Brainteaser - lens question

[jerry2]

I keep beating myself up on this quandray...

 

If I use two 80 mm lenses on a lens board, (stereo photography) with

a 6 x 12 roll film back, with a septum to prevent light spillage.....

I will get two "almost" identical images on film, each approx 6x6.

 

Now, in the second example, if I keep the tripod at the exact same

position and I will use a single lens with an Image Splitter to

produce two 6x6 images on the 6x12 back. (of course the septum has

been removed) An Image Splitter is a mirrored device to produce dual

images through one lens. It contains two inboard and two outboard

mirrors to reproduce what two lenses would see. It project a left

and a right image onto the film.

 

Here is the question.... what fl lens would I need with

the Image Splitter to get the same images as the first example using

dual 80 mm lenses? Same images, I mean the exact same scene on each

6x6 piece of film.

 

My first instinct was I need a 160 mm lens? Then I

started to think, why wouldn't 80 mm lenses produce the same images?

Thank you all in advance.

[billkantor]

Unless you change the film size or restrict the image size, I can't see why the answer would be anything other than 80mm.

 

FWIW, I don't know what your image splitter actually does but if it produces two images from one lens (as you say), you wont get a stereo image. The lens has to be offset to achieve the stereo effect.

[jerry2]

Bill, the Image splitter has been around for 100 years for stereo photography. I will try to attach an image to illustrate its purpose.

[silent1]

You'd need an 80 mm lens to get the same image size as with the two lens example -- but you won't get identical images for the two eyes to view with the splitting arrangement.

 

With an 80 mm on that very wide format, if it can cover the territory at all (unlikely unless you're using something akin to a Super-Angulon), you'll get images that are near the lens axis close to the dividing line, and well out in the edge of the lens coverage at the outer edges; you'll have significant keystoning, because the images will be projected from a single central point (the lens aperture) instead of two points axially central to each image, making them project a larger image toward the outer edges (possibly bad enough, with this short focal length, to actually prevent the brain from fusing the images at all when viewing). Aberrations will also be much worse further from the dividing line. Yes, a lens with very good linear correction and edge sharpness can avoid these problems to some extent, but such lenses probably cost more than two ordinary 6x6 coverage 80mm lenses.

 

I can get a pair of perfectly acceptable f/4.5 80mm triplet lenses in non-functioning folders (with bad bellows and needing CLA, for instance) for under $20 each at auction, and fabricate the dual lens board and septum, for less than the price of the stereo adapter you're talking about -- and then you might pay $1000 for the 80mm lens to cover 6x12.

[jerry2]

Donald....

 

> You'd need an 80 mm lens to get the same image size as with the two lens example -- but you won't get identical images for the two eyes to view with the splitting arrangement.

 

So my question is, what fl lens with the Image splitter would produce the same two images as a dual 80mm lens set up?

 

> With an 80 mm on that very wide format, if it can cover the territory at all (unlikely unless you're using something akin to a Super-Angulon),

 

I was planning on use a Super Symmar XL 80mm, 226 mm Image circle, whereas the 6x12 format is only a 123 mm diagonal. This should work fine, right?

 

 

> you'll get images that are near the lens axis close to the dividing line, and well out in the edge of the lens coverage at the outer edges; you'll have significant keystoning, .

 

So few people pick up on this keystoning.... I am completely aware of such and planned to overcome this by a distortion correction program after the film has been scanned. This works wonders for this issue.

 

> I can get a pair of perfectly acceptable f/4.5 80mm triplet lenses in non-functioning folders (with bad bellows and needing CLA, for instance) for under $20 each at auction, and fabricate the dual lens board and septum, for less than the price of the stereo adapter you're talking about -- and then you might pay $1000 for the 80mm lens to cover 6x12.

 

Donald, you are very correct.... however, one issue that is hard to overcome is shutter sync. This is critical for me, and I struggle with getting two lenses to fire accurately to 1/500 th second. In addtion, you have to deal with getting two lenses that have fl's within 1/4% of each other. Not always easy. Also, you have to set two lenses each time, vs. the IS method with one lens. So as I see it, there is some drawbacks, but there is advantages also. The biggest drawbacks with the IS is the fact you can't stop down past f11 due to vignetting. But for this application, that is acceptable.

