Stereo Photography

in the old days you'd use 2 identical cameras and wire the shutter buttons together, or you'd use one camera and a slider to take a second shot from a few inches away. today you just buy a 3D camera.
3D Cameras for Professional Photographers -
http://www.toptenreviews.com/electronics/photo-video/best-3d-cameras/

File: 3d camera.jpg (7KB, 259x194px) Image search: [Google]

7KB, 259x194px

1 lens bad, 2 lenses good.

you should be asking /p/ for this but as a resident from there, i can tell you the most you'd get is "3d gimmic gtfo"

Stereography is as simple as taking a photograph and introducing parallax in the second.
That's the technical term for when things move at different rates in a picture depending on the distance to the camera.

You have several ways to do this:
1. Have one camera, move the camera for a second shot.
2. Have one camera, split the image in two with mirrors.
3. Have two cameras, have them take images simultaneously.

The advantage of the first is that it's extremely cheap.
You don't actually require anything except photoshop for this.
Take a picture, hand hold the camera a few inches away but facing the same direction and take a second picture.
Then auto-align the two in photoshop and put them side to side and voila.
The second one, is as cheap as buying 2 or 4 mirrors but is harder simply on the basis that you'll need to actually make some sort of rig - miniature or not.
You're basically making one or two periscopes and assuring the camera sees into both at the same time.
Then it's as simple as realigning the images correctly.

And the 3rd on is simply using two cameras instead of splitting the image.
This is is generally used for acutal production cameras as it introduces almost no problems.
I was going to sketch the mirror thing but apparently someone already even made smartphone adapters using that technique.
Just watch this and google "3d camera mirrors" and you'll get all the info you need.
http://www.3ders.org/articles/20160927-kula-bebe-mirror-based-lens-adaptor-turns-your-smartphone-into-a-true-3d-camera.html

>>1192416
Yeah I considered simply taking 2 pics side by side but I want to photograph things that may move (eg trees swaying in the wind).

I also considered mounting two identical cameras on a mount and then somehow controlling both shutters from one button.

But that brings issues like making sure both are at the exact same focus, etc.

>>1192420
You could use CHDK compatible cameras slaved to USB.

Alternatively, you could bodge a Panasonic CLT1 3D to a stills camera.

File: s-l1600 3_zps9hiac8wg.jpg (85KB, 1024x768px) Image search: [Google]

85KB, 1024x768px

>>1192401

Look into 'mirror beam splitter'
It puts both images on a single frame

there are tripod addons that allow you to translate the camera left and right by a few inches since depending on the distance of the focal point in the picture the spacing between the two photographs will have to be different.

>>1192401
By adjusting the scale of how far apart the cameras are, you can make things look either massive or tiny because the brain assumes that the parallax is the same as our own eyes.
Good way to take a picture of GIANT bugs. If you can adjust it.
>>1192572
This and some tape would be my go to. However it may pervent you from using zoom.
Definitely use a digital camera for this, you want to see what the picture split will be, not what the viewfinder sees (becaise the lens view is split, not viewfinder), or use a beam splitter 3D lens meant for the camera.
The argument for two cameras wired together is that beam splitter = 2 pictures at once= half resolution.

>>1192736
Why? Your eyes don't move about like that do they? I hope not!

>>1192896
... so the camera can take a picture from the perspective of the left eye, move it a few inches right at the same height without going forward or backward, and take a picture from the perspective of the right eye.
getting parallax is why birds with an eye on each side of their head bob their heads back and forward.

File: IPD.png (9KB, 650x173px) Image search: [Google]

9KB, 650x173px

>>1193069
I think he was talking about >>1192736
>the spacing between the two photographs will have to be different.
(depending on distance to the subject)

Stereo cameras don't vary the spacing between the lenses because the spacing between people's eyes don't vary.

My personal IPD is 66mm

Images taken 66mm apart look exactly as I see them in real life.
If the images were taken with 50mm spacing the 3D effect will be lessened.
If the images were taken with 75mm spacing the effect will be more dramatic.
Commercial 3D is usually taken with the wider spacing to emphasize the 3D effect.

File: SDW3d.jpg (59KB, 480x248px) Image search: [Google]

59KB, 480x248px

Here is one I took about 25 years ago using a Minolta 110 Zoom SLR.
>snap
>move camera to right
>snap
Use PS to combine images after scanning

File: Stereo-3 - DSCN0918.jpg (234KB, 1024x555px) Image search: [Google]

234KB, 1024x555px

>>1192401
It is really easy to do.

