Another nail in the view camera coffin?

Leonard Evens wrote in message ...
Robert Feinman wrote:
Leaving aside film size, the two features that view cameras still
have over other formats are the ability to adjust perspective and
the plane of focus.
By using a digital editor one can generate the same perspective
effects from an image taken with a conventional camera afterwards.

I've been playing with this feature in Photoshop and have put up
an additional tip about this on my web site. This one shows the
creative uses the extreme perspective adjustments can yield.

As I'm sure you are aware, it is not quite that simple. In addition to
making verticals parallel, you have to adjust the vertical height. At
one point there was a long discussion of that in this newsgroup, and I
think I convinced everyone who understood the issues that there is no
way to do that accurately without using some additional information,
which is not in the picture, although in many cases it can be deduced
from what is there.

Just follow the tips link on my home page, if you are interested.

I still haven't solved the plane of focus problem, however...

Nor are you likely to.

Also, you can't apply a vertical or horizontal shift. It is so
fundamental that often view camera users forget its importance. The
single most important choice one makes when planning a picture is the
point of view. The second is just how to frame the subject. That
depends on the choice of focal length, but it also depends on shifts.
If you point the camera up and then digitally correct to create vertical
parallels, you don't change what is framed. One of your examples
indicates that pretty well. The road (empty of interest) in front of
the church is included in all the corrected versions. Of course, you
could crop it out, but that would reduce the resolution of fine detail
if enlarged to the same size final print. With a view camera you can
usualy frame the picture right to start. And of course, you can also
apply digital perspective or other corrections afterwards if you want.

Finally, I'm still thinking about it, but I am not sure by perspective
transformations of the plane image you can produce the same three
dimensional perspective relations you would be able to produce with a
view camera using shifts. My initial intuition is that you can't, but
further thought may show me otherwise.

I regularly go out with a digital camera to scout out interesting
pictures. I then digitally correct perspectives to get some idea of
what they would look like using my view camera. But the results never
seem up to what I can get directly with the view camera using movements.

The tools for simulating camera movements in Photoshop are admittedly
very crude. However, since perspective is defined by the camera
position (more exactly the entrance pupil of the lens), then any
geometrical effect you can produce by shifting the front or back or by
tilting the rear can be precisely duplicated in software. Its just a
matter of doing the mapping correctly, which as you point out
Photoshop by itself does not do. Use Panorama Tools along with a good
interface like PTAssembler if you want to get it right. Note that I'm
not talking about Scheimflug effects.


Brian
www.caldwellphotographic.com

2. Rises and shifts, avoiding reflections and the like.
Any of these movements put the lens somewhere else that the original
spot. To achieve this result with a fixed camera you put the lens at
exactly the same spot pointing in exactly the same direction. You then
adjust the image in the editor to simulate the back movement that is
needed to get the desired result.

This explanation ignores the fact that using LF for reflection reduction can
entail the use of a portion of the outer area of the image circle, so that the
frame includes the area of interest while the lens has been moved so as to not
'see' the offending reflection. Using Photoshop cannot 'create' a part of the
scene not captured into the digital CCD. Yes, the reflection may be gone, but
so too is the area in the scene which needs to be included, because it is
beyond the image circle captured by the CCD!

--wilt

2. Rises and shifts, avoiding reflections and the like.
Any of these movements put the lens somewhere else that the original
spot. To achieve this result with a fixed camera you put the lens at
exactly the same spot pointing in exactly the same direction. You then
adjust the image in the editor to simulate the back movement that is
needed to get the desired result.

This explanation ignores the fact that using LF for reflection reduction can
entail the use of a portion of the outer area of the image circle, so that the
frame includes the area of interest while the lens has been moved so as to not
'see' the offending reflection. Using Photoshop cannot 'create' a part of the
scene not captured into the digital CCD. Yes, the reflection may be gone, but
so too is the area in the scene which needs to be included, because it is
beyond the image circle captured by the CCD!

--wilt

Robert Feinman wrote:


4. Image quality issues. I explicitly stated that I was ignoring
film size as not relevant to the point I'm making.

The problem you don't seem to understand is even if you used a 4X5 scanning
back to create giant files, post shooting digital corrections create a huge
loss in quality over one shot with the perspective corrected at the time it
was shot.



