Friday, August 29, 2008

Adventures of an International Film Traveler

This blog is based on an article that I wrote for the Stereogram (13.1, with the same title).


<- If this donkey could talk, what would she say? Most probably she would ask about the weird two camera jig I am holding in my hands. [Photo taken by my son with my S1 -in Santorini]

As a kid, growing up in Greece, I did not like traveling or anything that would disturb my daily routine. This changed drastically when I discovered stereo photography. Now I look forward to traveling just because it is an opportunity for stereo photography.

What to Pack?

Packing stereo photo gear is a fine balance between taking all the equipment that I might need and keeping my camera bag small and light enough to carry around.

This year I decided to take the following:

1. RBT S1 and flash. This is my favorite all-around stereo camera.

2. RBT X2 with a Tokina 28-80mm zoom lenses as a backup to the S1 and also for its zoom lenses. I bought this camera from ebay and have only ran a test roll. It is a type C configuration (65mm lens spacing, only 13 pairs per roll but full frame). At it turns out, the lenses are well matched and very sharp. I am pleased overall.

3. Horseman 3D. This is a perfect portrait camera and took it to take portraits of my nieces, but I barely shot a roll through the camera. If I had to do this again, I would leave this home.

4. Pentax ZX-M cameras & twin camera bar that allows separation from 5.5 inches (cameras touching) to 14 inches. It has been a while since I traveled with a twin camera rig, and I am glad I did. I took some nice hypers with this set (and also with just one camera). I took two pairs of lenses, 40mm & 20mm, both very compact. The 40mm lenses saw the most use.

5. Misc accessories, including a compact Metz flash. I managed to fit all the equipment in a medium-size (some people would call this large) camera bag but also packed a smaller bag for lighter traveling within Greece.

In Greece we spent 3 days in Athens and then a week in Santorini (a popular Greek island) and then back to Athens for 4 days. While I was in my 4th day in Santorini my RBT S1 stop advancing (I had problems before I left, so this was not totally unexpected). Traveling lighter, I had not brought the RBT X2 so I ended up using only the Pentax SLR cameras for hyperstereos in the last few days in Santorini. I had the X2 back in Athens.

My 3d cameras attracted some attention from fellow tourists but no unwanted attention from security personnel in airports, museums, etc.


Slide Mounting & Airport Security

I like to utilize my travel time mounting stereo slides. My mounting procedure consists of two steps: First I cut and mount all the pictures in Albion aluminum mounts for preview. I then select the best and remount them permanently in RBT mounts. For this trip I decided to do the 2nd part and transfer the film chips from the Albion to RBT mounts.

For this I needed RBT mounts, a light table, viewer (Realist red button), scissors (to cut the edges of the full-frame film chips), gloves, markers, and related supplies. Before I left I checked the security information on line (http://www.tsa.gov/) and found that I can carry scissors, as long as they are shorter than 4 inches, which is fine. So everything went well. I was able to lay out my 3d mounting equipment/supplies and do a lot of mounting in a very small space, like the airplane food tray. Interestingly, no one asked me what I was doing or gave me any strange looks.

In the return trip (Greece to USA) the scissors were not allowed. So there was less slide mounting in the way back. Realist photographers might not need the scissors but for anyone mounting full frame slides they are important because the film chips often extend outside the mount and need to be trimmed.

While mounting en flight, I made an interesting observation: The low humidity environment in the plane results in a strong curling of the film chips (towards the emulsion side). This curling was so bad that I had to press the film chips with my fingers to keep them straight. Once we got back to the airport, the curling disappeared. Clearly this is the result of the low humidity.

Film Inspection & Processing

Do x-rays harm the film? Through my travels, I have not seen any clear signs of x-ray damage to the film. Every photographer has the right to request a hand-search for the film. The security personnel will immediately tell you that x-rays are safe for film speeds of 800 ASA or slower and try to change your mind and let the film be x-rayed. Some people carry high speed dummy rolls, mixed with their standard rolls, to counter this “slow film is OK” argument. In any case, it is your right to request hand-search. This will delay you (and maybe others around you) but it can be done. In my 2006 Greece trip I had all 50 or so rolls hand-searched. For every roll they touched the film leader to a machine looking for explosives. It took a while but it was done.

For this trip I brought with me about 35 rolls of film. I had more film waiting for me in Athens (had left it in storage in 2006). I let my film be x-rayed in the way to Greece just to avoid the hassle of hand-searching, but I asked to be hand-searched when traveling back. No problem with Greek security with this request. But some of my rolls were x-rayed 3 times (to Greece, to Santorini, and again in the USA from Atlanta to Cleveland).

In order to avoid carrying exposed film in the way back, I was planning to have all my film processed in Greece. I had done this in the past with good results. Price for slide film developing in Athens is half the price in the USA and the quality is equally good. I figured that despite the closing of film developing places in the USA, it would be "business as usual" in Athens. Well, I was wrong. The professional lab that I took my film in the past, was closed and out of business. So, I had to bring my film back to the USA.

I came back with 42 rolls of film (plus 3 more rolls in the cameras, that I finished shortly) which I have already processed and mounted. This might sound like a lot of film for Realist photographers but have in mind that I am getting about half the amount of pictures in a roll of film with my RBT cameras than a Realist (only 13 pairs from the X2 and 15 from the S1, vs. 29 pairs from a Realist).

The Bottom Line

Overall, other than the lab closing in Athens, and the malfunction of the S1, things went OK, and I have some very nice pictures from this trip to Greece. I am especially happy with the twin and single camera hyperstereos (from the airplane or hand-held shooting). If I had to do it again, I would have taken my “other” S1, left the Horseman back, and I would have taken more film with me to Santorini. I plan to return to Greece in 2009 to run the “Classic” Marathon.

RBT Film Formats

To understand the dilemma that RBT (or any person/company attempting to design a 35mm stereo camera) faces, consider the size of the full-frame 35mm film format: 24x36mm. The simplest solution for film economy, while maintaining full frame, is to have a full frame pair, separated by a full frame. This design will lead to the sequence: ABAB/CDCD/, etc. (the film can be cut every two pairs without splitting a pair) The only problem is that the minimum separation of a pair is 2 x 36mm = 72mm. That would have been the case if the pictures touched each other. With a bit of a margin between each pair, we arrive to the 75mm image (and lens) separation

This is configuration B, possibly the most popular configuration. The configuration leads to alternating advance sequence (1-3-1-3, etc). One advantage is that the roll looks similar to a standard roll exposed from a 2d camera, so the slides can be handled and mounted by automated machines. The disadvantage is that the lens spacing (75mm) is wider than the spacing of the eyes (which is around 65mm average).

If one does not like this 75mm wide separation of the lenses (dictated by the separation of the images on the film) then they have two options:

1) Reduce the image size. Keep the same basic format (ABAB/CDCD/, etc, alternating advance sequence) but make the size of each picture slightly smaller, 33mm instead of 36mm. By cutting 3mm from each picture, we can reduce the spacing of the images and lenses by about 9mm. The result is a 65mm lens spacing (desirable) but the image size is smaller and the roll cannot be processed by automated machines. This is configuration A.

2) Leave a gap between each pair (i.e. have no images recorded between two pairs). In this case we can keep any spacing of images (and lenses) at the expense of film economy. This is configuration C. This configuration only gives 13 pairs per 36 exp roll (vs. 20 for configuration A).



Which format is best?

It depends. Three questions to ask: 1) Is orthostereo (65mm) important to you? 2) Is automated processing important to you? 3) Is having full frame images (36mm wide) important to you?

Format B gives up orthostereo for automatic processing. If you want automatic processing, then B is the only choice, at the expense of orthostereo. If you want orthostereo but don’t care for automatic processing (because you are doing your own stereo slide mounting, for example) then go for A (film economy) or C (full frame size).

How important is the difference between 65mm and 75mm?

Some people will say that this difference is small. Some people will actually prefer the extra “depth boost” of the 75mm spacing. Those coming from the Stereo Realist format (70mm) should have no problem adjusting to 75mm. But, for me, coming from the gentle depth of the RBT S1 (59mm), the 75mm is too much. I do not like indoor shots taken with 75mm lens spacing. So I want 65mm. I also don’t care for full frame and I prefer film economy. The format of choice for me is A.