1 Camera
2 Tripods
1 Line Level
2 Sticks

Set up the 2 tripods so their tops are level using the line level between two sticks. Take a photo from each tripod while focusing at the exact same thing.

The further away the subject is, the further apart the tripods need to be. For close shots you just use a single tripod and a wooden jig, you can make, that allows the camera to slide across a wooden ruler. You take two pics at different positions along the ruler.

Here's what the 2-tripod method looks like. Though, the tripods were a bit too far apart, hence the discomfort. I just made this as example for this post. You can also use photoshop to overlay the images to align them, if you don't have embedded levels on your tripods, and ensure the angles of the photos are correct.

File: you're welcome.jpg (99KB, 480x248px) Image search: [Google]

99KB, 480x248px

>>1193087
For stereo pairs it's best to use the crosseye method, so the images need to be swapped. Using the parallel method (as with stereograms, as opposed to stereo pairs) is quite uncomfortable on this scale.

File: you too.jpg (402KB, 1024x555px) Image search: [Google]

402KB, 1024x555px

>>1193125
>>1193186
Same problem with this one. Have you both been busting your eyeballs all these years by having the images the wrong way around?

Seems a pretty basic mistake for someone who goes to the effort of making such pictures in the first place.

>>1193186
I prefer the parallel method so that's what I use.
The crosseye method is tiring and uncomfortable for me.
I can look at parallel for hours without discomfort.
Additionally, some cardboard and a couple of magnifiers make a simple viewer for those who can't be bothered with learning how to do it.

>>1193188
>Seems a pretty basic mistake
No mistake at all - it was purely intentional.
They're my images for me.
I promise to not direct you on how to live your life.

>>1193189
>>1193192

I find it hard to believe that anyone could comfortably view >>1193125 in parallel.

Try the altered ones in crosseye and see how it is. There's a reason that stereo pairs are all crosseye, parallel is reserved for stereograms because the distance is only about an inch.

File: Stereo-1 - DSCN0918.jpg (3MB, 7022x1504px) Image search: [Google]

3MB, 7022x1504px

>>1193188
True. Mine in >>1193125 was a rush job for the post. I did it correctly with the original full size image, but when re-cropping for the smaller one I forgot which layer was which evidently. I only re-cropped it because I have to sit like 5 feet from the screen to view the large one.

File: stereo.jpg (1MB, 3600x3264px) Image search: [Google]

1MB, 3600x3264px

I find setting up a tripod or a using a dual camera setup to be too cumbersome when trying to capture wildlife.

File: mas04a.gif (1MB, 480x640px) Image search: [Google]

1MB, 480x640px

>>1192442
Please consider this method. If you can find two identical Canon Powershots that are supported by CHDK, you should be able to use Stereo Data Maker. You can make or buy a USB trigger. I used an arduino, a usb breakout board from an old pc, a 9volt, a couple USB to mini cords, and a button. You can run scripts and do some cool things. http://sdm.camera/index.htm

>>1193069
i thought you meant two cameras on the tripod then you can move them both. ok cool.

Here's a good read on the subject when using a single camera:

http://nzphoto.tripod.com/sterea/stereotake.htm#shift

>>1192416
>you should be asking /p/ for this but as a resident from there, i can tell you the most you'd get is "3d gimmic gtfo"

I run a photography business. Occasionally I wander out to /p/ but usually the fad goes away quite fast. It's a nasty board full of fuckheads, hipsters, filmfags, gearfaggots and other riff-raff.
I like it here more.

File: DSCN0931b.jpg (178KB, 1024x384px) Image search: [Google]

178KB, 1024x384px

>>1193125
>>1193197
I did these with only one tripod. I actually got better results than with two when lining things up for some reason. I didn't need to rotate anything in photoshop, though it might need it a little bit.

Having a dedicated setup with two cameras taking pics at the same time would be best to prevent movement problems from wind or animals/people moving around.

File: DSCN0938b.jpg (156KB, 1024x384px) Image search: [Google]

156KB, 1024x384px

>>1193408

>>1193313
Thanks. I have some Arduino programming experience so this is up my street.

>>1192416
>>1193405
I've never been there and just went for a visit. Those are some miserable fucking people.

>>1193294
Mmm women's asses look so much better with a third dimension.

>>1193313
Is he the one?