 

So in the end, I am not sure of your answer...will a single 80mm lens with an IS produce the same two images as dual 80mm lenses? All else being equal of course. Thanks, Bill

[jnanian]

hi bill -

 

i am not sure if you have been here yet:

 

http://www.pauck.de/archive/mailinglist/tech-3d/tech-3d.html

 

it is a tech-forum for 3-D/stereo photography. they might be able to point

you in the right direction, since they are WAY into that sort of thing.

also, there a group of stereo enthousiasts in new england

 

http://users.rcn.com/copley.ma.ultranet/StereoNE/

 

you might want to try to contact a guy named jon golden, he makes very high end

stereo cameras, and he might be able to help you out too. ( he also posts on the

other forum as well ...)

 

good luck!

 

-john

[allen_whittier]

Another reason to get an image splitter working is the problem of the reversal of stereo pairs with separate lenses. A lens rotates an image 180 degrees around the central axis of the lens. With two lenses each image is rotated separately placing the two outside edges of a stereo pair next to each other in the center. This requires the two images to be separated and switched. With a splitter the entire pair is flipped and the right and left orientation are maintained When it is flipped back.

 

I shoot every still image I take in stereo using a slide bar. (What's another sheet of film once your set up.) For moving subjects that I want to be stereo I use a duel lens Realist for 35mm, or a Sputnik for 6x6. For the dual lens shots I have to separate and remount the images for viewing. If I were using the splitter they would be ready to view as they came out of developing.

 

By the way two 6x7 images side by side are exactly the old stereo card format and will contact print for stereo cards. Two 6x6s are to close together and need to be enlarged or separated. I use a 6x7 back with a slide bar to get the two images for contact printing. Nothing can move between shots! If you try this you have to get the R/L sequence right. You move the camera between shots the same direction the film is moving in the camera. For example; if the film is moving from the right roller to the left roller in the back take the right image first then the left. Just contact print and mount for viewing.

[jerry2]

Allen, interesting points. But since I plan to scan the film anyway, I am not bothered by the image position on the roll. I have considered a spud before, but the lenses seemed a bit inferior and more importantly you could not sync very fast. However, after the mess I am going through trying to build a camera, I may re consider. One of my goals was min. 1/500th second sync.

 

Since you are very familar with stereo, what is your answer to my original quesiton? And out of curiosity, what MF viewer do you use? If the world would only would stand still, I would love to shoot all my stereo with my M7's using the cha cha method on a slide bar! But everything I shoot moves....

[paul moshay]

To make some stereo photos, either transparencies, prints or Polaroid, I use a slider bar. I have made stereo with two Rolliflex cameras mounted on a 12 inch bar and with shutters timed with a dual cable release. Using transparencies I mount them on cutout cards, and with negatives I contact print them and mount them on flat cards and view them with an antique stereopticon viewer. With Polaroid, I have an SX70 camera and use a slider bar. The object cannot be moving like with the Rollie setup but for static objects the images are about 6x6cm and work fine in the antique viewer. I just purchased a second 8x10 Toyo and might try some giant stereo just for fun.

Paul

[jerry2]

Paul, funny, I have an 8x10 Toyo field camera.... I often contemplated shooting large format stereo...but one major problem, how do you view them? Its seems near impossible to get the virtual view of any stereo image greater than about 14".

 

Virtual View I define as, image size (one dimension) x magnification. I have explored this for quite some time and it does not matter which variable you play with, you end up in the same place... a severe limitation on virtual view size...

[struan_gray]

Bill, to get the 'same' images on film you need the same

magnification, which means you need the same focal length, i.e.

80 mm.

 

If the spacing of the mirrors is roughly the same as the spacing

of the dual lenses the images should be very similar. The

biggest difference is that with the splitter in front of a lens you are

not necessarily using the whole aperture to make the image, so

exposure may be less than the calculated aperture and bokeh

will be a bit odd at wide apertures. For small stops it won't be

such a big deal.

 

The other difference is that the splitter doesn't use the lens' optic

axis, where the image quality is best, so your two images may

have different asymmetric blurring from aberrations which in bad

cases might spoil the stereo effect (although my experience with

cheap binocular microscopes says the brain is good at

compensating).

 

I would expect all these differences to show up much more

prominantly with closeups and macros than with infinity

landscapes.

[jerry2]

Struan, very nice response, thank you. I think you just talked me out of this approach. It seems I would be taking a good, high priced lens, and then making the images look as if they were shot with a 1920's lens. Oh well.... If only I could find a focal plane shutter that would work for both images at 1/1000 th's, this would solve my problem. But trying to make two seperate shutters fire the simultaneously is very hard at high speeds...they just weren't designed for that task. The best I have done is Mamiya leaf shutter lenses which sync up to 1/250th.

