizontal
axis (e.g., 110+ degrees with horizon 202 on 24x56mm), i.e., you have an
18+ mm equiv. (on 6x6cm!) lens horizontal coverage being the limiting
factor for enlargement (not the vertical 28mm lens coverage of the fixed
lens, or 50mm lens if one is hacked and used etc.

The horizontal direction is only a limiting factor if you are too lazy to
figure out how to print...

quoting Roland:
view are very different. Just a glance will convince you they are very
different. For a start, in the centre of the image, the fixed lens
camera will show a much smaller image. the depth of field will be much
greater, therefore, for the fixed lens camera.
end-quote

How can the fixed 28mm lens show a smaller image in the film/image center
than the same 28mm lens rotated in a swing lens camera?

Because he's talkinag about your comparison of an 18mm fixed lens with a
28mm swing lens...

But the image height of the various scenic elements in a swing lens camera
are the same as that projected by the center of the fixed lens used in the
swing lens design. Some types of lens corrections may not be available in
the swing lens design because only the center portion of the fixed lens is
used.

Don't you mean: "Some types of lens corrections _are not needed_ in the
swing lens design because only the center portion of the fixed lens is used.

But the attraction of the swing lens design is not that it is a 28mm fixed
lens camera; rather, that it provides an angle of coverage (e.g., 110 to
120+ degrees) which is otherwise not available with rectilinear lenses (at
that price, anyway ;-). The image is one which simulates what we would see
with a much wider lens,

The projection is quite different...

and the enlargeability is determined by that
resulting image's qualities. Usually, as I noted, it is the longer axis
which limits our enlargeability capabilities first, so the CoC of the
short axis is not what is limiting us...

No, the CoC is X/Y symmetrical across the whole frame in a swing lens
camera. It's when a non-swing lens camera is used for a panorama aspect
ratio that you would see asymmetrical CoCs at the edges of the frame. (Wide
angle lenses show oval CoCs at the edges, and since the edges are off-axis
in only one direction when you crop to a panoramic aspect ratio, the edges
will have CoCs elongated in the X direction.)

David J. Littleboy
Tokyo, Japan

(Bob Monaghan) wrote in message ...
Why do you think the film diagonal is independent of field of coverage in
determining the enlargeability? If this were the case, then all photos
with the same diagonals would have the same enlargeability, and they
don't! ;-) The field of view is critical because it determines how big the
images are on film, and hence, how much enlargement they can sustain.

This is not my opinion. This is the way CoC is determined. You know
that the CoC is larger for medium format than it is for 35mm format
and you know that CoC is larger for large format than medium format?
It is all to do with the film diagonal. Not my opinion, I stress. This
is the way it is worked out.

Enlargeability is magnification factor(s); the image height is determined
by scale or size of the image; the limiting axis is usually the long axis
of the film where a large angle is compressed onto the film's horizontal
axis (e.g., 110+ degrees with horizon 202 on 24x56mm), i.e., you have an
18+ mm equiv. (on 6x6cm!) lens horizontal coverage being the limiting
factor for enlargement (not the vertical 28mm lens coverage of the fixed
lens, or 50mm lens if one is hacked and used etc.

quoting Roland:
view are very different. Just a glance will convince you they are very
different. For a start, in the centre of the image, the fixed lens
camera will show a much smaller image. the depth of field will be much
greater, therefore, for the fixed lens camera.
end-quote

How can the fixed 28mm lens show a smaller image in the film/image center

I was taking about your hypothetical fixed -- 18+mm -- lens.

than the same 28mm lens rotated in a swing lens camera? Does the lens know
it is being rotated and changes magnification modes? Of course not ;-) The
"cigar" distortion of a swing lens camera is due to the ends of nearby
objects (such as a bench) which are farther away than the center (with the
camera focused at the center of a bench in this case obviously). With wide
angle lenses (such as 28mm), there is a rapid falloff in image height with

Change of image height due to the edges of an object being further
away is only a small part of what causes the cigar shape. It is mainly
caused by the swinging of the lens so that the lens is at a different
angle to the edges of an object than it is in the centre of the
object. Think of taking a picture of Buckingham palace from the
entrance gate. The edges of the building are maybe only 25% further
away from you than the centre of the building but you will still get
an extreme cigar shape. The cigar shape is mainly due to the lens
being perpendicular to the centre of the building when you take the
shot and at an angle for the edges of the building.

distance. So the farther away edges appear rather smaller than the nearer
center of the object (e.g., a park bench here). The result is the familiar
cigar shaped distortion of the swing lens camera.