File: A1c0Fm[1].jpg (14KB, 320x199px) Image search: [Google]

14KB, 320x199px

>>1193870
I've never seen so many letters! He truly is the one!

File: 1379090796026.jpg (135KB, 1024x726px) Image search: [Google]

135KB, 1024x726px

I've decided to invest in a decent dual-lens digital camera. I'm looking at the Fujifilm Finepix 3D - are there any others I should consider?

>>1193196
You can do it with practice. Crosseye wasn't easy to view the first time you did it.

File: is it a potato.jpg (19KB, 240x219px) Image search: [Google]

19KB, 240x219px

>>1197721
>Crosseye wasn't easy to view the first time you did it.
I genuinely feel that it was for me.

Surely anyone can cross their eyes to quite a large extent. All they have to do is look at a point close to their eyes.

Conversely, moving one's eyes further apart is harder, not least because in practice your eyes are always partly crossed unless you're looking at something very far away (then your eyes are more or less parallel).

That's hard enough without a visual cue, but some of these stereo pairs - given their size, if viewed in the parallel method, would require you to point your eyes even further apart than as if they were trying to focus on something infinitely far away. Clearly this is something the eyes would *never* be called on to do in their regular use.

So in reality your eyes have had much more practice going cross-eyed simply by doing their job.

There is no natural precedent for using the parallel method on distances of this size.

>>1197721
>>1197947
That's not what he's talking about. The images in >>1193125 are transposed. Compare it with the corrected one in >>1193188 which is the correct order they should be viewed in.

File: 1379091445876.jpg (351KB, 1400x1136px) Image search: [Google]

351KB, 1400x1136px

>>1197953
>Compare it with the corrected one in >>1193188 (You) which is the correct order they should be viewed in.
I see. That was my post too.

File: 20170622_222245.jpg (105KB, 1342x503px) Image search: [Google]

105KB, 1342x503px

>>1192401
Just downloaded a free app onto my samsung s7 and took this picture, works fine for an amature.

File: animated-magic-eye-2[1].gif (3MB, 512x312px) Image search: [Google]

3MB, 512x312px

>>1197721

Yes. Yes it was.

In fact, when one of my gradeschool teachers introduced those magic eye pictures to the class, almost all the students had trouble seeing them properly because doing them parallel (which almost all of them are) was so much more difficult than crosseye.

I don't even know how the fuck a normal human could see stereoscopic images that big in parallel, at least on my screen. The centers of the images are further apart than my eyes, meaning I'd actually have to go lazy-eyed to see them properly, which I simply cannot do. And the vast majority of humans can't, because there's no situation in real life that requires you to focus your eyes _beyond_ infinity, while crossing your eyes slightly is commonplace in order to view anything closer than a few feet.

File: 1498112525213.png (563KB, 561x592px) Image search: [Google]

563KB, 561x592px

>>1197984
Its a way to store messages in secret.

>see poster of boat
>tell kids around me its a lion
>they all see the lion

File: 1498191572591.jpg (115KB, 500x667px) Image search: [Google]

115KB, 500x667px

>>1197984
rotating intersecting rings, what do I win?

>>1197984
Somehow disappointed it was not goatse.

=\

How far can you get and how fast did you get there on http://www.spotthedifference.com/ ? Free viewing methods makes this shit easy.

>>1197953
>corrected
You mean in the sense that cross eyed viewing is correct and parallel viewing is incorrect? Why should you dictate which viewing method people use? I'd get that logic if this were a cross eye thread, but this is stereography, people could post anaglyphs as well.

>>1198065
We are obviously talking about crosseye.

>>1198065
OP here, this reminds me of this nice little "racket" I had with those touch screen spot the difference games in pubs. I'd cross my eyes and all the differences would jump out immediately.

I'd always get a few quid. Then they must have cottoned onto that because they started to put the images on a tilt from one another.

File: 1379011333480.jpg (882KB, 914x720px) Image search: [Google]

882KB, 914x720px

>>1198065
>Why should you dictate which viewing method people use?
Because if you use the wrong one the depth is inverted and it looks whack as fuck.

Given this, why do you think that almost every stereo pair I've ever seen (including every one in my folder) is crosseye?

>>1197984
Basically what I said.

o/

>>1198179
>Given this
You mean that given parallel and crosseye are opposite viewing methods? That has only a little bearing on the question you posed. I'd assume that corsseye is more frequent because of how most camera rigs are set up. The left camera producing the image on the left and the right camera producing the image on the right.