So your choice of equipment is dictated by your ultimate goals
and expected audience.

If we were interested in accepting lower quality work, we wouldn't be
shooting 4X5 would we?



--

Stacey

Robert Feinman wrote:


4. Image quality issues. I explicitly stated that I was ignoring
film size as not relevant to the point I'm making.

The problem you don't seem to understand is even if you used a 4X5 scanning
back to create giant files, post shooting digital corrections create a huge
loss in quality over one shot with the perspective corrected at the time it
was shot.



So your choice of equipment is dictated by your ultimate goals
and expected audience.

If we were interested in accepting lower quality work, we wouldn't be
shooting 4X5 would we?



--

Stacey

Robert Feinman wrote:
In article ,

Lighten up guys!

Just because I made a teasing title for the post doesn't mean I'm
denigrating LF. The point of the post was to illustrate an example
of the extreme amount of distortion you can achieve with a digital
editor.
To take some of the points mentioned in the thread:

1. Can digital editing create the same image shape as that from
corrections with a camera. Yes. The proof is that there is a simple
linear transformation that can shift every point from the original
image to the place it would have been if tilts had been used. It may
not be easy to determine this empirically, but there is such a function.

It is actually called a projective transformation, not a linear
transformation, but you are essentially right about that. (A linear
transformation can't move a vanishing point to infinity as in making
verticals parallel.) But that only applies to tilts or swings and
assumes the entire image is essentially in focus. So if you are
talking about pictures taken with pinhole cameras, either fixed or with
movements, then what you say is correct. But you can't bring something
totally out of focus in focus.

2. Rises and shifts, avoiding reflections and the like.
Any of these movements put the lens somewhere else that the original
spot. To achieve this result with a fixed camera you put the lens at
exactly the same spot pointing in exactly the same direction. You then
adjust the image in the editor to simulate the back movement that is
needed to get the desired result.

No that is not quite right. The only way you could do it with a fixed
camera and have the same point of view (interpreted literally) is if the
image produced by the fixed camera were essentially the same size as the
image circle of the lens used in the view camera---after normalizing for
differences in format. In principle, you could do that by choosing a
sufficiently wide angle lens for the fixed camera, but suppose you are
already using an extreme wide angle lens in the view camera? What you
call similating back movement amounts to cropping an existing image and
of course you can do that in a photoeditor or by conventional darkroom
techniques. But you can't digital move outside the existing image and
magically produce what was in the scene.


3. The meaning of perspective. From dictionary.com
"The technique of representing three-dimensional objects and depth
relationships on a two-dimensional surface." (definition 4). So
perspective commonly includes point of view as well as geometric
considerations.

Agreed. But together with point of view---i.e., the point at which the
camera lens sits---there is a line of sight which can be defined as the
line through the lens perpendicular to the film plane. That determines
what appears in front of what among other things. In a camera
without movements it is the same as the lens axis. When you point such
a camera up, that line of sight points upward. It won't be the same as
the line of sight in a view camera picture in which the back has been
kept vertical and a shift has been used to include the top of a
building. So when you correct the perspective digitally in a
photoeditor, you will still have that line of sight.

If it were possible to change the line of sight then in one of your
examples you could have digitally manipulated the image so it was
exactly the same as looking down straight from above, which you point
out yourself would require an aerial photo to accomplish, and can't be
done digitally.

4. Image quality issues. I explicitly stated that I was ignoring
film size as not relevant to the point I'm making. There have
been never-ending discussions on how much quality is enough and
I was trying to avoid rehashing those arguments again. I think
it is generally agreed that larger film sizes produce better
print quality at moderate to large degrees of magnification. In
addition, there is great variability in what viewers perceive as
quality, with the number of discriminating observers seemingly
getting smaller each year as more people view images online and
never see real photographic prints. Many people are acutely aware
of quality defects in prints, but many more seem oblivious.
So your choice of equipment is dictated by your ultimate goals
and expected audience.

If you look at my final image of the building lobby in my second
perspective tip you will see an effect that could not be achieved
from that vantage point with a view camera.
The floor has been transformed into a perfect rectangle. This implies
that the back of the camera would have to be parallel to the floor.
The next adjustment would have to be to use the rise to clear the
balcony that I was standing on. There is not a camera made that has that
amount of rise or lens with enough covering power.
The final image has very pronounced distortions. The people are all
stretched. I was just trying to get photographers to experiment and not
be constrained by mechanical limitations. Whether the image is an
artistic success is for the creator to decide.