The RBT S1 camera follows a different format. For some reason, RBT designed the RBT S1 with a 59mm spacing. This leads to exactly half frame gap between each stereo pair. The film advance is fixed (20 perforations). Maybe this choice was dictated by the details of the film advance modification when joining the cameras. The RBT S1b uses a 45mm lens spacing but apparently the same advance sequence (and film waste) as the S1. This reinforces the idea that the details of the camera advance modification have led to this unsual format. I’ve heard that the RBT S1b was made especially for underwater stereo photography. Some people think that the 59mm spacing of lenses in the S1 is already too small, but personally I have grown accustomed to it.

RBT Stereo Cameras

The RBT Company in Germany has been joining 2d cameras to create 3d stereo cameras for quite a few years now. They probably started inspired by the work of Fritz Ochotta, linking Yashica cameras. Their web site is: http://www.rbt-3d.de/. The importer of RBT products in the USA is 3D Concepts: www.stereoscopy.com/3d-concepts/. Here is a list of the RBT cameras I am aware of:

A. RBT SLR cameras – Usually prefixed by “X”

RBT 108: Yashica 108 bodies, Yashica/Contax (Y/C) mount
RBT 109: Yashica 109 bodies, Y/C mount
RBT X2 – Ricoh KR 10M bodies, Pentax K mount
RBT X2II: Ricoh XR-X3000 bodies, Pentax K mount
RBT X3 – Ricoh XR-X3PF bodies, Pentax K mount
RBT X4 – Cosina Cs1 bodies, Pentax K mount
RBT X5 – Cosina FM10 bodies, Nikon F mount

B. RBT Rangefinder cameras – Prefixed by “S”

RBT S1 – Konica Hexar bodies, fixed 35mm f2.0 lenses, 59mm lens spacing
RBT S1b – Like above with only 45m spacing of lenses
RBT S2 - Voigtlander Bessa R2 bodies, Leica M mount
RBT S2a - Voigtlander Bessa R2a bodies, Leica M mount
RBT S2b - Voigtlander Bessa R2b bodies, Leica M mount
RBT S3a – Voigtlander Bessa R3a bodies, Leica M mount
RBT S3 – Zeiss Icon bodies, Leica M mount

From what I gather from the RBT web site, only the 109, X2II, X4, X5 and S2, S2b, S3, are currently produced. The 3d-Concepts web site lists as S3 a Voigtlander camera with Bessa R3 bodies. I am calling S3 the Zeiss Icon camera, following the RBT web site.

Most RBT cameras come in one of three configurations: A, B, C. More about these configurations in the following blog.

I have personally owned the following cameras: 108, X2II, X3, X4, S1, and now S3. The RBT S1 remains my favorite stereo camera.

Friday, March 7, 2008

Stereo Drawings



Back in 1988-1990 I wrote a program in “C” (that’s a programming language), to create stereoscopic drawings. It worked like this: The program will read a data file with the three coordinates of an object in space (X, Y, Z) plus simple instructions for plotting them, and then it would create a stereo pair. A fancy (for 1990 computer technology) interface allowed the user to rotate the image in space, while viewing it in 3d.

This was a nice learning experience for me and took quite a bit of time. See here some examples of the drawings I produced (click to enlarge) and also of the keyboard layout that allowed me to rotate, translate (even in the z-direction, i.e. affecting the stereo window), magnify the image, and also change the two important variables, perspective and parallax.

I am amazed that I did this work 18 years ago and then abandoned it. I am now feeling an urge to resume this, but what tools to use today? The reason I want to resume this work is that I want to study some interesting situations, like “keystoning” using simple drawings instead of actually trying to take pictures.



The program did a lot of things, like scaling, centering, rotating, etc., but the most important operation for 3d purposes is the projection of the point on the “plane of the screen”. Here is how this goes (see figure on the left):

We start with a point in space with coordinates (Xo, Yo, Zo). The observer is sitting at the point Z in the z-axis. If from this point Z we project this point to the projection plane (x, y), we get these projection coordinates: (Xc, Yc). We then move the observer to the Right and Left by B/2 each time (B is the stereo base) and project. Only the x-projections change. In the end, we get two pairs of points (XR, YR=Yc), (XL, YL=Yc). These are the coordinates of our starting point in the right and left projections.



I soon realized that there are two variables involved: One is the stereo base (B), which controls the stereoscopic deviation in the pair. This is clear. The other variable is not so obvious: For a given object, it is possible to come closer and photograph it with a short FL lens, or stay back and photograph it with the long FL lens. The difference between these two situations is the perspective. So I decided to use two variables in my program:

- PAX (Parallax) defined as B/D
- PER (Perspective) defined as S/D

B is the Stereo Base, S is the size of the object, D is the distance of the “camera” from the object.

The Parallax value is what I have called in previous postings “convergence ratio”, B/I in my current terminology. The Perspective variable is inversely proportional to the distance. It is related to the angle of view that the lens sees, which is also related to the size of the film and inversely related to the focal length. We can call this variable the “field of view ratio”.

It is interesting that I note as reasonable values 0.3 for PER (a “normal lens” has a ratio closer to 1. A ratio of 0.3 corresponds to lens of focal length between 80 and 120mm for 35mm film(depending on how you define the film angle of view, horizontal, vertical, or diagonal – the diagonal dimension is normally considered, but in stereo the horizontal is more appropriate, but should this be 36mm or maybe 28mm-7p or 21mm-5p?). So my choice of 0.3 is definitely conservative, meaning that it leads to milder perspective changes.



I also note that 0.06 is a good convergence ratio. Written as a ratio, this is 1/17, which is stronger than the 1/30 rule of thumb for normal stereo photography, but more in line with close-up/macro stereo photography.

It is interesting to study the formulas that give the coordinates of the projected points. There is a variable (which I called M because it looks like a magnification factor since it affects both variables the same degree.) This variable is not constant but it depends on distance (z) only. This expresses the perspective. Object points close to the camera, project further away, i.e. they look larger. The true stereo part of the coordinates depends also on the distance. As expected, objects closer to the camera have more depth. As we have discussed, distance affects all three metric aspects of the stereo image: Size, perspective, and depth.

Two interesting situations: If we set PAX = 0 (B = 0), there is no stereo! The two projections will be identical. The 2d images will have perspective but it will be flat. If we set PER = 0 we can still create a stereo pair, but there will be no perspective. This is known as “orthogonal projection”. Now, to have zero perspective we must be far away from the object, but then how do we have depth? By increasing the stereo base. The bottom line is that by changing our two variables we can create stereoscopic drawings that have various amount of depth (stereoscopic deviation) and perspective.

Infinite Road Problem

--- Background:

This is something that has been in my mind for a while, since we discussed the effects of changing the stereo base and focal length, and their interaction.

The fundamental stereoscopic formula (P = FB/I, P: parallax, or stereoscopic deviation, B: Stereo base, F: Focal length, I: near object distance, assuming far object is at infinity) clearly shows that increasing the focal length increases the stereoscopic deviation *for the same near object distance*.

In practice however, the near distance does not stay the same as we "zoom into the scene", but it is pushed back. I have watched Jay in our stereo club, zoom into a scene using digital stereo projection. This zooming is equivalent to increasing the focal length of the recording lens (yes, it is!) and as he zooms into the scene, it appears that the magnified scene is perfectly balanced depth-wise. So, it appears that zooming into a random scene is not a problem and stereoscopic deviation is under control. While it is difficult to analyze a random scene, it is easy to analyze this “Infinite Road” situation.

--- The Problem:

I set my camera on a tripod, at height H from the ground, in front of a road which is flat and extends far away. The only near point to the camera is this road. When I use a wide angle lens the near point is at I1 (see figure at left) If I switch to a longer focal length lens, the near point is now I2. The far point is at infinity. Now, by increasing the focal length I am increasing the stereoscopic deviations according to P = FB/I, but I am also increasing I, which is the nearest point in the road seen by the camera. Question: What happens to the stereoscopic deviation? Does it increase, decrease or remain the same?