>>1198208
>The left camera producing the image on the left and the right camera producing the image on the right.
>crosseye

>>1198311
Yup, that's how you make crosseye images. I just tried it with a camera to confirm. Take one image of something still, move the camera a little bit to the right and take another. You take the first image as the left of the pair and the second image as the right of the pair. Then you cross your eyes.

File: pedstereo.jpg (207KB, 1600x1200px) Image search: [Google]

207KB, 1600x1200px

There are parallel and crosseye photo setups exist because there are different methods to view them.

Stereoscope viewers work with parallel images. They obviously require a viewer to work, but it's simple to just plant the device to your face and it instantly comes into focus.

Crosseye images that a bit more effort to view correctly, but they are widespread because you don't need a stereoscope viewer on hand. Just cross your eyes and boom.

So y'all can stop bickering about which format is correct. In the end, it's piss easy to swap the images around to whatever format you need.

>>1198527
>They obviously require a viewer to work
Parallel images don't require a viewer to work. If that were the case then all of those magic eye books would come with a viewing device.

to view large format images in stereo, you need to use cross-eye mode
to view small format images in stereo, parallel is easy - look through the picture

>>1192401
Can anyone explain why crossing my eyes inward shows the image clearly with correct depth, but if I cross my eyes outward it appears as if the mountain is closer than the snow mound at the bottom right?

>>1198565
Magiceye books are not parallel.

>>1198799

They almost always are. I exhausted my library's stock of them when I was a kid.

But the patterns used in magic eye are a small fraction of the whole image. It's much easier to view those in parallel than whole page/half-page images.

>>1198799
If they're not parallel, then why don't people cross their eyes to view the books?

>>1198804
>>1198985
see
>>1198796
and stop being morons

>>1198502
Doesn't seem right to me. Surely the word "cross" implies that your eyes' paths are crossing to see the images, hence the right eye looks at the left one and vice versa.

>>1198799
Yes they fucking are. There's no point in even arguing the point. It seems there's an awful lot of people in this thread who've been viewing stereograms wrong all these years.

I remember in school there was a kid who'd clearly been doing that - when I told him, the amount of sheer, aggressive denial was staggering. You'd have thought I'd just told him he'd been using his dick to shit through his whole life.

It seems he's not the only one who's touchy about this sort of thing...

File: IMG0011B.jpg (206KB, 1600x1200px) Image search: [Google]

206KB, 1600x1200px

Exhibit A. The actual instructions from an actual Magic Eye book.

File: IMG0013A.jpg (208KB, 1600x1200px) Image search: [Google]

208KB, 1600x1200px

>>1200347
Exhibit B: stereogram in another book I have called 3D Planet. Note the icon in the lower right, indicating to use the parallel method.

File: IMG0014A.jpg (135KB, 1600x1200px) Image search: [Google]

135KB, 1600x1200px

>>1200349
Exhibit C: another page from the same book. Note the larger stereo pairs are to be viewed crosseye (because the distance is too large for parallel) - whereas the smaller pairs are meant for parallel.

Now can someone recommend a good stereo camera?

>>1200344
I think your eyes are misaligned or something.

>>1200349
clearly that image is, "crossed axes" shown in >>1198796

If you focus on it and cross your eyes to view it, hold your finger out in front of your face, maintain focus on the image, not the finger. Move the finger away from your or towards you until it becomes 1 image. This should be a few inches in front of your face to 1 foot in front, depending on how far away you are from the photo. That means it is crosseyed, just like in >>1198796

>>1200357
>>1200359
Are you two just trolling or just that stupid?

>>1200347
It says right here to use parallel.

>>1200349
And there's even an icon indicating such for this image. As evidenced by the parallel lines - as opposed to the image here >>1200350 where not only can you see the other icon for crosseye (hint: it's an X), but it even demonstrates the differing conditions under which to use parallel or crosseye (i.e. larger distances use crosseye because your eyes are not meant to go beyond parallel to the point of actually diverging).

Again, I refer you to my old school pal and his fervent denial.

>>1200363
You were just proven wrong in >>1200359 via the finger test, yet you are the one in denial.

>>1200373
>You were just proven wrong in >>1200359
All that post did was provide instructions to "trick" your eyes into doing crosseye. With reference to a stereogram which is clearly marked with an angular U-shaped icon, which tells the viewer to use the parallel method.