[struan_gray]

I don't think the degredation wil be *that* bad. Not unless you

are using a lens that is known for residual errors like a Dagor or

one of the older classics. But you won't get as good results as

you would with two MF lenses in parallel. The beauty of the

splitter approach is that you can play around with kiddie

periscopes and get reasonable results with little financial

commitment.

 

One solution to the shutter problem would be to settle for 2"

images side-by-side on 4x5" film or a pair of 645 frames on

rollfilm. A speed graphic or similar shutter would then expose

both images the same way. With a splitter you could also use a

curtain shutter behind the lens - look for D. Foschi's posts about

mounting a Speed Graphic shutter behind large lenses.

 

What's the application? It almost sounds as if you're trying to do

high-speed stereo photogrammetry. Regular stereo imaging

can tolerate a fair amount of slop.

[jerry2]

Struan....

 

> One solution to the shutter problem would be to settle for 2" images side-by-side on 4x5" film or a pair of 645 frames on rollfilm. A speed graphic or similar shutter would then expose both images the same way.

 

This would be ideal, and I have researched this, however, those old Speed Graphic shutters are very slow and cause lots of vibration. If I remember, they max out at 1/60th of second, I need a min. of 1/500 th. Also, the vibration will kill some of the sharpness which is so critical in stereo. I also looked into Packard shutters, but once again, 1/25th was max speed, and they were too big anyway.... but your on the right page here....

 

> What's the application? It almost sounds as if you're trying to do high-speed stereo photogrammetry.

 

I am trying to produce many of these cameras for several types of events that have moving subjects. Perfect sync is the hard part of stereo imaging... So I was trying to find "off the shelf" products to make such. Hence why I was so attracted to an image splitter, it solves the sync problem perfectly, but it seems like its other drawbacks are too great. As in some of the old Leica and Leitz cameras, they made custom stereo lenses to work with their Image Splitters. They recognized the shortcomings of using a normal lens with an IS. That would be the ultimate solution.

 

> Regular stereo imaging can tolerate a fair amount of slop.

 

Not true of moving subjects. Even a leaf falling down from a tree will ruin a stereo pair if the sync is not perfect, and fast. I have hundreds of rolls of film to attest to this!

 

going to 35mm solves lots of problems, RBT makes such stereo cameras, but from what I here, even their lens syncs have problems. The ultimate solution would be using the Xpan camera body, as it has a 65mm long focal plane shutter, so two lenses could be used with perfect sync.... Too bad the 6x17 cameras do not use focal plane shutters!

[Bill C]

>> I have researched this, however, those old Speed Graphic shutters are very slow and cause lots of vibration. If I remember, they max out at 1/60th of second, I need a min. of 1/500 th. <<

 

I think your research hit a glitch on this one; I'm pretty sure at least some Speed Graphics could expose at 1/1000 second. I'm not saying this is the best way to go, just calling into question an assumption that this is not feasable.

 

If you're just looking at things like a falling leaf (I presume from a distance away; not macro), it seems to me that simultaneously firing two solenoid release equipped cameras would be good enough. Solenoids are probably available which will thread into a cable release socket on the camera. No specific source, but within the past couple of years, Horseman, for example, had such units. You would probably want to use some batteries and a single push button switch to trip both solenoids simultaneously (ie, an "old time" flash gun as used on a press camera has outlets for this).

 

I don't know for sure what kind of actuation times the solenoids would have, but an off the cuff guess would be a couple of milliseconds. Even if longer, the times would probably be a pretty close match to each other. If the camera itself introduces a variable time, switching to the most straightforward mechanical leaf shutter may help; I would expect your Mamiya leaf shutter lenses would be fine.

 

You could fine tune the system, if necessary, by adding resistors to one solenoid circuit. You could probably even use an electronic flash unit to evaluate the timing; trip the flash with a sync cord from one shutter; look through the other to see if the flash is visible. If you used a long exposure on the non-flash shutter, but saw no flash, the indication is that this is the slow shutter; so switch the sync cord to this shutter. Repeat process, always looking through the faster shutter. By progressively shortening the exposure time on this shutter while looking through it for flash, you should get a good idea of how well the two shutter trip events match.

[grant_.]

ud probably need a lens that has glass, and some aperture blades in it...

[jerry2]

Hi Bill

 

> I think your research hit a glitch on this one; I'm pretty sure at least some Speed Graphics could expose at 1/1000 second. I'm not saying this is the best way to go, just calling into question an assumption that this is not feasable.