But the image height of the various scenic elements in a swing lens camera
are the same as that projected by the center of the fixed lens used in the
swing lens design. Some types of lens corrections may not be available in
the swing lens design because only the center portion of the fixed lens is
used.

You seem to be forgetting about this 18+mm lens of yours. That was the
whole crux of your argument. What happened to it?

But the attraction of the swing lens design is not that it is a 28mm fixed
lens camera; rather, that it provides an angle of coverage (e.g., 110 to
120+ degrees) which is otherwise not available with rectilinear lenses (at
that price, anyway ;-). The image is one which simulates what we would see
with a much wider lens, and the enlargeability is determined by that
resulting image's qualities. Usually, as I noted, it is the longer axis
which limits our enlargeability capabilities first, so the CoC of the
short axis is not what is limiting us...

again, my $.02 ;-)

bobm

You seem to have changed the subject and forgotten about your 18+mm
lens. Reemmber you maintained that the DoF of a 28mm swing lens photo
could not be scientifically determined because it should be more like
that of an 18+mm rectilinear lens that covered the same field of view
as a 28mm swing lens? I hope you haven't lost your 18+mm lens. They
are precious items.

quoting David: This is dead wrong. When you take a photograph, you take a
photograph with a composition. How much enlargement that composition can
sustain has nothing to do with the size of the images of the individual
objects on the film. end-quote:

Huh? If how much enlargement is possible has nothing to do with the size
of the images on the film, what does it have to do with?

Isn't this why we use MF film, as the image sizes can be larger (up to 56
mm high?)? And LF, up to 4x5"?

My whole starting point on this discussion was that DOF/CoC were only
aspects of image magnification limitations. Clearly we disagree, based on
your above quote ;-)

If I shoot a portrait with a 110mm kowa 66 portrait/macro lens, full
frame, 56x56mm, that image can be enlarged without falling apart a lot
more than the same image shot at the same distance/perspective with my
35mm rectilinear Kowa 66 wide angle lens, where the face is now maybe 10mm
high, right? ;-) Image height on film is controlling enlargeability,
right?

In this same thread, Roland also takes me to task by observing that CoC
varies for 35mm, MF, and LF. This is quite true. But why does it vary by
format? Why should the 35mm CoC be 0.03mm while the 6x6cm CoC should be
0.06mm as in one DOF calculator? The film is the same, right? ;-)

The answer is that these "relaxed" CoC for MF and LF are due to the lesser
enlargements usually required by such MF and LF negatives to produce a
similar quality (vs. 35mm) enlargement, typically 8x10", yes? ;-)

Again, CoC isn't an aspect of the film format, but the expected
enlargement size (here, 8x10") and the degree of enlargement or
magnification required by that format (35mm, MF, LF) to produce that size
(8x10" here).

So again, I state that DOF and CoC are based on the magnification required
rather than constants as often assumed in simplistic DOF calculators. The
better DOF calculators let you put in your own COC for this reason etc.
;-)

I don't agree that the cigar shaped distortion is primarily due to the
edge effects of the swinging lens perspective, as I think Roland is
suggesting. I still think it is due to wide angle effects as I noted. The
edge effects produced by the swinging lens are similar to LF camera lens
swings and tilts, as with the classic studies of box edge effects in
Stroebel and Stone and other LF texts. Again, these are not seen in the
fixed 28mm lens until it is put into a swing lens camera design ;-) The
reason wide angle distortion is less obvious in our fixed w.a. lenses is
largely the result of rectilinear lens designs, which specifically work to
keep lines straight and a flat focal plane etc. etc. ;-) If you look thru
an afocal 0.42x superwide adapter, which isn't corrected for rectilinear
projection, you get a better feel for wide angle dropoff effects ;-)

Finally, I haven't abandoned the 18mm equiv. lens case, I just worked out
the equivalent swing lens factors (e.g., for the horizon 202s, an 18mm+
horizontal equiv on 24x56mm with a vertical enlargement (or cropping)
factor equiv. to the 28mm lens used (or 50mm lens or whatever is used on
the swing lens camera).

my $.02 again

bobm
--
************************************************** *********************
* Robert Monaghan POB 752182 Southern Methodist Univ. Dallas Tx 75275 *
********************Standard Disclaimers Apply*************************

(Bob Monaghan) wrote in message ...

snip

Finally, I haven't abandoned the 18mm equiv. lens case, I just worked out
the equivalent swing lens factors (e.g., for the horizon 202s, an 18mm+
horizontal equiv on 24x56mm with a vertical enlargement (or cropping)
factor equiv. to the 28mm lens used (or 50mm lens or whatever is used on
the swing lens camera).