--- Solution:

As we change the focal length, we have in front of the camera a triangle with H as its height, and I as its length (angle φ is the half of the angle of view). The schematic here shows what is happening in front and behind this triangle (this schematic is upside-down, compared to the previous one). From similar triangles we have H/I = h'/F. h' is the image of H in the film, or half the image height. Rearranging we get: h'/H = F/I. The stereoscopic formula is P = FB/I. Substituting the ratio B/I, we get:

P = (h'/H) B (1)

This is interesting: The stereoscopic deviation is independent of the focal length! Since H and h' are constants, P depends on B only, not F. Conclusion: The stereoscopic deviation remains the same. It is does not change with focal length.

--- Discussion:

This is actually a “constant magnification problem in disguise. The ratio H/h' is actually the magnification (H is the object photographed and h' is its image in film). We already know that P = M B (this formula is actually more general than the stereoscopic deviation formula P = FB/I) and we know that if the magnification is the same, then the stereoscopic deviation only depends on B, not F. But, there is no actual object of height H in front of the camera! H is the height of the camera above the road. What magnification am I talking about? It does not really matter that there is no object of height H. The way the geometry works, as we change the focal length, it is as if there is always an (imaginary) object of height H at distance I from the camera that fills the frame. Think about it!

--- An application:

Let’s say that I am using my RBT camera (B = 75mm) and I want to achieve 1.2mm total deviation on film. How high should I raise my camera? Let’s plug some numbers to (1): P = 1.2mm, h' = 12mm (half the film height), B = 75mm. Solving for H = (h’/P) B = 12x75/1.2 mm = 750mm (about 30 inches). So if I raise my tripod by 30 inches (3/4m) I will get good depth, no matter what lenses I use.

--- Implications:

It is easy to see that instead of a road we can have any type of surface (a tunnel, etc) that follows simple perspective geometry as it recedes from the observer and the conclusion will be the same, i.e. stereoscopic deviation is independent of focal length as long as the near object is this ground and not a tree or something else. We have also seen that the same conclusion applies when there is an object like a tree or a person or animal and we are then *moving* so that this object is framed to “fill the frame” (constant magnifications). So there are more than one situations where we arrive at the same conclusion. Maybe having a camera with fixed stereo base and variable focal length (like an RBT camera with fixed lenses) is not so bad after all!

--- An extension:

What happens if the ground is not level but slopes up or down (uphill/downhill, see picture). In this case, I worked out the math as:

P = (h’/H) B tan (φ-θ)/tanφ, where θ is the slope, and tanφ = h'/F

Notice that if θ = 0 then we get (1). The sign of θ is important (positive is slope down in my formula). This result shows that P now depends on the focal length. If the ground is sloping down, then the deviation decreases as we increase F. Note that if φ = θ, then the camera never sees the ground because of this slope (the lower filed of view runs parallel to the road) so P = 0 (no near object, only infinity, thus no depth!) If the slope is negative (uphill) then the deviation increases with focal length. Finally, the same result applies if the camera is tilted by θ with respect to the ground (ground flat, camera tilted).

Monday, February 25, 2008

Long FL stereo photography - PePax Principle


To neutralize the compression resulting from the use of longer lenses we can increase the stereo base. This leads us to the “PePax Principle”, advocated by H. C. McKay in the '50s. According to this principle, if you increase the focal length of the recording lens you should increase the stereo base proportionally to get a picture that resembles ortho stereo.

This technique is used for wild animal photography and assumes two things: 1) The near object is far away from the camera, and this is the reason for using long FL lenses. 2) The background is limited (there is no infinity and the depth range of the scene is restricted). These two conditions ensure that the stereoscopic deviation remains under control. (As we showed in the previous posting, under constant magnification and narrow depth range, p ~ B/F, so doubling B and F results in the same stereoscopic deviation.)

Two examples are shown here: The bear was photographed in the Cleveland zoo using 135mm lenses. At first, I took the picture with the camera side-to-side (6 inches). I was not happy with the depth compression so I tried increasing the stereo base to follow the PePax principle (10 inches). The results looked better. When I entered this picture in Detroit, someone asked me how could I come so close to the bear with my Realist, an indication that the picture looks like it was photographed closer with shorter FL lenses.


The three bowling pins have been photographed and are shown so that the front pin has the same size (constant magnification). The depth range is also limited. The middle picture is taken at F, B, I. For the top picture the camera is moved twice away from the subject (2I) and the focal length is doubled (2F) but the stereo base remains the same. This picture (2F, B, 2I) shows depth compression, compared to the (F, B, I). In the bottom picture, we have doubled the stereo base (2F, 2B, 2I). This bottom picture compares well with the original one.

So, What's the Catch?

By increasing F and B proportionally we make further objects, photographed with a longer lens, look as if they were photographed closer with a normal lens. This sounds too good to be true. As a matter of fact, it is too good to be true! To understand how the resulting pair differs from the “ortho pair”, compare the top and bottom figures of the bowling pins. Even though the depth is increased and the compression seen at the top is partially eliminated, the bottom picture is still different from the middle (which was recorded from closer). The difference is in the perspective, or the relative sizes of objects at different distances from the camera which has not changed. As a result, distant objects appear larger than nearby objects or nearby objects appear smaller than distant objects. This is direct result of trying to fool the brain into thinking that we are closer to our subject when we are really further away.

Effect of Focal Length on Stereoscopic Deviation

What happens to the stereoscopic deviation as we increase the focal length? We have mentioned that the focal length acts as a magnification factor and magnifies the stereoscopic deviation, per the formula: p = FB (1/Imin – 1/Imax) (1)

So, according to (1), if we double the focal length, the stereoscopic deviation will be doubled. This assumes that the distances of the near and far objects do not change as we change the focal length. Now, what are the chances of this happening? Pretty slim, I think. Unless if the near and far objects are in the line of sight near the center, as we zoom into the scene we will be moving past near objects, thus keeping the deviation under control. I first noticed this while watching zooming during digital stereo projector. As our projectionist was zooming into the scene, the range of depth changed and the deviation and sense of depth seemed well-balanced and under good control.

Constant Magnification

So far we have examined the effect of the focal length with a fixed distance from the subject. What happens if we change both the focal length and the distance to the subject so we can have the subject fill the frame? “Fill the frame” implies constant magnification. In this case we can either use a short FL lens and come close to the subject or use a long FL lens and stay far back. There are many situations (wild life photography, portraits, etc) where we use long FL lenses in order to stay further from the subject and still fill the frame. In the case of constant magnification, instead of using (1) we should go back to the original formula: p = M B (2)

This formula shows that if the magnification remains the same, the stereoscopic deviation is independent of the focal length. This formula assumes that there is infinity in the picture. But in many real situations not only there will not be any infinity, but the scene will have a rather narrow depth range.

Narrow Depth Range

From (1) we have: dp = F B dI / I**2 (dI is the depth range). Substituting M = F/I, we get:

dp = M**2 B (depth) / F (3)

This formula shows that for constant magnification and a narrow depth range, the stereoscopic deviation varies inversely to the focal length. So, if we want to maintain the same amount of deviation, while we are increasing the focal length, we need to increase the stereo base. This is the basis of the “PePax” principle.

Effect of Focal Length


In previous postings we examined the effect of the stereo base. Now we will see what happens if we change the focal length of the recording lens or the viewing lens (or viewing distance).

These effects are summarized in the Table here. Note that what matters for these effects is not the absolute value of the focal lengths of the individual lenses, but the relationship between the two.

If no lenses are used (as in the case of mirrors or projection) we can substitute the focal length of the viewing lens with the viewing distance. If the original film chips are magnified (as with prints) then the focal length of the recording lens must be multiplied by the magnification. A better way to treat this subject is to talk about angles. The condition for ortho stereo is that the scene is viewed from the same angle as it is recorded.

Why does this mismatch of recording/viewing affect the perception of depth?