I even posted a pic with stereo pairs using both methods (>>1200350), with the icons clearly indicating crossed lines or parellel lines. What do you think they mean?

>>1200388
>>1200350
I'll even quote the instructions (partially visible in the pic):
>Use the crosseye method to view the top and middle stereo pairs, and use the parallel method to view the stereo pairs at the bottom.

File: diagram.png (15KB, 640x480px) Image search: [Google]

15KB, 640x480px

>>1200343
That's what I described. Maybe a diagram will help people understand.

File: parallel .png (14KB, 176x483px) Image search: [Google]

14KB, 176x483px

>>1200388
>What do you think they mean?

This is parallel method. The image in >>1200349 works perfectly with cross-eye.

>>1200441
>The image in >>1200349 (You) works perfectly with cross-eye.
No it doesn't. How can you even know that without seeing the full image? You can't even see the main feature in that pic. For your information it's birds flying in front of the trees.

Viewing it in crosseye inverts the depth map, meaning what are meant to be birds hovering in front of the plane are now inverted bird-shaped holes eating back into the said plane. It looks completely wrong because that's now how you're meant to view it.

If you've been viewing stereograms in crosseye all this time then you've been doing it wrong. It's that simple. Every single one you've viewed has been a gross parody of the author's intention, because of your misconception of how to do it.

Even the example pic you used (which I posted) SAYS to do it parallel.

You fucking idiot.

File: download (5).png (812KB, 1800x581px) Image search: [Google]

812KB, 1800x581px

http://3d.jespertheend.com
Check out this website if you want to make your own. They can turn out pretty cool

File: IMG0015A.jpg (266KB, 1600x1200px) Image search: [Google]

266KB, 1600x1200px

Here's the complete image. The other one, in addition to being cropped, was also zoomed in too far to be doable in parallel (hence why you probably did it crosseye). Hopefully this one is clearer.

>>1200453
That's just the same image twice but I'm sure you know that.

>>1200450
You can see part of it, in the upper left corner and that's all you need to know it is working for the entire thing.

>>1200455
It is clear as day using cross-eye.

I think your book is a load of shit.

>>1200461
>that's all you need to know it is working for the entire thing.
Aw shit, it's as if you think that all you've got to do is make the texture overlap in your vision, regardless as to which method you use, and it's golden.

I'm sorry but you are completely wrong. There is no scope for personal preference. If you use the wrong method then the depth information is inverted.

What should be close is far away, and vice versa.

Just because you got the tree texture to intersect itself doesn't mean shit.

Also, the part of the image visible in my first pic carries almost no depth information. It's simply the flat plane against which the birds fly. So yeah, in this special case, it doesn't matter - each method will yield a flat plane.

As I say: doing it in crosseye means the birds are impressions carved into the plane (which I'm sure you can see). What you should see are birds standing proud from the plane.

File: crosseyedtopparallelbottom.png (649KB, 500x646px) Image search: [Google]

649KB, 500x646px

>>1200461
You're wrong, take this image as an example. When you view it in cross eyed the top half of the image is correct, with the mountains in the top as the furthest away from the viewer; the bottom half will have the mountains the closest to the viewer, an inverse of the top depth so to speak. When viewed in parallel view the depth of the bottom half is correct while the top is inverse of the bottom.

>>1200466
While true I don't think you can get him to believe it.

>>1200456
Yeah, just giving an example. You can change all sorts of settings around to make it actually do something

>>1193186
Wanted to ask about this. Cross eye is so much easier for me, but the vast majority of pairs I see are the parallel type. Is there a reason? Is parallel easier for most people?

>>1193186
>For stereo pairs it's best to use the crosseye method,
FOR. LARGE. IMAGES.
The pic you swapped sides on is a small image perfectly suited for parallel viewing.

>my way is better than your way - you should use my way
>stop doing what I don't like to do

>>1200755
>Is parallel easier for most people?
I wouldn't have thought so, for distances of more than about 2 inches. See: >>1197947

>>1200880
Not for me, personally. I suspect that this is the case for most people but I'd be open to being corrected if I'm wrong.

Either way, parallel is generally only used for stereograms where a texture of about an inch in width is repeated.

Unlike >>1200755 I don't think I've ever seen a stereo pair intended for parallel. All the ones in my books are crosseye.

>>1201087
>All the ones in my books are crosseye.
Tell a lie - there are the bottom ones on this page >>1200350. But as we can all see, they are very small.

Right, so... about stereoscopic cameras then...