I'm glad you are still in there with the 18+mm lens. How would it
apply in this case.....

Suppose somebody had a 35mm camera with a 28mm lens. They have filed
away at the edges of the frame so there is no gap between the frames.
They take a picture, turn through an angle of 65° and take another
picture. So this combined picture covers an angle of 130° and the
frame size is 24mm x 76mm. Now consider the depth of field. Would the
depth of field be that for an 18+mm lens (or an even shorter focal
length) used with a 6x9 camera such that it could cover a view of 130°
on 76mm wide? Or would the depth of field be that for a 28mm lens, but
adjusted for a larger CoC size than for the 35mm format due to the
double width?

(RolandRB) wrote in message . com...
(Bob Monaghan) wrote in message ...

snip

Finally, I haven't abandoned the 18mm equiv. lens case, I just worked out
the equivalent swing lens factors (e.g., for the horizon 202s, an 18mm+
horizontal equiv on 24x56mm with a vertical enlargement (or cropping)
factor equiv. to the 28mm lens used (or 50mm lens or whatever is used on
the swing lens camera).

I'm glad you are still in there with the 18+mm lens. How would it
apply in this case.....

Suppose somebody had a 35mm camera with a 28mm lens. They have filed
away at the edges of the frame so there is no gap between the frames.
They take a picture, turn through an angle of 65° and take another
picture. So this combined picture covers an angle of 130° and the
frame size is 24mm x 76mm. Now consider the depth of field. Would the
depth of field be that for an 18+mm lens (or an even shorter focal
length) used with a 6x9 camera such that it could cover a view of 130°
on 76mm wide? Or would the depth of field be that for a 28mm lens, but
adjusted for a larger CoC size than for the 35mm format due to the
double width?

Bob? Are you there?

again, to me DOF is all about magnification and enlargeability potential

the COC is determined by the degree of magnification or enlargement factor
needed, which is why 35mm have more restrictive COC (0.03mm) than 6x6cm
(e.g., COC of 0.06mm) etc. These are for enlargements to the same size
(generally assumed to be 8x10" print in most DOF calculations it seems)
from different formats.

Since 6x6cm is 3.8-4.2X area of 35mm, the linear axis is about 2X, and the
COC allowed is also 2*0.03mm of 35mm case, right? In fact, the COC is
determined from allowable enlargement quality factors (e.g., 8 lpmm on
print) at one film format size and then the rest fall out from simple
geometry, yes?

you could take the negative and slice it up, then tape together and get
a similar situation to the one you suggest; or use a shift lens (my nikkor
shift lens can produce 24x72+mm images at max. shifts) etc. But if you end
up with a MF sized negative, then it is as if you were shooting on MF
film. You will only enlarge by half the factor as for a 35mm sized film.

finally, my argument has continued to be that swing lenses are special
cases which are not well modeled by DOF calculators aimed at fixed focal
length lenses. One more reason is the underlying assumptions in fixed lens
DOF calculators that COC is related to format size (e.g., 24x72mm) rather
than lens coverage (e.g., 35mm format). It is after all the film at a
given format size that is being enlarged to some limiting quality
magnification level and size (hence, determining DOF).

Most DOF calculators use the film format size to pick the magnification or
enlargement factors which in turn determine COC (at a given print size),
then use these COC limits with the incident light cones on the film (from
lens focal length) to determine DOF values within those limits, right? So
the 24x56mm or 24x72mm+ case uses different format size than 35mm,
different diagonals, different enlargement or magnification factors, and
different COC than for the smaller 35mm format, right? That's another
reason you aren't likely to get the correct answer from a fixed lens DOF
calculator, since none that I've seen handle swing lens situations like
this ;-)

grins bobm
--
************************************************** *********************
* Robert Monaghan POB 752182 Southern Methodist Univ. Dallas Tx 75275 *
********************Standard Disclaimers Apply*************************