Parallel shift and Scheimpflug lens tilt
Parallel shift and
Scheimpflug lens tilt
The use and ideal design of technical cameras
ALPA Capaul & Weber Ltd.
Neptunstrasse 96
POB 1858
8032 Zürich / Switzerland
Phone: +41 44 383 92 22
Fax: +41 44 382 01 80
Layout: ALPA of Switzerland
Text & Graphics: Walter E. Schön, Munich/Germany
Version: 06.2012
All content © ALPA Capaul & Weber Ltd./Walter E. Schön, 2012
Parallel shift and Scheimpflug lens tilt
The use and ideal design of technical cameras
by Walter E. Schön
Adjustable technical cameras are intended to allow the
photographer to correct perspective distortion (“converging
lines”) when taking shots with a slanted view and to gain
more depth of field for the object to be imaged without having to stop down overmuch, that is to avoid diffraction blur.
Such technical cameras therefore offer a mechanical parallel shift of the back and/or of the lens to correct perspective distortion and a mechanical tilting device of the
back and/or of the lens to displace the depth of field in
space for a better adaptation to the extent of the motif.
What does “perspective” mean in photography and how is it influenced?
The word “perspective” comes from the Latin perspicere =
to look through and in photography it is used to express how
a three-dimensional object is imaged in a plane such that a
viewer will still experience a spatial impression.
To obtain ideal results in this respect, the viewer has to see
every point A‘ of the image in the same direction of gaze as
he would have seen the corresponding point A of the real
spatial object from a specific (possibly different) location.
The “location” here is the position at which the viewer’s eye
is located, more precisely the center of the pupil of his viewing eye. The image will appear differently depending on the
inclination of the image plane.
If we transfer this principle to a camera, what is decisive for
the perspective of the object in the photo is where the center of the lens entry pupil (or aperture) is located and how
the image plane is orientated relative to the object (parallel
or slanted) and to the direction of gaze (perpendicular or
This image, which is viewed perpendicular to the image
plane, has a correct perspective:
Each point of the object is on
the same line of vision as the
corresponding image point.
The vertical lines also remain
vertical and parallel in the image
because the image plane is parallel
to the vertical lines of the object.
When viewed as here, this image
also has the correct perspective
even though the image plane
is slightly slanted.
However, this image then
has „converging lines“ if it is
not viewed obliquely, but perpendicular to the slanted image plane.
How does parallel shift work and why is it needed?
If e.g. a high building is photographed with a camera pointing upward at a slant to get the whole image in the shot,
converging lines are produced: The vertical edges converge toward the top of the image to a vanishing point; the
building narrows at the top and is also compressed vertically (relative to its width).
If the building is photographed so that the image plane (=
sensor plane or film plane) of the camera is aligned parallel
to these perpendicular edges, they also remain perpendicular, and so parallel to one another, in the image. With
non-adjustable cameras, this is only possible with a horizontal taking direction (= vertical image plane).
In architectural shots, the upper part of the building is
often cut off and too much foreground appears in the
image. To take the whole building with less foreground, a
non-adjustable camera has to be held at a slant or the taking
location has to be at around half the height of the building
instead of close to the ground as with the natural viewpoint
of a pedestrian.
Exactly the same problems occur with product shots of objects with perpendicular edges which are usually taken
obliquely from above; the only difference is that then the
perpendicular edges converge toward the bottom in the image instead of toward the top.
If, however, the camera provides the possibility of moving
the lens or the back up and down, it is possible to take
shots with the camera obliquely from above or below within certain limits (which depend on the maximum movement
distance and the angle of view of the lens) while still
keeping the image plane perpendicular. The corresponding shots show the object without converging
lines. The ratio between height and width remains correct,
that is with natural proportions.
In the first instance, it is irrelevant whether the lens is
moved upward or the back downward for a shot
obliquely from below because it is only important that the
center of the lens is higher than the center of the image area
(of the frame in the back). This also applies in an analogous
way (reverse upward and downward) for a taking direction
obliquely from above. It is therefore initially irrelevant
whether the technical camera offers a parallel shift of
the lens or of the back or even both.
Image plane slanted
The camera, which is pointed obliquely upwardly, admittedly images the whole building and with only a little foreground, as desired, but does show converging lines.
Image plane vertical
The horizontally aligned camera reproduces the building
without distortion at a normal eye level, but not completely and with too much foreground.
Back shifted downwards
If the back is moved down (upper image) or if the lens is
moved up (lower image), a complete image without
distortion is possible.