 

Wow, if this is the case, you may be on to something... I may be able to overcome the vibration issues. I will try to investigate this... I remember a web site dedicated to speed Graphics ...

 

> If you're just looking at things like a falling leaf (I presume from a distance away; not macro),

 

yes, even at a distance, say 15ft....

 

> it seems to me that simultaneously firing two solenoid release equipped cameras would be good enough. Solenoids are probably

available which will thread into a cable release socket on the camera.

 

I have investigated this...and oddly enough, Fuji makes such a solenoid which is sold on the their newer version 6x17 camera. I assume this was to make the camera more "hand holdable". This was my last resort, as I have been told that two of them firing at once will cause lots of vibration. In addition, these noids are not designed to have specific lag times since they only fire one single lens at a time. I could not get specs, so not sure how accurate they would be. I thought about independent noids, and then retrofitting them to have them screw into a cable release threads, but lots of tedious work, hard to get this stuff done.... not to mention have it performed as desired. Lots of intricate work here and the vibration issue may not be overcome. So if I did take this route, I would try the Fuji noids first, $200 each...but considering they will install with no hassles, not too bad.

 

> No specific source, but within the past couple of years, Horseman, for example, had such units. You would probably want to use some batteries and a single push button switch to trip both solenoids simultaneously (ie, an "old time" flash gun as used on a press camera has outlets for this).

 

Yes, a simple double pole momentary push button switch will peform this task well. I synced my mamiya 7's together this way, but even with these very responsive leaf shutters, the best I can sync is 1/250 th, it shows how hard this can be. LF lenses are all mechanical, so I would expect even less success. Of course these M7's are fabulous stereo cameras, except side by side, the min. lens spacing is 165mm, just too damn wide for the intended purpose. Bases wider than human eye spacing (avg 65 mm) introduce too many obstacles in the viewing, such as miniaturization effect, cardboarding, difficulty fusing, etc. There is one guy who hated this so much, he actually fused two M7's together, creating a Siamesed M7, with 78mm lens spacing...not too bad... but he charges $6k just for the labor..you provide camera and lenses. And even then, I still don't have the base seperation I desire and lens sync is 1/250th. If that Speed Graphic fl shutter works at 1/1,000 th, that would be ideal. FL shutters guarantee perfect sync!

 

> I don't know for sure what kind of actuation times the solenoids would have, but an off the cuff guess would be a couple of milliseconds. Even if longer, the times would probably be a pretty close match to each other. If the camera itself introduces a variable time, switching to the most straightforward mechanical leaf shutter may help; I would expect your Mamiya leaf shutter

lenses would be fine.

 

I agree leaf shutters are better than Copal shutters in LF lenses...but are you suggesting installing an electronic leaf shutter in a LF lens? That would be awesome, but I never found any that would fit? Are you aware of such?

 

> You could fine tune the system, if necessary, by adding resistors to one solenoid circuit. You could probably even use an electronic flash unit to evaluate the timing;

 

Actually, a simpler and more effective test is to point the lens fronts at each other, shine bright light through the back of one lens, then a shutter speed tester in the back of the other lens. If both lenses fire at 1/500 th, then when pointing at each other, if they have perfect sync, the shutter tester will also read 1/500th. If they are off, you will always get less than 1/500th, but you won't know which lens is off.... but you can experiment to find out.

 

Off to the Speed graphics site....

 

[Bill C]

I didn't realize the apparent scarcity of screw-in solenoid releases until you talked about the $200 Fuji release. I did a little internet looking, and was kind of surprised that I didn't find anything. Definitely such things USED TO BE readily available, and should have probably been priced in the $20-40 US range. Unfortunately, I'm not finding anything now.

 

If you wanted to experiment with the concept, you might find a couple of old 4x5 press cameras with solenoid releases. This used to be pretty common; if the photog was using the button on the flashgun to trip, a solenoid was doing the work. I've seen Crown Graphics and Super Graphics that had these; I thought all were so equipped, but maybe not.

 

It looks like these people might be selling replacement solenoids; they may be of some use to you.

 

http://www.paramountcords.com/graflex.htm

 

At any rate, if you could find a pair of solenoid equipped cameras, you could experiment. Again, my idea is that if one trip slightly leads the other, you would insert a resistor in series with the solenoid on the fast one; this should slow it down.

 

BTW, I was not suggesting that you swap shutters, etc; just that a mechanical in-lens shutter would probably be more consistent in timing than, for example, an SLR camera.

[silent1]

<i>So my question is, what fl lens with the Image splitter would produce the same two images as a dual 80mm lens set up?</i><p>

 

That's the problem -- to get the same image scale, overall, you'd need the same focal length, but then you'd get all the keystoning and projection scaling issues I talked about originally. I personally think shutter synchronization is easier to solve, unless you're shooting something that's moving so rapidly you <i>need</i> 1/500 to freeze the motion. Otherwise, a siamese cable release and a good, firm hand on the button should get the shutters within 1/60 or so.<p>

 

And if you're scanning and using software already to correct keystoning, you have no reason to need the lenses within 1/4% on focal length; you can just scale the images to match and crop off the extra border on the one that had to be enlarged; that, IMO, is much easier than a keystone correction. It's also trivial to swap the images to convert cross-eye viewing to wide-eye (standard stereoscopic) or vice versa, and (if necessary) adjust vertical registration at the same time. In fact, with a copy of Photoshop between the taking lens(es) and the stereo viewer, shutter synch is the <i>only</i> reason not to prefer two lenses. And I still don't think you'll find an 80 mm that can cover a dual 6x6 frame from a single lens, unless you're working close to 1:1 macro focus (which would throw your stereo splitter into a cocked hat in any case).<p>

 

FWIW, I've seen a web page for a fellow who's been taking stereo video with two synchronized cameras, and combining the two captures into an anaglyph that can be viewed on the web. I don't recall the URL offhand (it's at home, and I'm not), but if you search Google for "Puppet Kite Kid" with the quotes, you'll find him and his web pages. This pretty well bypasses the whole issue of shutter synch, at the cost of huge file sizes, requiring a computer for viewing, and significantly reduced resolution.

[jerry2]

My research also did not reveal any solenoid makers for LF lenses, except the Graflex, which as you can tell by the link you provided, is not well suited for LF lenses. The Fuji noid is made to fit modern LF lenses. It would be more expensive to modify a noid vs. a just buying the Fuji noids...and so far, I can't even find them for sale anywhere, it seems they are only used at Fuji factory repair...

 

It looks like these people might be selling replacement solenoids; they may be of some use to you.

http://www.paramountcords.com/graflex.htm

 

I think these actual noids used on the Graflex....

 

> would insert a resistor in series with the solenoid on the fast one; this should slow it down.

 

It would be easier to use adjustable noids, less guess work with the resistors.... but I follow your point.... I have explored this.

 

> BTW, I was not suggesting that you swap shutters, etc; just that a mechanical in-lens shutter would probably be more consistent in timing than, for example, an SLR camera.

 

I am not so sure this is true.... mechanical lens shutters are pure mechanical, whereas many fl shutters on slr have noids that fire them...beter chance of getting noids synced vs. full mechanical shutters. At least that has been my experience so far...

 

> Otherwise, a siamese cable release and a good, firm hand on the button should get the shutters within 1/60 or so.

 

That's about right.... I have done this many times... and the only way to acheive any consistency with a dual release is by setting up a single cable for each lens, into a dual machined box, wheras a single cable release goes into this piece and fires both shutters. The machined piece needs to have an adjustment piece for sync tuning. Then you must secure the lens cables to the lens board so they never move. However, this device is not easy to make, and in the end, you still have mechanical shutters in lenses which will limit sync to about 1/30 th to maybe 1/60 th max.

 

Agreed if images go digital, one can compensate for mismatched fl's.

 

> It's also trivial to swap the images to convert cross-eye viewing to wide-eye (standard stereoscopic) or vice versa, and (if necessary) adjust vertical registration at the same time.

 

There is software ready made for all these functions... check out www.pokescope.com very good program....

 

> In fact, with a copy of Photoshop between the taking lens(es)and the stereo viewer, shutter synch is the only reason not to prefer two

lenses.

 

TOTALY AGREED!!! BUT THIS IS MY SHORTCOMING...SORRY FOR YELLING :-)

 

> And I still don't think you'll find an 80 mm that can cover a dual 6x6 frame from a single lens

 

See above.... I covered this....

 

> FWIW, I've seen a web page for a fellow who's been taking stereo video with two synchronized cameras,

 

As you know, syncing video is nothing like syncing still shots. It seems the digital revolution may soon hold the solution to this quandry.... Some small digicams will allow for center lens spacing of 65mm, but only in the vertical format, which I might be able to live with. Howver, after cropping, the MP remaining is bit shy vs. film...well, a lot shy. The new Sony RGBE 8 MP cameras hold some promise.... assuming I can jury rig the sync.