Stereo camera to tilt or not to tilt

himself · May 12, 2012

Mornin'

I'm thinking of building myself a stereoscopic camera - just for the fun of it - and I have a pretty simple design that shouldn't be too much trouble once I have the parts.

The design will be built around the back of either a 6x9 or 9x12 plate camera back, this is so I can use either paper negs, roll film or use the instant back I modded and have instant 3d prints. Basically my plan is to built a box with the 2 lenses (off 2 old 6x9 folders) at the front and then just attach it to the already assembled back. The front will be split down the middle with a piece card or thin wood so that each lens will only exposes half the frame. Doing it this way will give me 2 images on the same frame so I won't have to mount them side by side later.

The thought I had then today would be, what if I added a tilt to the lenses - anybody know if this has been done or if it will enhance or hinder the effect of the 3D.

Any other problems you might foresee in my design, let me know.

thanks

Old-N-Feeble · May 12, 2012

Two thoughts:

1. I've always wondered why no one bothered to build stereo cameras in horizontal format instead of square or vertical.

2. You'll need matched lenses.

Okay... a third thought:

3. Adding tilt should work fine with matched lenses provided they have enough coverage and you have them precisely aligned during all movements.

Photo Engineer · May 12, 2012

Kodak advised against any tilt and included a bubble level in the viewfinder to ensure any tilt was eliminated.

PE

Old-N-Feeble · May 12, 2012

PE... Any examples of why a small amount of tilt won't work? Does it have to do with size relationships? If so then could one have a "set" front tilt to be used in conjunction with a "set" rear tilt to compensate?

Photo Engineer · May 12, 2012

I have no further information except to say that I was warned by the EK "guru" of stereo about this same thing. So, I stayed away from it. There is an example in the instruction book IIRC. But that was back in the 50s.

PE

himself · May 12, 2012

for the matched lenses I thought I'd just get of the same folder and use one off each, or does it need to be more precisely matched?
coverage shouldn't be a problem because it will be a 6x9 (probably) lens covering 6x4.5

and PE, thanks for the heads up, I won't bother with the tilt then - less work too now, I won't have to build another bellows and can save the material for another LF.

MattKing · May 12, 2012

This is just a guess .....

I'd be willing to bet that the apparent stereo effect would be quite disorienting if tilt is applied.

Something like what happens if you tilt your head back and then look down at the subject from the "bottom" of your eyes.

himself · May 12, 2012

MattKing said:
Something like what happens if you tilt your head back and then look down at the subject from the "bottom" of your eyes.

I have sore eyes now, thanks

Photo Engineer · May 12, 2012

Several camera companies make stereo attachments that fasten on the front of a lens and by means of a prism create 2 3D frames on one standard frame of film. Nikon (IIRC) makes one. A friend produces stereo slides that he shows on his TV in stereo.

PE

filmamigo · May 15, 2012

himself said:
for the matched lenses I thought I'd just get of the same folder and use one off each, or does it need to be more precisely matched?
coverage shouldn't be a problem because it will be a 6x9 (probably) lens covering 6x4.5

Two of the same lens should be matched well enough. I have three different stereo cameras, and two of them are hardly "precision matched": I have an ancient 127-format camera with two simple meniscus lenses, and I have the new Holga stereo with a pair of Holga lenses. Both are perfectly capable of forming decent stereo photos. If anything, I find that stereo improves the performance of a given lens, as your brain combines two separate low-resolution images into one image that appears to have more resolution.

The bigger problem with stereo images occurs when the exposure is not exactly the same between images; our brains don't like seeing the same image with two different brightnesses. I don't know what kind of shutter system you are planning; if it's a shared shutter, or some kind of focal plane shutter, you will be fine. If you are using the leaf shutters on two separate lenses, then I suggest giving them a CLA and getting them to fire at the same speeds.

Diapositivo · May 15, 2012

I think you should in theory be able to match two "shifting" lenses by applying the same degree of shift.

If for tilting you mean exploiting the Scheimpflug law, then I don't see it very feasible as the two lenses would have to be tilted by different angles in order to cover, more or less, the same field of view, that I suppose would mean that the brain would not be able to "reconstruct" the image without some headache as the two images would be different.