One way to explain it is this: A longer focal length lens (which essentially magnifies the image) makes it look as if the camera was closer to the subject than it really was. When we view this picture we mentally compare it with the one recorded from closer. In this case, to be compatible with the reduced perspective (2d) and deviation (3d), the depth must be reduced. Hence the perceived depth compression.

Some people think that the focal length affects the perspective (relative size of near vs. far objects). This is not correct. Only the distance affects perspective. Let’s say that we record a scene with a wide angle lens, and then, without changing position, switch to a long FL lens and take another picture. If we then enlarge the picture from the wide angle lens to match the size of the objects in the picture from the long FL lens, the two pictures will be identical! The focal length acts only as a magnification factor and this is true both in 2d and 3d. What creates the compression/stretch “illusion” is the mismatch of the viewing distances. If both the wide angle and telephoto lens pictures are viewed from the same distance, then they will result in a different impression.

A Simple Experiment

Here is an experiment anyone can do right now: While viewing a stereo image (for example, freeviewing some of the images in this newsletter) move the image away and see how this affects the perceived depth. You should see the depth increase (“stretch”). By bringing the image closer, the depth should decrease (“squash”). The effect is rather subtle but most people notice it. Another experiment is to move back and forth during stereo projection. There appears to be more depth in the projected image when viewed from further away. Note that these effects are not perceived proportionally.

Seen Also in 2D

This effect is seen in both 2d and 3d images but it is more noticeable in stereo. In 2d it is the change in perspective with distance that creates this impression. Most of us are familiar with races in which, when filmed straight-on with long lenses, it appears that all the runners are in the same line and we are surprised to see how far apart they are when the camera changes angle of view. The reason the runners appear in the same line (depth compressed) is that their sizes are the same (zero perspective). That's the result of filming the race from far away and not the result of using long lenses, but the long lenses help get a larger image. To be more exact we should say that this compression is the result of viewing the image from much closer than it was recorded. We should always have in mind that it is not the recording lens or distance that creates the effect but the mismatch between the recording and viewing distances.

What is the BEST Stereo Base?

Given the freedom to achieve any stereo base (as it is the case when using a slide bar and a single camera) what is the correct stereo base to use for a given scene?

This has been the subject of some debate. My answer to this is that there is only one “correct” stereo base and this is Bv (B=Bv, equal to the spacing of the eyes). Anything else will result in an impression that alters reality, in which case there is no right or wrong.

If we decide however that we want to alter reality then there are formulas and rules of thumb which guide us into producing stereo pairs with a decent amount of depth (not too little or too much). There are two schools of thought: One advocates having a constant/maximum on-film deviation. It uses the basic stereoscopic formula, plugs the distances Inear, Imax, also F, and maximum-on-film-deviation, usually 1.2mm (for 35mm film) and calculates B.

I find this approach very artificial. The stereo base will change any time the distances of near/far objects change. Imagine that the spacing of our eyes changes as we move around, thus changing the distances of near/far objects. It is crazy!

The other school of thought advocates a constant convergence angle (expressed as ratio: B/I). One example is the well-known rule of thumb the “1/30 rule” which says that the stereo base should be equal to 1/30 the distance of the nearest object (B/I = 1/30). I prefer this approach for my stereo photography, but I understand that the convergnece angle can change, depending on the subject. For example, close-ups and macro photography generally requires a larger convergence (1/20 to 1/10). The reason I like the convergence approach is that it is easier to measure (divide the stereo base by the distance of the near object, or, multibly the ratio with the distance, to get the stereo base) and easier to visualize.

Is "More" Always "Better"?

Related to the this topic is the frequently asked question: “Wider stereo base means more depth. Stereo photography is about depth. So a wider base (and therefore more depth) is always better, right?”

I hope it is clear that the answer is “not necessarily”. More is not always better. Sometimes less can be better. The effect of putting more depth into the scene will result in the scene appearing smaller in size. This can lead to unusual and impressive images, like a “toy model” impression of a building or Grand Canyon. But many times making an object appear larger in size is equally, or more impressive. And many times just reproducing a scene in near-ortho (as seen by the eyes) is best. It all depends on the subject, application, and personal taste.

By all means experiment with different stereo bases but it would be a mistake to assume that more is always better!

Ortho Hyper Hypo Stereo



By considering the stereo base with respect to the interpupillary distance, we can talk about three special types of stereo images: ortho stereos, hyperstereos, and hypostereos, defined in the Table here.

Sunday, February 24, 2008

Space Control

We saw in the previous posting that changing the stereo base changes the distance of the subject from the observer. The amount of movementis proportional to the change of the stereo base. For example, doubling the stereo base will pull the scene twice as close. Reducing the stereo base by half will push the screen twice back.

This leads to the concept of space control. By controlling the distance of an object to the camera and the stereo base, it is possible to alter and control the apparent dimensions and distance of this object.

Those who have mastered space control have produced fascinating illusions by multiple exposures, like the ones shown by Tommy Thomas in the book “The Stereo Realist Manual”. Examples include the miniature girl inside the wine glass or the artist who is painting a picture of a live model (see left).

Here is how this illusion is produced: Start by taking a picture of the artist with the canvas covered by the dark cloth. Then proceed to double-expose the model in the area inside the dark cloth. To achieve this, the model must be photographed from a distance (to reduce the size) with increased stereo base (to be pulled closer). This requires very good planning and lots of patience and luck!

Digital photography can certainly make these kinds of effects easier to produce in the “digital darkroom”, but what do I know? :)

Stereo Base

In a previous posting we looked at the three variables that affect a stereo image, the stereo base (distance between lenses, B), the focal length of the lenses (F) and the distance from the subject. A conventional stereo camera has fixed lenses so the only way to affect the image is by changing the distance to the subject. Going beyond the stereo camera usually means changing B or F. From these two variables, the stereo base is perhaps the easiest variable to experiment with, since any camera (including a stereo camera) can be used in two successive exposures to record stereo pairs with any desired stereo base.


The effect of changing the stereo base are summarized in this statement: Increasing the stereo base increases the deviations, pulls the scene closer to the observer and makes the objects within the scene appear smaller. Reducing the stereo base decreases the deviations, pushes the scene away from the observer and makes the objects appear larger.

The effect on the deviations is pretty clear from the fundamental stereoscopic deviation: p = FB/I, but how increasing B makes the scene come closer or appear smaller? The figures on the left might will help clarify this.

Why Appear Closer?

Consider the series of pictures of the 3 bowling pins, taken with different stereo bases. These pictures are shown here exactly as recorded (with a slide bar) with no attempt to adjust the stereo window. It is clear that the image is shifting towards the inner edge of the frame with increasing base. From our discussion of the stereo window we know that when the film chips are pushed closer, the scene appears to be moving towards the observer. Stereoscopic viewing of the bowling pin pictures confirms that this is indeed the case. Increasing the stereo base does make the scene appear closer to the observer.

Why Appear Smaller?

This is a direct consequence of the previous effect. The actual size of the objects does not change but when the stereo base increases they appear to be closer, so our brain concludes that they must be smaller.

This is a little more subtle to observe and is more pronounced in stereo projection than freeviewing on this page. But let's try it: Consider the series of letters in the stereogram here. All letters have the same size. The relative shift of the letters (artificial deviation) makes the ones in the center (“Smaller”) appear closer to the observer. Hence they look smaller. The ones in the back (“Larger”) appear the furthest from the observer hence they look larger. It might take a little while for this to be clearly seen. Stereoscopic viewing is absolutely essential to experience this. If instead of parallel, you use cross-freeviewing then the opposite effect will be perceived, i.e. the “Larger” will be forward and smaller and the “Smaller” will be pushed the back and appear larger.

Stereo Photography Viewing Variables


The three "recording" variables F, B, and I, affect the way the stereo image is recorded on film but they also affect the way the stereo image is perceived, i.e. how it appears during stereoscopic observation.

To understand the stereoscopic impression when we view a stereo image, we also need to know the focal length of the viewing lens, Fv, and the interpupillary distance (eye spacing) of the observer, Bv. Fv and Bv are now our viewing variables.