It can't be accomplished in a view camera without using digital
techniques. But who said you can't use a view camera and then also
digitally edit it? That way you would have the best of both worlds.
In fact, some of us do that regularly. Each photograph presents a
challenge in its own right. My feeling is that you should get as much
right in the camera to start. After that, it is a matter of choice as
to how much further you want to go.

Needless to say few large format photographers never use any other kind
of camera. So it is not a matter of either/or, but rather of when. So
far, you haven't been convincing in the (tongue in cheek or wherever it
s) suggestion that view cameras are not obsolete because of the advent
of digital editing.


Just so we are all on the same page, here is the link:
http://robertdfeinman.com/tips/tip32/tip32.html

Robert Feinman wrote:
In article ,

Lighten up guys!

Just because I made a teasing title for the post doesn't mean I'm
denigrating LF. The point of the post was to illustrate an example
of the extreme amount of distortion you can achieve with a digital
editor.
To take some of the points mentioned in the thread:

1. Can digital editing create the same image shape as that from
corrections with a camera. Yes. The proof is that there is a simple
linear transformation that can shift every point from the original
image to the place it would have been if tilts had been used. It may
not be easy to determine this empirically, but there is such a function.

It is actually called a projective transformation, not a linear
transformation, but you are essentially right about that. (A linear
transformation can't move a vanishing point to infinity as in making
verticals parallel.) But that only applies to tilts or swings and
assumes the entire image is essentially in focus. So if you are
talking about pictures taken with pinhole cameras, either fixed or with
movements, then what you say is correct. But you can't bring something
totally out of focus in focus.

2. Rises and shifts, avoiding reflections and the like.
Any of these movements put the lens somewhere else that the original
spot. To achieve this result with a fixed camera you put the lens at
exactly the same spot pointing in exactly the same direction. You then
adjust the image in the editor to simulate the back movement that is
needed to get the desired result.

No that is not quite right. The only way you could do it with a fixed
camera and have the same point of view (interpreted literally) is if the
image produced by the fixed camera were essentially the same size as the
image circle of the lens used in the view camera---after normalizing for
differences in format. In principle, you could do that by choosing a
sufficiently wide angle lens for the fixed camera, but suppose you are
already using an extreme wide angle lens in the view camera? What you
call similating back movement amounts to cropping an existing image and
of course you can do that in a photoeditor or by conventional darkroom
techniques. But you can't digital move outside the existing image and
magically produce what was in the scene.


3. The meaning of perspective. From dictionary.com
"The technique of representing three-dimensional objects and depth
relationships on a two-dimensional surface." (definition 4). So
perspective commonly includes point of view as well as geometric
considerations.

Agreed. But together with point of view---i.e., the point at which the
camera lens sits---there is a line of sight which can be defined as the
line through the lens perpendicular to the film plane. That determines
what appears in front of what among other things. In a camera
without movements it is the same as the lens axis. When you point such
a camera up, that line of sight points upward. It won't be the same as
the line of sight in a view camera picture in which the back has been
kept vertical and a shift has been used to include the top of a
building. So when you correct the perspective digitally in a
photoeditor, you will still have that line of sight.

If it were possible to change the line of sight then in one of your
examples you could have digitally manipulated the image so it was
exactly the same as looking down straight from above, which you point
out yourself would require an aerial photo to accomplish, and can't be
done digitally.

4. Image quality issues. I explicitly stated that I was ignoring
film size as not relevant to the point I'm making. There have
been never-ending discussions on how much quality is enough and
I was trying to avoid rehashing those arguments again. I think
it is generally agreed that larger film sizes produce better
print quality at moderate to large degrees of magnification. In
addition, there is great variability in what viewers perceive as
quality, with the number of discriminating observers seemingly
getting smaller each year as more people view images online and
never see real photographic prints. Many people are acutely aware
of quality defects in prints, but many more seem oblivious.
So your choice of equipment is dictated by your ultimate goals
and expected audience.

If you look at my final image of the building lobby in my second
perspective tip you will see an effect that could not be achieved
from that vantage point with a view camera.
The floor has been transformed into a perfect rectangle. This implies
that the back of the camera would have to be parallel to the floor.
The next adjustment would have to be to use the rise to clear the
balcony that I was standing on. There is not a camera made that has that
amount of rise or lens with enough covering power.
The final image has very pronounced distortions. The people are all
stretched. I was just trying to get photographers to experiment and not
be constrained by mechanical limitations. Whether the image is an
artistic success is for the creator to decide.

It can't be accomplished in a view camera without using digital
techniques. But who said you can't use a view camera and then also
digitally edit it? That way you would have the best of both worlds.
In fact, some of us do that regularly. Each photograph presents a
challenge in its own right. My feeling is that you should get as much
right in the camera to start. After that, it is a matter of choice as
to how much further you want to go.

Needless to say few large format photographers never use any other kind
of camera. So it is not a matter of either/or, but rather of when. So
far, you haven't been convincing in the (tongue in cheek or wherever it
s) suggestion that view cameras are not obsolete because of the advent
of digital editing.


Just so we are all on the same page, here is the link:
http://robertdfeinman.com/tips/tip32/tip32.html

brian wrote:
Leonard Evens wrote in message ...

Robert Feinman wrote:

Leaving aside film size, the two features that view cameras still
have over other formats are the ability to adjust perspective and
the plane of focus.
By using a digital editor one can generate the same perspective
effects from an image taken with a conventional camera afterwards.

I've been playing with this feature in Photoshop and have put up
an additional tip about this on my web site. This one shows the
creative uses the extreme perspective adjustments can yield.

As I'm sure you are aware, it is not quite that simple. In addition to
making verticals parallel, you have to adjust the vertical height. At
one point there was a long discussion of that in this newsgroup, and I
think I convinced everyone who understood the issues that there is no
way to do that accurately without using some additional information,
which is not in the picture, although in many cases it can be deduced
from what is there.


Just follow the tips link on my home page, if you are interested.

I still haven't solved the plane of focus problem, however...

Nor are you likely to.

Also, you can't apply a vertical or horizontal shift. It is so
fundamental that often view camera users forget its importance. The
single most important choice one makes when planning a picture is the
point of view. The second is just how to frame the subject. That
depends on the choice of focal length, but it also depends on shifts.
If you point the camera up and then digitally correct to create vertical
parallels, you don't change what is framed. One of your examples
indicates that pretty well. The road (empty of interest) in front of
the church is included in all the corrected versions. Of course, you
could crop it out, but that would reduce the resolution of fine detail
if enlarged to the same size final print. With a view camera you can
usualy frame the picture right to start. And of course, you can also
apply digital perspective or other corrections afterwards if you want.

Finally, I'm still thinking about it, but I am not sure by perspective
transformations of the plane image you can produce the same three
dimensional perspective relations you would be able to produce with a
view camera using shifts. My initial intuition is that you can't, but
further thought may show me otherwise.

I regularly go out with a digital camera to scout out interesting
pictures. I then digitally correct perspectives to get some idea of
what they would look like using my view camera. But the results never
seem up to what I can get directly with the view camera using movements.


The tools for simulating camera movements in Photoshop are admittedly
very crude. However, since perspective is defined by the camera
position (more exactly the entrance pupil of the lens), then any
geometrical effect you can produce by shifting the front or back or by
tilting the rear can be precisely duplicated in software.

Look. I believe I'm the mathematician here. I know all about
projective mappings, which is what you are talking about. I've taught
courses about such things for years.

But it doesn't require a mathematician. When you take a picture you
select some part of what is theoretically viewable from the camera
position. If you haven't selected something, you can't recreate it by a
transformation. Of course, if you arrange focal lengths and formats
properly, you could produce a very large image with a "fixed" camera and
then crop rather than doing the selection in the camera by
rise/fall/shift. There is nothing digital about that. You could do it
by conventional cropping in a darkroom. But there are practical
limitations to doing it that way. In a view camera, you do it in the
camera by adjusting the position of the film frame with respect to the
lens axis.

In addition, you can't change the line of sight, which refers to three
dimensional relations by a two dimensional projective transformation of
the image. (You could do it by a three dimensional transformation, but
you don't have that option.) As someone pointed out, if you see
yourself in a photograph of your mirror, you can't change the line of
vsight so you no longer see yourself by a projective transformation.
(You can of course erase your image, but you could also digitally
simulate the image from scratch without ever taking a picture.)