Since a vertical movement of the lens causes a visible
change in the perspective in close-up shots, it is usually
more favorable to move the back, at least in the near
zone or with a very close foreground. If only the lens can be
moved (e.g. with 35 mm cameras with a “shift lens”), it is
possible that the entire camera on the tripod will have to be
lowered or raised by the vertical movement of the lens for
close-up shots after the lens movement to correct the converging lines so that the lens is in its original position.
Lens shifted upwards
How can the depth of field be utilized better with a Scheimpflug tilt?
Most objects which are photographed are three-dimensional. However, the lens only provides a sharp image for a
certain distance depending on its focus. The range at either side of the adjustment distance in which the increasing
blur is still too small to be disturbingly visible is called the
depth of field. Its boundaries are defined by the maximum
permitted diameter of the circle of confusion d which
has amounted to 1/1400 of the image diagonal for around
80 years in analog photography. In a sensor format of 24 mm
x 36 mm, this is approx. 0.03 mm. All standard depth of field
scales and tables are based on this.
The better lenses and films have become, the more questionable this value has become because a point image of
0.03 mm diameter is already perceived as blurred relative to
the maximum resolution available today when viewing
greatly magnified images. The latest high-resolution digital
backs and lenses rather require half this value, that is
around 0.015 mm for 35 mm. It is necessary to stop down
by two more f-numbers for the same depth of field.
Unfortunately, the diffraction blur increases greatly on
stopping down (by a factor of 2 for two f-stops). This sets
limits on the stepping down: If you make very high demands on the image definition and so stop down less,
you have to accept lower depth of field.
Theodor Scheimpflug proposed one way of escaping this
dilemma more than 100 years ago: It is not possible to
gain more depth of field by tilting the lens or the back, but
it can be arranged differently (obliquely) in space and can
so in many cases (but unfortunately not always) be
better adapted to the spatial extent of the motif to be
imaged with high definition.
If the image plane I and the main lens plane L on the image
side (we can consider the plane of the diaphragm blades for
this purpose) are parallel to one another, the sharply imaged
plane P and the planes A and B bounding the depth of field
from the front and from the back are also parallel:
Depth of field
In some technical cameras, the lens can be tilted so that the
main lens plane L intersects the image plane I at the straight
line C (image below). The sharp plane P in the image then
likewise extends obliquely through the straight line C and
the depth of field (yellow zone) previously bounded by parallel planes becomes wedge-shaped. The edge W of the
wedge is before C where the non-tilted lens plane intersects
the plane P. When the focus is changed, the depth of field
wedge rotates about the edge W of the wedge: Focusing at
a greater distance (image plane I is moved from the red to
the blue position) tilts the wedge down. Paragraph 7 on
page 5 explains how the Scheimpflug lens tilt has to be
carried out in practice.
.I L
Different focusing (red/blue),
same lens tilt angle
How do I have to focus to ensure the best depth field?
Without a Scheimpflug lens tilt, the image plane I and the
lens plane L are parallel to one another and to the plane P
in the motif sharp in the image.
It is first impossible to know which plane P to focus on and
which f-number to step down to in order to get the depth of
field over the total motif to be imaged as sharp. The popular
rule that the plane P lies at the end of the first third between
the front boundary A and the rear boundary B of the depth
of field is unfortunately usually incorrect.
There is only one third depth of field in front of and two
thirds behind the plane P when the rear border B is exactly
twice as far away as the front border A. Otherwise different
ratios of almost 1:1 (with macro shots) to 1: ∞ (at depth of
field up to infinity) result.
As well as being incorrect, the above rule was also rarely
practical because it was necessary to know the distance
between the front boundary A and the rear boundary B, to
determine the position of the focal plane P from this and
also to have an object to focus on at this point.
This is the correct focusing method: There are always
objects for focusing at the front boundary A and at the rear
boundary B of the depth of field (otherwise the boundaries
A and B would not be exactly there). First A is focused on
and then B or vice versa.
1. With cameras on an optical bench and with focusing
by moving the back standard, both focus positions can be
marked or read off a scale and noted as numerical values,
e.g. 37 and 40. The best focus is at the center between
these two values; i.e. in this example with 37 and 40, at
the scale value 38.5.