Finally, even if we know the recording variables and the viewing variables, what we actually perceive also depends on our brain & experience, what we call "peception". So, we can say that:

3d image perceived = (recording variables) + (viewing variables) + (Perception)

There are two conditions that, when satisfied, viewing the stereo image most closely imitates viewing directly the original scene: 1) Stereo base is equal to the interpupillary spacing (B=Bv, approximately 65mm or 2.5") and 2) focal length of the recording lens is equal to the focal length of the viewing lens (or viewing distance), F=Fv. This is known as “ortho stereo”.

Ortho Stereo: B = Bv & F = Fv

General-use stereo cameras are well-suited for this type of stereo photography which explains the choice of lens separation in Realist-format cameras. The focal length of the recording lens is not important as long as it is matched by the viewing lens. Most 35mm film viewer lenses have a FL of 40-50mm. The 35mm FL lens in many stereo cameras is a compromise, offering good depth of field, decent field of view, and near-ortho viewing conditions.

Any deviation from these conditions will result in a visual impression that deviates from reality. We will explore some of these situations in subsequent postings.

Stereo Photography Recording Variables

This blog is based on my Tutorial “Beyond the Stereo Camera”. You can purchase the entire collection of my stereo Tutorials by going to: http://www.stereotutorials.com/

There are three variables which affect the way images are recorded on film:

1) Focal length (F) of recording lens.
2) Stereo base (B) of stereo system.
3) Distance (I) of the camera to the subject.

These three variables affect three “metric” (measurable) aspects of the recorded image:

1) On-film size of an object (or magnification).
2) Relative sizes of objects at different distances from the camera (this is also known as linear or geometric perspective).
3) Stereoscopic deviation.


These effects are summarized in the Table reproduced here. Note the formulas that express the relationship between the recording variables and the metric aspects of the recorded image:

  • Magnification: M = s’/s = f(I-f) ~ f/I, or on film size s’ = s f / I, only depends or object size, focal length and distance.
  • Perspective: ds/S = dI/I, only depends on subject distance. (ds is a change in image size due to a change in image distance dI)
  • Stereoscopic Deviation: p = FB/I, depends on F, B and I

    Some comments:
    • The focal length acts as a magnification factor. It magnifies the size of the recorded image without altering the perspective. It also increases the stereoscopic deviations.
    • The stereo base is the only variable unique to stereo photography and it only affects the stereoscopic deviations, which is the only metric aspect unique to stereo.
    • The distance of the camera to the subject, essentially the only variable available in a standard stereo camera, affects all three aspects of the recorded image. The effects are proportional to the inverse distance (1/I) which we can call “closeness to the subject”. By coming closer to the subject you 1) increase the on-film size of the subject, 2) intensify the perspective (make closer objects appear larger than further objects) and 3) increase the stereoscopic deviations. That's a good argument for getting closer!

    Basic Stereoscopic Equation


    Consider an object A at a distance I from the lenses of the stereo camera, which are separated by B (stereo base). An object at infinity is formed at O1 on the left side and at O2 at the right side, while the image of A is A1 and A2. The situation is symmetric so half the stereoscopic deviation (or parallax) is P/2. From similar triangles we have:

    B/2 / I = P/2 / I' or P/B = I'/I (1)

    From our previous posting we know that the ratio I'/I is the magnification M. So we get:

    P = M B (2)

    This is the basic stereoscopic equation. I cannot help but make the following analogy: Variables in the image space (with prime ') are related to variables in the object space through the magnification. For example, I' = M I, s' = M s, and here we have P = M B, so we can think of P as B', in other words, the stereoscopic deviation is the “image space” equivalent of the object space stereo base.

    If the subject is far away from the lens we can use the low magnification approximation and write (2) as follows:

    P = FB / I (3)

    Equation (3) gives the parallax with respect to infinity (remember, we measured P from point O which is a point at infinity). If we have a near object at Imin and a far object at Imax, then the stereoscopic deviation equation can be written more generally as:

    P = F B (I/min – Imax) (4)

    The stereoscopic deviation is proportional to the focal length, the stereo base, and inversely proportional to the distance.

    Basic Lens Equation


    I find myself using the basic lens equation quite a bit so I would like to derive some important formulas. Consider a lens of focal length f. The object is at distance I from the lens, while the image is formed at distance I'. The size of the object is s, the size of the image is s'. See the diagram here.

    The basic lens equation is: 1/f = 1/I + 1/I' (1)

    The magnification by definition is M = S'/S = I'/ I (2)

    If we use equation (2) to solve for either I or I' and substitute it in equation (1), we obtain these two useful formulas:

    M = f/x (3) and M = x'/f (4)

    From (3) and (4) we can write (1) as: f**2 = x x'

    If the subject is far away from the lens (low magnification) then I >> f and I = x, I' = f, so the magnification is approximately equal to M = f/I. This is the low magnification approximation.

    At high magnifications I gets close to f, and I' gets very large, so I' = x' and M = I'/f. this is the high magnification approximation.

    An interesting situation occurs at M = 1, then x = x' = f, and the subject is at distance 2f from the lens and the image is formed at distance 2f from the lens. In this case the total distance from the object to the film plane is the smallest possible (4f).

    Monday, February 11, 2008

    Close ups with the HORSEMAN 3D Camera




    The Hosreman 3D camera with its shorter stereo base (Spacing of lenses, B = 34mm) is well suited for close ups. As a matter of fact, I have only used it for close ups at the near focus (0.7m) and also with +1 and +2 lenses (at near focus again). For most indoor pictures I use a flash in auto mode (the camera does not have TTL flash) at f8. As I said in an earlier post, the results are just great! I am very pleased.

    Here are some comments/specifications of interest to me for close ups (for complete technical specs, see the info above – from the instruction manual, click to enlarge – and also the official Horseman 3D web page):



    • The Horseman is based on the Hasselblad Xpan II and it has been modified by installing the lens plate, containing two 38mm f/2.8 lenses. So F = 38mm. Minimum aperture is f16.

    • The spacing of the lenses (stereo base) is B = 34mm.

    • The filter threading is 62mm and covers both lenses. This makes it very convenient to use filters. You only need one 62mm filter and this will cover both lenses.

    • What is unusual about this camera is that it has only one (metal-blade focal plane) shutter, long enough (65mm) to cover both lenses. As you can see from the picture here, there is a bar that separates the two images. The bar has a notch in the bottom of the right side. This notch will show up in the top of the left image and helps differentiate the left from the right image.

    • I measured the size of the film gate openings as carefully as possible and concluded that each film gate opening is 29.5 x 24mm. The size of the bar at the center is 5mm. This makes the total width of the shutter area 29.5x2+5 = 64mm.

    • Measurements on the film show that each image is 30.5x34.5 mm. The separation between a stereo pair is 3.8mm and the separation from pair to pair is 1.7mm. Based on these measurements, the total distance from the left side of the left image to the right side of the right image is 30.5x2+3.8 = 64.8, while separation from stereo pair to stereo pair is 64.8+1.7 = 66.5. I measured the sprocket holes above the image and each image is about 7 sprocket holes wide, while from pair to pair we have exactly 14 sprocket holes. Given that spacing of sprocket holes is 4.75mm, 14x4.75 = 66.5mm. It appears that the camera advances 14 sprocket holes between each shot, using a sensor located in the sprocket hole path.

    • The camera gives 20 stereo pairs in a 36 exposure film. Another unusual feature is that when you load the film, the camera advances the entire film to the right side, and then it releases it back to the canister as you take picture after picture. One advantage of this system is that if you accidentally open the back, you will not ruin the pictures taken so far (because they are already inside the canister).


    There are two questions in my mind, based on the comments above:

    1. How is it possible that the film opening is 29.5mm wide, while the image is 30.5mm wide? I suspect that this is happening because the film is sitting a bit further than the film openings. The light comes from the lenses as a cone and the cone is being cut a bit higher than the film openings, thus a larger area is being cut.

    2. Does the camera have a built-in stereo window? From the measurements it appears that the spacing of the film gates is 34.5mm. Considering that the spacing of the lenses is 34mm, this will create a stereo window with 0.5mm parallax, which, for this camera, it places it at 2.5m (8 ft).

    The table here summarizes the important metrics of the Horseman 3D camera, when used for close ups, focused at the near distance, with or without close up filters. Because I am not sure about the built in window, I am using x as the “built in window parallax”. This could be 0.5mm as my measurements indicate, or it can be zero if there is no built-in window.

    To mount my Horseman 3D slides I use either 7p (28mm) mounts or 5p (21mm) mounts. I use the narrower 5p mounts for composition purposes because most portraits are framed better in square or vertical mounts. If you must use 7p mounts then you can tolerate 31.5-28 = 3.5mm of image loss. As you can see from this table, you can mount +1 shots in 7p mounts even without a built-in window, but for +2 shots you are forced to use 5p mounts.

    As I said earlier, I use the camera mostly for close ups. It nicely bridges the gap between a standard stereo camera and a macro stereo camera. Here is how I carry the camera with me: I stack the two close ups filters (+1 closer to the lenses, then +2) with a metal cover over the lens. This way I always have with me the two close up filters. If I want to use the +1 filter, I remove the cap and +2 filters together. I also carry a measuring tape, marked with the three near distances and I use this to quickly frame my subject. Even though the camera has a rangefinder window, this does not work when using close-up filters, and it is also faster to use a frame even for the near shots without filters. Finally, I carry a flash with f8 auto mode.

    Pictures (mainly portraits of people or shot of my kitty) work well with the camera focused at the near distance at f8 with a flash. Even though the convergence is 1/18, stronger than 1/30, the pictures look very natural. With the close up filters some care is needed to select the subject and it is better to use f16.

    Update (2/26/08): The camera certainly has a built-in stereo window and it appears to be around 0.5mm as I suspected when I wrote the original post. This helps a bit by reducing the image loss in close ups, but it is not of much practical value for me. But it is good to know that this is something that the designing engineers took into account.

    Sunday, February 10, 2008

    The Macrolist



    The Macrolist is a specialized macro stereo camera, a replica of the Macro Realist Stereo camera. It was designed and built by David Burder in England. David is a world-renowned stereo personality who has been involved in many stereo projects, including unique modifications of the Nimslo stereo camera. I estimate that 30-50 Macrolists were built. This makes this camera a rare stereo collectible, in addition to being a very practical camera to use for those interested in macro stereo photography.

    Technical Specifications:

    • The Macrolist is based on an original Stereo Realist body with a custom lens and shutter assembly.
    • Image size is standard 5p ("Stereo Realist format") or 23x24mm. Unlike the Realist-format, the images are side by side (not interlocked). You still get about 29 pairs from a 36 exposure roll of standard 35mm film. The film advance is the original film advance systems of the Stereo Realist camera.
    • The lenses are air spaced triplets, effective focal length 35mm, coated optical glass elements. Effective aperture is f/40 and it is fixed. This camera gives a greater depth of field than the original Macro Realist, which is considered an advantage.
    • Lens spacing (stereo base) is 16 mm. Field size is approximately 50-65mm and the magnification about 1:2.
    • The shutter is a Copal shutter with a full range of shutter speeds. Recommended exposure for the macro pictures is via electronic flash. Focusing is simply done by placing your subject between the two prongs of the camera.
    • Compared to other macro stereo systems, the Macrolist is rather “strong”, as you can see from the convergence of 9 degrees so some care is needed to keep stereo deviation in check.


    Close-ups with a Stereo Realist – Closer than 2.5 ft


    The Realist can focus as close as 2.5ft. If you want to focus closer then you have to either fool around with the focusing wheel or attach close-up (supplementary) lenses.

    The limit of 2.5 feet is imposed by the rotation of the focusing wheel, which is limited by a small screw. This screw can be removed (see picture) allowing the Realist to focus even closer. It has been reported that the “second round” of the focusing wheel corresponds to focusing distances from 20 inches (at INF) to 14 inches (at 2.5ft) but testing should be done to get the exact focusing distance. This testing can be done by using a ground glass and a magnifier and observing a test image.

    If you decide to experiment with this method make sure you don’t lose track of which round the wheel is at. Also, because the focus plane is pressuring the film, you should turn focusing to infinity before advancing the film.



    We discussed close-up lenses in previous postings. With a little imagination you can attach close up filters (or any filters) to your Stereo Realist camera. For example, I have used a film cap with a 20mm hole, with the filter attached to it, over the Realist lenses (see picture). Some people have used the Realist Film ID unit for macro pictures. This unit is essentially a strong close-up lens with the advantage that it attaches to the Realist lenses and offers a holder right at the near sharp focus distance. You must find some way to shift or rotate the subject for stereo relief and you need to move the unit from one lens to the other without disturbing the setup. Sounds like a challenge.



    Here is a Table that summarizes key variables when using the +1 and +2 close up lenses with the Realist focused at 2.5 ft.. In-between magnifications can be achieved by focusing the camera further. For example if you use the +1 lens and focus the camera at 10ft, then the actual focus point is at 30 inches (2.5ft) which is the same as the near focus without lenses.

    The drawback of these extreme close-up settings is realized when we look at the image loss. Using the +2 lens at 2.5ft focus we lose 8mm of image. This leaves us with 23.5mm – 8mm = 15.5mm of useable image width. There is no slide mount that can mask this (4p/Half Frame/Nimslo mounts have 16mm openings)



    One way around this problem is to use prismatic close-up lenses. These have the shape of a wedge and must be aligned over the lenses. In addition to changing the focus, they also shift the image to reduce/eliminate image loss. The drawback is that they introduce aberrations and distortions.

    I am familiar with the “Stereo Angle-Lens”, a prismatic close-up lens manufactured by Photo-Liz Inc in Long Beach NY. As seen from the copy of the instructions reproduced here, these filters come in 3 different strengths. Of interest is the “neutral” filter which is a simple prism. This is used to push the window back in close ups without filters or also to push the window forward when taking hyperstereos!

    I have tried these filters for an extreme close-up portrait of my daughter. The results were interesting, but the resulting image has too much stretch and distortion for my taste, even when viewed in the viewer.

    The bottom line is that for extreme close ups with the Stereo Realist, you need to reduce the stereo base. Some people have used the Realist and a slide bar for close ups. This procedure will reduce the stereo base in a stereo pair, without wasting any film: Cover the right lens, and take the first picture. Do not advance the film. Shift the camera slightly to the left (to reduce the spacing of the lenses). Cover the left lens and take the 2nd picture. I have seen a portrait taken with this method! (The model must stay still for along time). This method might have novelty value, but if you are going to use a slide bar, it is much more convenient to use an SLR camera, not the Stereo Realist.

    Close-ups with a Stereo Realist – 2.5 to 7 ft

    In the range of 2.5 to 7 ft, all you have to do is point the camera to your subject, focus and shoot. Sounds simple, however, if you want to do this successfully, you have to take certain precautions: First you have to limit your background. Second, you have to crop the final images for the correct stereo window.

    I see this in our stereo club. Some people take pictures with objects from 4 ft to infinity. This is too much for me. Such pictures cannot be viewed comfortably in projection. It is also impossible to mount them properly using standard stereo mounts. (Note: Some people attempt to mount such pictures using a technique called “double depth”. This technique eliminates “window violation” while keeping this excessive amount of depth. Personally, I prefer to keep the background a stereoscopic deviation within limits.)

    How do you block the background? By altering the composition. For example, put your subject closer to a wall or other natural barrier, use a flash (background is dark), shoot a picture of a dog or cat looking down, etc.)

    We can construct a Table which shows the maxium object distance for a certain near distance. I have done this in the previous posting for 4 focus distances. For example, if the near object is at 4 ft (1.2m), the far object should be no further than 10 ft (3m).



    As it turns out, you don’t need this Table because it is built into the Stereo Realist’s Depth of Field (DOF) scale. If you line up the near object distance with one of the f8 marks, the other f8 mark will show you the far object distance. For example, in the picture here one f8 mark is at 3ft (near distance). The other is at 5 ft (far distance). (Note: If your Stereo Realist does not have a DOF scale - earlier models did not come with one - I have these for sale in my web page)

    I credit Charles Piper for pointing this out. As it turns out, both the stereoscopic deviation and the depth of field have an inverse relationship to the distance. So we can use the DOF information to derive stereoscopic deviation information. There is nothing mystical about this. It is a coincidence that allows us to kill two birds with one stone.

    The f8 pointers correspond to 1.2mm, which is generally accepted at the maximum allowable on-film stereoscopic deviation. Charles Piper followed a more conservative approach and used the f5.6 pointers. These correspond to 0.8mm deviation and it is safer for stereo projection (in general, “less is better” when it comes to stereo projection).



    When it comes to stereo slide mounting, the image loss will often dictate what image size to use. The openings of the Realist film gate are 23.5mm wide. The 3 relevant RBT slide mount sizes are 21.5mm (5p), 19.5mm (5p close up), and 16mm (4p/Nimslo). So these sizes can accommodate an image loss of 2mm, 4mm and 7.5mm. Considering that film cutter is not always perfect, we should give a margin of 0.5mm, so the three slide mounts will cover an image loss of 1.5, 3.5, and 7mm. These correspond to a near object distance of 940mm (3ft), 556mm (1.8ft, 22in), 333mm (13in).

    Close-ups with a Stereo Realist - Theory

    The Stereo Realist is designed to take pictures with the near object at 7 feet (2.1m). For “normal” stereo photography, the following advice applies: “Do not let the near objects come closer than 7ft to the camera”. So, before taking a picture, make sure that there is nothing closer than 7ft to the camera.

    You can of course take pictures of objects closer than 7ft, and the camera will focus as close as 2.5ft. This is the area of close-up Realist photography. One advantage of the Stereo Realist for Close-ups is that the viewfinder is centrally located and parallax-free. So the framing is always accurate and you don’t need a focusing frame or other aids. You will appreciate this if you try close ups with other stereo cameras. One disadvantage is that distance of the lenses (Stereo Base, 70mm) is too wide for close-ups.

    There are two precautions not only for Realist but all close-up & macro photography: As you get closer and closer to your subject, you have to block distant objects. If you don’t, then your stereo pair will have too much deviation and it will be impossible to project and possibly difficult to view in the viewer.

    Also, you have to be prepared to make certain adjustments (cropping) in mounting to set the proper stereo window. For the slide film user, this means that you might have to use close-up or half-frame mounts, instead of standard (5p) realist-format mounts.

    Without any aid, the Stereo Realist camera will focus as close as 2.5 ft (0.76m), which is surprisingly close (the near focus of most 35mm lenses is 0.9m, or 1m). This near focus can be extended in two ways: The first way is by using a close-up (supplementary) lens. This is a universal way, and can be used with any camera/lens. A +1 lens will bring the near focus to 0.43m (1.4 ft). The second way is Realist-specific. You can remove the little screw that stops the focus wheel at 2.5ft and focus even closer. We will discuss these practical issues in the next posting.


    Here are the basic metrics of the Stereo Realist:
  • Focal Length, F = 35mm
  • Spacing of lenses (Stereo Base), B = 70mm
  • Near focusing distance, Inear = 2.5ft (0.76m)
  • Internal cropping = 1.2mm

    • This last item needs an explanation: The distance between the film gates is 71.2mm, which is 1.2 mm longer than the spacing of the lenses. This shift is done on purpose to block out a small part of the image (left edge of left image and right edge of right image, remember the images are reversed inside the camera) and set the stereo window at 7 ft.

      The Table below gives basic measurements for the Stereo Realist close up photography. I have elected to list the following near object distances: 1) 7 ft, that’s where close up photography starts, 2) 4 ft 3) 2.5 ft (near unaided focus), 4) 1.4 ft (focusing with the Realist at 2.5 ft and +1 close up filter).

      I have tabulated the following quantities of interest (all values are in mm):
      • Inear = near object and where the camera is focused.
      • Ifar = the recommended maximum distance of the far object, in order to keep the total stereoscopic deviation to 1.2mm.
      • Back Plane Extension, BPE: The Realist focus by moving the back plane away from the lenses. This value shows you how much the plane is moved back. We have called this quantity x' in our formula derivations here. Mostly a curiosity item.
      • Magnification M: This is expressed as a ratio, for example 1/21 at 2.5ft. This can show you the field of view. Considering that the height of the film gate is 25mm or about 1 inch, a magnification of 1/25 means that you can get 25 inches of your subject matter inside the picture. For comparison, a portrait usually requires 12 to 16 inches so the Realist is not really capable for a tight portrait even at the closest focusing distance.
      • Convergence: This is the ratio of the stereo base over the near distance, B/Inear. This ratio is 1/30 at 7ft. Values from 1/10 to 1/30 are generally acceptable.
      • Image Loss: This is the amount of film that needs to be cropped in order to fix the stereo window. It takes into account the internal cropping of 1.2mm so at 7ft there is no image loss. This quantity is of practical mounting value for slide film users.

      Limit to Stereo with Translation – Converge or not Converge?

      As the magnification increases, translation cannot be used effectively any more. Consider for example M = 5 (5 times magnification).

      I = 1.2F, I' = 6F, Cropping = 5B

      How much should I translate? My rule of thumb B = I/20. Assuming F = 100mm, B = 6mm. But cropping = 30mm. If we use 35mm film, all we have is 36mm to record the image. If we crop 30mm, we have very little image to work with. We need to work this the other way around. Let’s say we need to maintain half the film width. Then cropping = 18mm, B = 1.8mm. We can only translate by 3.6mm, not 6mm we originally had planned. That might not be enough and result to a flat image.

      In practical terms, we are forced to converge the lenses or use tilt instead of translation for magnifications of 10x or higher. I mentioned earlier that you should avoid convergence because it leads to keystone distortion. As it turns out, at high magnifications the working distance becomes equal to F and the extension equal to MF. Because of the large extension, perspective is minimized, which means that we can use tilt with no problem. (I will try to explain this “perspective is minimized” in a subsequent post.)

      The same is true of very long focal length lenses are used. Perspective is minimized and convergence is not only acceptable, but also the only practical way to record the image.

      Image Loss In Macro 3d Photography

      This is an important issue that can cause problems for a beginner. Every time you take a stereo picture by shifting a camera/lens parallel, the “stereo window” is placed at infinity. This creates two wide bands on the left side of the left image and the right side of the right image, which not only do not contribute anything but they also do not help to properly locate the image in space. To restore the proper location of the stereo window, some cropping is required.

      Consider the picture shown here. This is a close up of a 10 inch doll taken by shifting a digital camera. Can you see these wide bands? A rule for proper placement of the stereo window (3L rule) says that “The Left eye should see Less on the Left side”. In this case clearly there is more to see on the left side of the left image, not less. We need to crop out these bands. By doing this, the entire picture will now properly be behind the stereo window, as you see in the corrected pair.

      How do you get rid of these bands? It is easy to do it in digital photography or when making prints (cut off part the left side of the left print and the right side of the right print). But what do you do if shoot slide film? In this case you use the stereo mount to crop the edges, by shifting the film chips away from each other until the bands are hidden behind the mount.

      The amount of cropping needed depends on the stereo base (B) and the magnification (M). It is given by basic stereoscopic formula: Cropping = P = M B. See this blog for derivation.

      Consider some special cases: As infinity (M=0) no cropping is needed. At low magnifications (regular stereo camera distances) M = F/I, and the formula becomes Cropping = FB/I, where I is the distance of the subject. This is about 1.2mm for most stereo cameras.

      At M = 1 we have the interesting situation where a shift of B in the object space creates the same shift B in the image plane, so cropping = B. In a previous slide bar example we showed that for M=1, a good value for B is 5mm. The final image will require cropping by 5mm. So our 36mm length of 35mm film image area is now reduced to a useable 31mm length. This is important for slide film stereo photographers because it means that you cannot use 33mm or 31.5mm (all available by RBT) to mount this stereo pair and the next available size is 30mm. If you want to preserve more of the stereo image, you might want to consider less stereo base, just to reduce image loss. You might need to work this the other way around. Let's say that you want to mount your image in a 31.5mm mount. You cannot crop more than 4.5mm. To have some room in mounting, you put a limit of 4mm cropping. This corresponds to 4mm shift. So, instead of 5mm, you shift by 4mm only, the choice being dicated by cropping considerations only.