Its just a
matter of doing the mapping correctly, which as you point out
Photoshop by itself does not do. Use Panorama Tools along with a good
interface like PTAssembler if you want to get it right. Note that I'm
not talking about Scheimflug effects.


Brian
www.caldwellphotographic.com

brian wrote:
Leonard Evens wrote in message ...

Robert Feinman wrote:

Leaving aside film size, the two features that view cameras still
have over other formats are the ability to adjust perspective and
the plane of focus.
By using a digital editor one can generate the same perspective
effects from an image taken with a conventional camera afterwards.

I've been playing with this feature in Photoshop and have put up
an additional tip about this on my web site. This one shows the
creative uses the extreme perspective adjustments can yield.

As I'm sure you are aware, it is not quite that simple. In addition to
making verticals parallel, you have to adjust the vertical height. At
one point there was a long discussion of that in this newsgroup, and I
think I convinced everyone who understood the issues that there is no
way to do that accurately without using some additional information,
which is not in the picture, although in many cases it can be deduced
from what is there.


Just follow the tips link on my home page, if you are interested.

I still haven't solved the plane of focus problem, however...

Nor are you likely to.

Also, you can't apply a vertical or horizontal shift. It is so
fundamental that often view camera users forget its importance. The
single most important choice one makes when planning a picture is the
point of view. The second is just how to frame the subject. That
depends on the choice of focal length, but it also depends on shifts.
If you point the camera up and then digitally correct to create vertical
parallels, you don't change what is framed. One of your examples
indicates that pretty well. The road (empty of interest) in front of
the church is included in all the corrected versions. Of course, you
could crop it out, but that would reduce the resolution of fine detail
if enlarged to the same size final print. With a view camera you can
usualy frame the picture right to start. And of course, you can also
apply digital perspective or other corrections afterwards if you want.

Finally, I'm still thinking about it, but I am not sure by perspective
transformations of the plane image you can produce the same three
dimensional perspective relations you would be able to produce with a
view camera using shifts. My initial intuition is that you can't, but
further thought may show me otherwise.

I regularly go out with a digital camera to scout out interesting
pictures. I then digitally correct perspectives to get some idea of
what they would look like using my view camera. But the results never
seem up to what I can get directly with the view camera using movements.


The tools for simulating camera movements in Photoshop are admittedly
very crude. However, since perspective is defined by the camera
position (more exactly the entrance pupil of the lens), then any
geometrical effect you can produce by shifting the front or back or by
tilting the rear can be precisely duplicated in software.

Look. I believe I'm the mathematician here. I know all about
projective mappings, which is what you are talking about. I've taught
courses about such things for years.

But it doesn't require a mathematician. When you take a picture you
select some part of what is theoretically viewable from the camera
position. If you haven't selected something, you can't recreate it by a
transformation. Of course, if you arrange focal lengths and formats
properly, you could produce a very large image with a "fixed" camera and
then crop rather than doing the selection in the camera by
rise/fall/shift. There is nothing digital about that. You could do it
by conventional cropping in a darkroom. But there are practical
limitations to doing it that way. In a view camera, you do it in the
camera by adjusting the position of the film frame with respect to the
lens axis.

In addition, you can't change the line of sight, which refers to three
dimensional relations by a two dimensional projective transformation of
the image. (You could do it by a three dimensional transformation, but
you don't have that option.) As someone pointed out, if you see
yourself in a photograph of your mirror, you can't change the line of
vsight so you no longer see yourself by a projective transformation.
(You can of course erase your image, but you could also digitally
simulate the image from scratch without ever taking a picture.)

Its just a
matter of doing the mapping correctly, which as you point out
Photoshop by itself does not do. Use Panorama Tools along with a good
interface like PTAssembler if you want to get it right. Note that I'm
not talking about Scheimflug effects.


Brian
www.caldwellphotographic.com

"Leonard Evens" wrote in message
...
brian wrote:
Leonard Evens wrote in message
...

Robert Feinman wrote:

Look. I believe I'm the mathematician here.

True! I laughed out loud with relief when I read that, Leonard. Be
assertive! You ARE the mathematician! There is no doubt. Gosh, after all
these decades, to find a scholar making the difference here is so affirming.
Thanks for being here, sir.