Focusing after a Scheimpflug lens tilt: The depth of field
has the shape of a wedge which tilts on focusing onto the
edge W of the wedge (see figure on page 3). For focusing on
A and B, a respective object point must be taken which
contacts these planes (in the bottom figure on page 3 and
on page 5 e.g. the head at the top and a shoe of the woman
at the bottom) and subsequently the focus must be at the
center between them. The correct aperture can be
read off the depth of field scale of focusable lenses
(the distance values there are, however, no longer correct
after the Scheimpflug lens tilt and only serve as aids for the
respective focus at the near point and at the far point) or by
using a slide rule or using a depth of field scale.
Depth of field
The depth of field is then only 1/3 in front of the focusing
plane P and 2/3 behind it when the rear depth of field
boundary B is twice as far away as the front boundary A.
Depth of field
5/11 6/11
If the depth of field is smaller than the distance a from the
front depth of field boundary A, the depth of field behind the
focusing plane P is only slightly larger than before it.
Depth of field
If the depth of field is larger than the distance a from the
front depth of field boundary A, the depth of field behind the
focusing plane P is much or very much larger than before it.
2. With cameras with a focusing at the lens, both focus
positions can be read off from the distance scale of the lens
after the focusing to A and to B. The correct focus here is
also the one at the center between the two focus positions (not to the average of the given distance values).
The correct aperture can be located using tables, using a
slide rule or using a depth of field scale.
Depth of field
The depth of field then reaches exactly to infinity when the
focusing plane P is exactly twice as far away (2a) as the
distance a from the front depth of field boundary A.
a < b but A
The depth of field a before the focusing plane P is also always smaller than the depth of field behind the focusing
plane P after a Scheimpflug lens tilt, with the ratio a:b likewise varying. However, the depth of field wedge splits into
two angles of approximately the same values a and b.
In what order and how are the individual settings to be made?
If converging lines have to be straightened and a Scheimpflug lens tilt becomes necessary as well for “expanded”
depth of field, it is first necessary to correct the distortion and then to carry out the tilt. The following order of
the worksteps allows the fastest and most reliable reaching
of the best result:
1. Fixing the ideal camera position
How your motif will appear on the image, which foreground
will hide which background, how the size ratio between the
foreground and the background will be and many many more
aspects depend primarily on the location of the camera.
2. Arrange the objects in wedge shape if possible
If you are setting up the motif yourself (e.g. product shots),
try to arrange all the higher objects in the rear so that they
lie within an imaginary depth of field wedge to allow a
Scheimpflug lens tilt for less stopping down (e.g. for a larger
aperture) with minimal diffraction blur.
3. Setting up the camera with a vertical image plane
If “converging lines” need to be avoided, the image plane of
the camera must ran parallel to the perpendicular edges of
the motif even if the taking direction has to run obliquely
upwardly or downwardly at the end.
4. Carrying out the parallel shift
Any slanted taking direction which may be required is now
established by parallel shift of the back. If the camera (e.g.
an SLR camera) only allows a parallel shift of the lens, it
may subsequently be necessary for close-ups or with a very
close foreground to lower or raise the entire camera accordingly to restore the original lens position and perspective.
5. If desired: Setting the residual perspective
If converging lines should not be completely straightened,
but rather some “residual perspective” should remain for a
natural effect, the parallel shift initially carried out for the
slanted taking direction is reduced by about 1/4 and then the
entire camera is tilted so far that the desire picture section
is again present. The slight slant which is thus caused generates the residual perspective.
6. Focusing and determining depth of field
If the camera does not allow any Scheimpflug lens tilt or if
the motif does not require or permit any Scheimpflug lens
tilt, it is now necessary to focus and stop down so that the
depth of field includes the total motif to be imaged as sharp.
For this purpose, one respective motif detail has to be defined as the near point N in the front plane A and as a far
point F in the rear plane B between which everything should
be sharp and outside of which blur is permissible.
As we have seen, focusing on the end of the first third between the near point N and the far point F is incorrect. The
correct procedure is to focus on both after one another and
then to set the center between both focusing positions. The
required aperture is determined from the focus difference
(between the focusing positions for N and F) using a table
or a slide rule or read off from the depth of field scale with
cameras having a helical focusing facility at the lens.
7. Carrying out the Scheimpflug lens tilt
If the spatial depth of the motif to be imaged as sharp is
considerably larger than its height and if you can imagine a
depth of field wedge surrounding the whole motif, you can
use the Scheimpflug lens tilt. Image (as in the image below)
a plane A’ lying loosely on the motif to be imaged as sharp.
It contacts a respective “highest” object H 1 and H 2 respectively closer and farther away. Extend the plane A’ to the
intersection line W‘ with the base B‘. The plane A’ and the
plane B‘ (base) now form our provisional depth of field
wedge (which is a little bit different from the final depth of
field wegde with W below the camera as shown on page 3).