      Most stereo cameras from the '50s are constructed with a built-in stereo window. This is achieved by shifting the film gates with respect to the lenses. For example, the Realist lenses are separated by 70mm while the film gates are separated by 71.2mm. This shift creates a window at 7ft from the camera, which saves film and makes automatic mounting easier (by just centering the chips in the mount we get a window at 7ft.)

      Consider the schematic here (click at it to enlarge it). In the camera on the left, the film gates are centered under the lenses. The stereo window is at infinity and you will notice that at any distance from the camera the left eye sees more on the left side instead of less, which means that this point is in front of the stereo window. The correct stereo window can be set by trimming parts of the final image. The camera on the right creates a stereo window by separating the film gates wider than the lenses. The left eye now sees less in the left side for objects past the stereo window. In this case, if the film chips are centered, the window is set automatically and less film trimming is required for close-ups.

      Saturday, February 9, 2008

      Stereo Close-Up with a Slide Bar

      This is a great way to get started in macro stereo photography. For best results and convenience, I recommend using an SLR camera (with TTL flash capability if you plan to use a flash), a macro lens, and a slide bar. Here is why:

      1) An SLR camera allows you to see exactly what you are photographing. 2) Flash is often used in macro pictures and in close-ups it is difficult to determine flash settings. TTL capability means that the light output from the flash is measured through the lens, and this is an accurate and convenient way to use flash. 3) A macro lens might be a bit more expensive than ordinary lenses but it is built for optimum optical and mechanical performance in close ups. My personal preference is a 50mm macro lens. 4) A slide bar (on a tripod) keeps the camera well aligned between the two shots.

      The basic idea is to set up the shot, take the first picture, shift the camera, and take the second picture. As with all single-camera stereo photography, the subject must remain stationary between the two exposures. This limits the use of this technique, but there are still plenty of areas (tabletops, etc) to use it. All you need is some imagination. As an example of stereo photography with a slide bar and plenty of humor and creativity, consider Stan White’s View-Master set “Beyond the Third Dimension”.

      The most common question/concern during slide bar stereo is: “How much should I shift the camera?” This shift is what we usually call “stereo base” (B).

      One school of thought uses the formulas given in a previous post and aim to achieve a constant on-film-deviation (this is P, or Parallax or stereoscopic deviation). I personally find this solution very artificial because for a subject with lots of depth it calls for a small stereo base B, while for a flat subject it calls for a large and some times excessive stereo base. Why would B change with the depth range of the subject? There is no physical justification for this.

      Instead, I recommend looking at the “convergence angle”, that is, the ratio B/I of stereo base (B) to object distance (I). For macro stereo photography this ratio is usually between 1/10 and 1/30. I recommend 1/20 as a good starting point. If you are beginner, you can bracket stereo base, just like photographers bracket exposure, but you will soon find that the choice of stereo base is not critical. You will get a good, natural-looking 3d close-ups for a wide range of stereo bases.

      So, here is what you do: 1) Measure the distance of the subject to the lens. 2) Divide this distance by 20 and use this figure as your stereo base B. 3) Slide the camera parallel by B. I do not recommend converging the lens to the subject. This creates “keystone distortion”.

      Note: At magnifications greater than 0.5x where you usually know the magnification (can be read from the lens’ barrel for example) you can use this formula: I = F (1 + 1/M) to calculate I. For example, when using a 50mm lens at M=1, I = 100mm, B = 100/20 = 5mm.

      Macro Photography with Close-Up Lenses

      When extension is not an option (because the lens is fixed to the camera body, for example) an alternative way to focus closer is to use close-up (or supplementary) lenses.

      Close up lenses are marked in diopters. A diopter is 1/F, where F is the focal length of the close-up lens expressed in meters. So a +1 lens has a focal length of 1 meter (40 inches), a +2 lens has a focal length of 0.5 meters (20 inches) etc.

      Assume that the lens of focal length F is focused on an object at distance s. The image is formed at a distance d behind the lens. From the fundamental lens equation we get: 1/F = 1/s + 1/d. (1)

      Now, let’s insert a close-up lens with focal length f, in front of this lens. Without touching the focus of our lens (so d does not change), now we have to place our object closer to the lens, at a distance s’, in order to have it in focus. The combined strength of the two lenses now is F + f, and the fundamental lens formula now becomes: 1/F + 1/f = 1/d + 1/s; (2)

      Subtracting (1) from (2) we get: 1/f = 1/s' – 1/s, or 1/s' = 1/s + 1/f (3)

      This formula can be used to calculate the sharp focus distance (s’) when using a supplementary lens (f), while our original lens is focused at distance (s). What is interesting is that the focal length of the original lens is not involved. So, we can construct tables that universally apply for all lenses. Another interesting observation is that if our original lens is focused at infinity (1/s = 0) then the sharp focus distance is equal to the focal length of the supplementary lens.

      Based on (3) we can construct this useful table for lens whose close focus is 1m (about 3 ft, common close focus distance in many lenses):

      Close Up Lens / Infinity Focus / Close Focus
      --------------------------------------------
      None infinity 1m
      +1 1m 0.5m
      +2 0.5m 0.33m
      +4 0.25m 0.20m


      Two more comments: 1) Close-up lenses can be stacked. In this case, their powers are added. If we stack a +1 and +2 lens, we get a combined strength of +3. 2) It is recommended that the aperture of the lens is kept closed (f8 or smaller) when using supplementary lenses, both to get a decent depth of field and also to keep the aberrations down.

      When using close-up lenses with rangefinder cameras, a focal frame is useful for framing the subject. An example is shown in the picture here. This is taken from Rodolf Kingslake’s book “Optics in Photography”. I highly recommend this book for anyone interested in this topic.

      Macro Photography with Extension



      When an object is at infinity, the image of this object is formed at a distance equal to the focal length (f) behind the lens. As the camera is focused close and closer, the image is formed further and further behind, at a distance x' from f.

      Most camera lenses are designed to operate satisfactorily with an object lying between infinity and about 10 focal lengths. The maximum lens focusing distance is then about 1/10 of the focal length, representing a maximum magnification of 0.1. One exception are macro lenses who are designed to work at higher magnifications, usually 0.5 or 1.0.

      To make a lens focus closer than its minimum focusing distance, one can use extension tubes or bellows. This is of course possible only if the lens can be removed from the camera. The basic equations in work here are the fundamental lens formula and the definition of magnification. I will write these as follows (see also this blog):

      1/f = 1/I + 1/I'
      and:
      M = s'/s = I'/I = f/x = x'/f

      I am using this notation: f = focal length, x' = lens extension (distance of lens from film plane), I = Subject distance (distance of subject from lens), also known as working distance because it shows you how much room to you have to work with your subject, M is the magnification.

      An interesting situation exists at M=1. In this case, I = I' = 2F, i.e. the extension is equal to the working distance and it is equal to two times the focal length. For example, if you are using a 50mm macro lens at M=1, then the subject is at 100mm in front of the lens and the image is formed 100mm behind the lens. As it turns out, the total distance from object to film plane (4F or 200mm) is the smallest possible. As you increase the magnification beyond 1, the image moves quickly behind the lens and a bellows system will be needed. For best optical performance, it is recommended that the lens is reversed when M > 1

      Another note of practical importance: At magnifications close to 1 or higher, focusing is not done by changing the distance of the lens to the film plane, because this changes the magnification. Focusing is done but my moving the lens and camera closer to the object. So, in macro stereo photography, first you set the magnification by fixing the extension I', and then you focus by changing the working distance I. For stereo photography with a slide bar, you might need a stereo bar with two axes. The vertical axis is used for stereo translation. The horizontal axis is used for focusing.