Besides, I suppose a 3d camera would be optimized for a certain focus distance (let's say: 20 metres, or infinite). If you try to obtain a 3d image of a near object, the two lenses should be mounted so that they converge (just like your eyes do when you observe a pencil that you move toward your nose). Tridimensionality of non-far objects becomes, thus, quite a complicated matter, and one might expect that you want to add tilt to your images because you intend to include some near objects in the frame.

JerryWo · May 15, 2012

Actually, I shoot a lot of stereo using a Manfrotto slide bar and a (Gasp) digital camera. Good stereo requires that both images are taken aimed straight ahead, with the film plane for both photos parallel to the scene. Your camera as drawn, looks fine, I would not tilt the lenses, but you can certainly tilt (aim) the camera in any direction/elevation.

The Manfrotto slide bar keeps the camera parallel to the scene and it can then be moved laterally. Typically the distance is the same distance as between your two eyes....I think that's about 65mm.

It takes a pretty steady hand to take stereo photos manually (using one camera) without a slide bar. With your camera, it would be a snap (pun intended). You would need to ensure that both imaging units are pointed exactly straight ahead, otherwise there will be modest change in scene perspective between the two photos.

Small irregularities are compensated by the eye/brain combo. Your eye/brain combo is definitely NOT used to having two slightly different images. I have a stereo pair where the wind blew (moved) a branch in one photo and not in the other. It literally makes my head hurt to view this pair.

I print 7 x 7 photos and mount them, with the proper spacing, on opposite pages of a scrapbook (or whatever) and view them with a Geoscope. See:
http://www.berezin.com/3d/geoscope.htm

There is a formula for the correct spacing vs. distance to the object being photographed. We humans (unlike owls, etc.) have evolved to perceive stereo at pretty modest distances.

There is an absolute TON of stuff out there on stereo photography. Google stereoscopic.

Enjoy
Jerry

himself · May 16, 2012

filmamigo said:
Two of the same lens should be matched well enough. I have three different stereo cameras, and two of them are hardly "precision matched": I have an ancient 127-format camera with two simple meniscus lenses, and I have the new Holga stereo with a pair of Holga lenses. Both are perfectly capable of forming decent stereo photos. If anything, I find that stereo improves the performance of a given lens, as your brain combines two separate low-resolution images into one image that appears to have more resolution.

The bigger problem with stereo images occurs when the exposure is not exactly the same between images; our brains don't like seeing the same image with two different brightnesses. I don't know what kind of shutter system you are planning; if it's a shared shutter, or some kind of focal plane shutter, you will be fine. If you are using the leaf shutters on two separate lenses, then I suggest giving them a CLA and getting them to fire at the same speeds.

I was just going to use the leaf shutters.
the stereo camera I use now is a piece of junk and one side is usually overexposed, but it's never been a problem before. I'll keep it in mind tho', thanks.

himself · May 16, 2012

Diapositivo said:
I think you should in theory be able to match two "shifting" lenses by applying the same degree of shift.

If for tilting you mean exploiting the Scheimpflug law, then I don't see it very feasible as the two lenses would have to be tilted by different angles in order to cover, more or less, the same field of view, that I suppose would mean that the brain would not be able to "reconstruct" the image without some headache as the two images would be different.

Besides, I suppose a 3d camera would be optimized for a certain focus distance (let's say: 20 metres, or infinite). If you try to obtain a 3d image of a near object, the two lenses should be mounted so that they converge (just like your eyes do when you observe a pencil that you move toward your nose). Tridimensionality of non-far objects becomes, thus, quite a complicated matter, and one might expect that you want to add tilt to your images because you intend to include some near objects in the frame.

not what I meant by tilt, well, sort of not... I was thinking more of the "miniature" effect (can't be bothered to google the real name) - though it just might look cool.

himself · May 16, 2012

JerryWo said:
Actually, I shoot a lot of stereo using a Manfrotto slide bar and a (Gasp) digital camera. Good stereo requires that both images are taken aimed straight ahead, with the film plane for both photos parallel to the scene. Your camera as drawn, looks fine, I would not tilt the lenses, but you can certainly tilt (aim) the camera in any direction/elevation.

The Manfrotto slide bar keeps the camera parallel to the scene and it can then be moved laterally. Typically the distance is the same distance as between your two eyes....I think that's about 65mm.

It takes a pretty steady hand to take stereo photos manually (using one camera) without a slide bar. With your camera, it would be a snap (pun intended). You would need to ensure that both imaging units are pointed exactly straight ahead, otherwise there will be modest change in scene perspective between the two photos.

Small irregularities are compensated by the eye/brain combo. Your eye/brain combo is definitely NOT used to having two slightly different images. I have a stereo pair where the wind blew (moved) a branch in one photo and not in the other. It literally makes my head hurt to view this pair.

I print 7 x 7 photos and mount them, with the proper spacing, on opposite pages of a scrapbook (or whatever) and view them with a Geoscope. See:
http://www.berezin.com/3d/geoscope.htm

There is a formula for the correct spacing vs. distance to the object being photographed. We humans (unlike owls, etc.) have evolved to perceive stereo at pretty modest distances.

There is an absolute TON of stuff out there on stereo photography. Google stereoscopic.

Enjoy
Jerry

I was thinking of mounting them both in one lens board so that they would be pointing in the same direction no matter what.

oh, and I know what you mean about the stray object in just one of the images. I have a number of them where an odd branch blows into view at the wrong moment, but I'm hoping that using two lenses firing the same time onto one piece of film will help stop that happening.

my original idea was just to use paper negs or sheet film, so I could just contact print them straight away without needing to mount them latter, but now I'm thinking that some instant pack film will also give me images large enough to just use straight out of the camera.

I did google it, but couldn't (read - gave up too soon) find anything on the effect of tilt on the stereo effect.

filmamigo · May 16, 2012

himself said:
my original idea was just to use paper negs or sheet film, so I could just contact print them straight away without needing to mount them latter, but now I'm thinking that some instant pack film will also give me images large enough to just use straight out of the camera.

The instant film is a cool idea, it could look really good. Remember though, you need to swap the left-right images on the print. So the instant print will need to be cut, swapped and remounted.

At the risk of angering the Hybrid gods, I will say that I just use StereoPhotoMaker software. It's free and is a boon for printing stereocards or creating stereo output of any kind.

Also, many folks expose slide film and view the images with a slide viewer. Holga now have a 120-format 3D slide viewer and they sell cheap stereo mounts for 120 film.

himself · May 16, 2012

filmamigo said:
The instant film is a cool idea, it could look really good. Remember though, you need to swap the left-right images on the print. So the instant print will need to be cut, swapped and remounted.

I did not know I had to do that...
shoit

Diapositivo · May 16, 2012

filmamigo said:
The instant film is a cool idea, it could look really good. Remember though, you need to swap the left-right images on the print. So the instant print will need to be cut, swapped and remounted.

I don't understand why this would be necessary.

The image which is going to be seen by the right eye is bound to be formed in the right side of the paper, or negative (inverted, but on the right half) by the right lens.
The images which is going to be seen by the left eye bound is to be formed in the left side of the paper by the left lens.

When you want to reconstruct the stereo image, you must show the image of the left lens to the left eye, and the image of the right lens to the right eye, or headache will certainly ensue, as - imagine being in front of a convex building corner - the right eye will see the left side of the corner and the left eye the right side.

The father, or grandfather, of an uncle of mine took stereo B&W positive pictures of the front during WW1 on glass plates. Those were to be seen through a stereoscopic viewer with two oculars, which looks like a binocular. The two images were impressed on a single glass plate, which was then put - once developed - into the stereo viewer.

No need to invert the two images, and no possibility either. And, by the way, the images are absolutely spectacular.

Maybe I misunderstood what you mean.

Fabrizio

himself · May 16, 2012

Diapositivo said:
I don't understand why this would be necessary.

The image which is going to be seen by the right eye is bound to be formed in the right side of the paper, or negative (inverted, but on the right half) by the right lens.
The images which is going to be seen by the left eye bound is to be formed in the left side of the paper by the left lens.

When you want to reconstruct the stereo image, you must show the image of the left lens to the left eye, and the image of the right lens to the right eye, or headache will certainly ensue, as - imagine being in front of a convex building corner - the right eye will see the left side of the corner and the left eye the right side.

The father, or grandfather, of an uncle of mine took stereo B&W positive pictures of the front during WW1 on glass plates. Those were to be seen through a stereoscopic viewer with two oculars, which looks like a binocular. The two images were impressed on a single glass plate, which was then put - once developed - into the stereo viewer.

No need to invert the two images, and no possibility either. And, by the way, the images are absolutely spectacular.

Maybe I misunderstood what you mean.

Fabrizio

that's what I had originally thought having seen some original stereoscopic glass plates...

Silverhead · May 21, 2012

Two of the same lens should be matched well enough. I have three different stereo cameras, and two of them are hardly "precision matched": I have an ancient 127-format camera with two simple meniscus lenses, and I have the new Holga stereo with a pair of Holga lenses. Both are perfectly capable of forming decent stereo photos. If anything, I find that stereo improves the performance of a given lens, as your brain combines two separate low-resolution images into one image that appears to have more resolution.

The bigger problem with stereo images occurs when the exposure is not exactly the same between images; our brains don't like seeing the same image with two different brightnesses.

Agreed with all of this. I too shoot with the Holga Stereo and have noticed the "increase" in resolution (even though technically there isn't any) when the two images are merged into one by the brain. And yes, the brain doesn't like it when the two images aren't matched or in sync...it struggles to put the images together but can't, which becomes frustrating. Thus, I would steer away from tilt...too much room for things to go wrong and mess up the 3-D effect, IMHO.

johnielvis · May 21, 2012

I think tilt would be a benefit. There is NO perspective distortion using tilt--keep the BACK level and straight and you can do whatever you want to the front, as long as the SAME thing is done to both lenses. Try it out first using a view camera with a shift bar. The tilt on both (just hinge the front with both lenses at the bottom to allow this---same hinge--the lensboard should hold both lenses) will allow you to get more foreground in focus--you can use less stopping down to get some subjects in focus. Try it and you will see. Tilt changes only the plane of focus and has nothing to do with perspective.

OH--as for usefulness...only for when you want far away and VERY close in exact focus--this is only really needed with larger formats--the smaller formats--shorter lenses--they will have enough depth of field.

to tell the truth--building a camera is a real undertaking--lots of things to get working--I'd make it as simple as possible first (no tilt) and then ADD the tilt facilty later---should be an easy add.

I'd advise to build it first...THEN add the doo dads later AS NEEDED--the simpler the lighter and more rigid.

himself · May 22, 2012

johnielvis said:
I think tilt would be a benefit. There is NO perspective distortion using tilt--keep the BACK level and straight and you can do whatever you want to the front, as long as the SAME thing is done to both lenses. Try it out first using a view camera with a shift bar. The tilt on both (just hinge the front with both lenses at the bottom to allow this---same hinge--the lensboard should hold both lenses) will allow you to get more foreground in focus--you can use less stopping down to get some subjects in focus. Try it and you will see. Tilt changes only the plane of focus and has nothing to do with perspective.

OH--as for usefulness...only for when you want far away and VERY close in exact focus--this is only really needed with larger formats--the smaller formats--shorter lenses--they will have enough depth of field.

to tell the truth--building a camera is a real undertaking--lots of things to get working--I'd make it as simple as possible first (no tilt) and then ADD the tilt facilty later---should be an easy add.

I'd advise to build it first...THEN add the doo dads later AS NEEDED--the simpler the lighter and more rigid.

all that is what I thought in the first place. but it's good to have it confirmed.

I've already built a full movements LF camera, so the actual building will be no problem and I'm not that worried about it - so I was really just hoping for confirmation that sticking a tilt on it won't cause any problems after the fact you know, especially because building the bellows is a bit of a pain.

Old-N-Feeble · May 22, 2012

I say "try it" without bias from any previous information.

himself · May 22, 2012

there is one thing I just thought of, how will the front tilt over or under the "wall" separating the 2 images... I don't fancy making 2 bellows

johnielvis · May 22, 2012

2 bellows is the cadillac way--OR you can just put in a septum divider toward the back--the divider only has to be big enough where one of the 2 views doesn't "see" the other lens--or, should I say--the light from the other lens--if it sees the lens but there's no coverage (vignettted) then it's not a problem---size the back divider based on the closest focus distance--anything closer will have the lens closer into the divider and provide ample "shading" is guaranteed...

draw it out on a piece of paper and you shall see.....

Stereo camera to tilt or not to tilt

Moderator

Privacy & Transparency

Privacy & Transparency