The angle bisector of this wedge indicates how the focusing
plane P should run. Since you do not know the angle by
which the lens has to be tilted for this purpose, simply tilt it
by a small estimated angle, e.g. by 2°. Focus on the detail X
in the plane P near the front in the motif at half the height
of H 1. Then check where H 2 is the sharpest: If that is also at
half the height Y, you have completed the tilting. If the best
sharpness is higher, increase the tilt angle, e.g. to 3°. If it is
lower, reduce the angle, e.g. to 1.5° or 1°. Refocus on X and
check the sharpness again at H 2. Repeat until you have the
best sharpness in each case at half the height at H 1 and H 2.
8. Final focusing after the tilting
Focus to the sharpness as described under point 6 but so
that the near point N is the highest point of H 2 in the plane
A’ and the far point F is a point at the bottom in the plane
B‘. Please follow the instructions at the bottom of the left
text column of page 4
How to find the lens tilt angle and how to focus for best depth of field
Provisional wedge
of depth of field
Parallel shift and Scheimpflug lens tilt with technical cameras
We are looking for a technical camera with which converging lines can be corrected and where the depth of field can
be adapted to the motif extent by a Scheimpflug lens tilt.
Such a camera must offer devices for the parallel shift or
tilting of either the back or of the lens or of both. Which
design solution is the best for practice?
However, the tilt of the camera causes the plane P previously adapted to the motif by the tilting of the back to tilt by
the same angle (blue arrows). Only a subsequent focusing
to a shorter distance (= larger distance between the image
plane and the lens, see second image on page 3) allows the
plane P again to pivot back into the dotted starting position.
The parallel shift is intended to shift the image section
transversely to the taking direction so that a slanted taking
direction results without tilting the image plane. Generally,
this is possible with the back or the lens. So that the image
sharpness is not changed, the back or the lens has to be
shifted parallel to the image plane.
The back shift and the lens tilt are therefore generally
the best conditions for simple, fast working without subsequent correction of the picture detail, image sharpness
and camera position. The principle of a good technical camera is then as follows:
The back shift is better because the position of the lens
does not change, that is there is no vertical offset between
the foreground and the background in the near zone which
may require a subsequent vertical change of the camera (by
a column shift at the tripod) for retaining the perspective.
The lens tilt is better because the image plane aligned
perpendicular (or almost perpendicular in the case of the
residual perspective) for the correction of converging lines
remains unchanged. A further advantage is that on a subsequent correction of the image framing after the lens tilt, a
further parallel shift does not change the image sharpness
because the image remains in the same plane.
If the back has to be tilted because the camera does not
allow a lens tilt for design reasons, the image plane aligned
perpendicular (or almost perpendicular for residual perspective) initially to avoid converging lines is moved to a slanted
position: Converging lines would again appear - even in the
false direction with shots obliquely from above! For this
reason, the camera has to be tilted using the tilting head of
the tripod after the back tilt so that the image plane is again
perpendicular (or almost perpendicular).
Very good
Tilted lens
Tilted back
for a vertical
A further solution would be the parallel shift and tilt
possibility at the front at the lens (as with shift lenses of
35 mm SLR cameras). In this case, the parallel shift should
be on the camera side and the pivot possibility in front of it
because a further parallel shift taking place after the lens
tilt to recompose the subject within the frame will only then
not change the image sharpness.
If the tilt range of the lens is strictly limited due to a
lack of space, e.g. to 4°, a wedge-shaped intermediate
piece with a wedge angle (as large as the maximum tilt
angle) could be inserted between the camera housing and
the lens or between the camera housing and the back
which doubles the pivot range. If it can only be inserted
between the housing and the back, the camera must thus be
aligned before all other camera adjustments to avoid converging lines so that the picture plane is perpendicular.
However, then (as already described for the back tilting), the
sharpness will change and refocusing will become necessary on any subsequent parallel shift which may become
necessary if you want to recompose the subject.
Sports car © Can Stock Photo Inc. / XupaArts · Woman © Can Stock Photo Inc. / vujkekv · All other drawings © Walter E. Schön
The Scheimpflug tilt by a tilting of the lens plane L toward
the image plane I (in the back), or vice versa, should form
the depth of field otherwise bounded by two parallel planes
into a slanted wedge shape which can be better adapted to
the shape and extent of the motif. - online since 1996
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF