3D-Printed Anaperture Single-Shot Anaglyph Aperture

(Note: I've also posted a version of this announcement in the DIY forum.)

Would you like to be able to capture "3D" images using your current film or digital, still or movie, camera? Anaperture is a Customizer-compatible OpenSCAD program to design special dual-aperture discs that allow single-shot, single-lens, anaglyph stereo capture using a conventional still or video camera. Images captured can be seen in 3D using the usual funny-colored glasses... in fact, you can even see the camera live view in 3D that way!

This 3D-printed design, freely available as Thingiverse Thing 1574072 , creates filter masks that can fit inside the threaded filter ring of a typical DSLR lens. The color filter material itself is cut from the colored gels in a pair of viewing glasses that you can buy for under $1.

It looks like this (mounted on the front of an old Takumar 50mm f/1.4):

and makes images like this:

In a single shot with a single lens -- with no postprocessing required.

This is really an improved version of the technique I described in this Instructable ....

Did you never get any replies or inquiries about this post.  Have you ever heard of the Vivitar Q-DOS lens?

Boris Starosta wrote:

Did you never get any replies or inquiries about this post.

Nope... which is a bit odd, because it did get significant traffic on Thingiverse and my related Instructable got lots of traffic.

Have you ever heard of the Vivitar Q-DOS lens?

Of course -- I even bought one to do comparisons. The Q-DOS lens is shockingly terrible at anaglyph capture. There are two reasons:

1. The effective aperture shape isn't circular for the Q-DOS, but mismatched opposing parts of a circle that make bokeh terrible.
2. The filters come together in an excessively clever way, leaving a seam in the middle of the optical path. This seems to seriously disrupt focus quality.

Interestingly, the Q-DOS is actually one of the best of the Vivitar zooms when NOT used in Q-DOS mode, and my front-mounted filters actually produce very good quality anaglyphs with it.

If you look at the patent for the Q-DOS, it's pretty clear they didn't understand exactly how this works. No big deal -- most photographers and most image processing libraries have very wrong ideas about OOF PSFs (out-of-focus point spread functions). Honestly, I did too until I noticed the major differences between the models in image processing research papers and what I measured with actual lenses. My standard reader-friendly overview is A Poorly Focused Talk.

Oleg L K wrote:

Wow! Two "heavy-weights" of stereo-3d posting in one thread !

Regarding the single-shot anaglyph approach, I must admit I didn't try the anaperture. But I did experiment with DIY analogs of QDOS lens - I used to glue red-blue filters on the back of (generic) 35/2.8, Minolta 24-104 f/3.5-4.5 and Minolta 35-70 f/3.5-4.5. All are full-frame lenses used with APS-C cameras.

The filters shaping the OOF PSF need to be positioned such that they form an aperture rather than vignetting, and for most lens designs there are generally several potentially appropriate places for that. I prefer in front of the lens because it is easier to access, but behind the lens is also possible in some cases. Replacing the original iris of the lens always works, but is very inconvenient....

I confirm the bokeh being odd, but generally the results are interesting.

The bokeh are only odd if the two color filters aren't resulting in identical size and shape apertures. For example, just making opposite sides of the aperture different colors produces the "misaligned" bokeh seen with the Q-DOS lens.

In any case the area where both left- and right frames appear with proper "strength" is larger when the lens diameter is larger - e.g. faster aperture and/or larger focal distance.

Actually, it's not about how fast the lens is, but how little it vignettes.

In my opinion the most appealing is 35/2.8 (or faster), since "normal" lens is desirable for stereo, not telephoto. 35/3.5 and 29/2.8 proved to be unusable.

The key problem is that the color-coded aperture needs to be smaller than the largest non-vignetting aperture -- which is usually 2-3 stops down from wide open. With a 35mm f/2.8 lens, that would typically mean around f/5.6, and 35/5.6=6.25, so you'd have to fit both colored apertures within a 6.25mm diameter opening. That can be done, for example using two 2.75mm diameter circular openings separated by 0.75mm (baseline of 3.5mm), but that's pretty touchy for homemade stuff. The resulting effective aperture would be f/12.7... which is actually still ok for FF (or even APS-C, but would be past the diffraction limit for typical MFT).

Correct me if I'm wrong, but the main catch with this approach applies to anaperture too: you cannot make ghost-reduced anaglyph. E.g. if your scene doesn't include saturated primary colors, the result is fine, otherwise the photo is non-viewable.

Fundamentally, all dye-based filters leak. Thus, there is always some ghosting... but there also is potentially enough information to computationally correct it. I got pretty far along on the correction back around 2011-2014, see Reprocessing Anaglyph Images . However, there was significant interest in the technique from some key movie-industry players at the time, and I was never able to produce full-color stereo pairs with color quality sufficient to make cinematographers happy (which is a VERY high standard).

If I was to do this now, I'd probably combine the "color guessing" algorithm I originally used with some AI methods as well as the more typical techniques (e.g., deconvolution and stereo matching, which by themselves produced inferior results). Here's the really short explanation of the color guessing algorithm from that paper:

In a typical anaglyph, some fraction of the scene is naturally aligned to within a pixel. For example, a baseline of a few inches is negligible relative to a background that is miles away. If the naturally aligned regions can be recognized as such, the color information they provide is known accurate. Not only is it accurately describing the full color of each aligned pixel, but it also is establishing that those specific RGB triples actually occur in the luminant-colored scene. For example, if RGB value 0x802a1c is known to occur frequently in the aligned portion of a scene, it seems likely that a pixel in a not-aligned region with R=0x80 and B=0x1c should have G=0x2a.

Suppose we are examining a region of an anaglyph which has a relatively constant color for a horizontal distance which exceeds the maximum possible stereo misalignment between left and right views. The object in this region might not be fully aligned – there may be color fringes on either or both sides of it. However, the colors that appear on the wide overlapping portion of the object are approximately correct. In fact, if there are fringe colors on either side of this region, the actual alignment is most likely a shift corresponding to the width of the fringe, so even more precise color samples can be obtained.

The identification of color fringe areas is easy for humans, and approximately identifying them can be fairly easy for a computer. A simple approach is to look for areas with colors that are heavily biased toward either of the anaglyph coding colors; for example, in a red/cyan anaglyph, strongly red or strongly cyan pixels are good candidates for fringe areas. More accurately, a region with an RGB color tuple in the anaglyph that is relatively far from any RGB tuple known to occur in the image is probably fringe. The accuracy can be further improved by imposing the rule that the pixels can only be fringe if they are in a horizontal group less wide than the maximum shift. We do not know the true colors of fringe areas, but it is likely that they do not introduce color tuples that are dramatically different from any seen in other portions of the anaglyph.

I generally focus my research on things that seem to be most useful to the community, and frankly the interest in anaglyph capture was way lower than I expected... or wanted a level of color accuracy that I couldn't find a way to deliver. Cinema folks are VERY picky about color. 

I could easily pick this work up again, but lacking both research funding and public interest for it, that's very hard to justify....

What would really be nice, is a device that takes full-color left-right photos sequentially through anaperture holes. I'd think of shutter-glasses synchronized with 30fps 4K video capture.

Actually, if Canon hadn't botched the dual-pixel stuff so badly, it would be pretty easy to use it to capture a full-color high-quality stereo pair at a time. I bought a 5D IV as soon as it came out just to test using the dual-pixel raws as an aid to this. Unfortunately, the print-through on Canon's dual pixels is giving only about 2 stops of SNR, and they also botch the exposure computation so by default you get pixels clipping. I'm amazed Canon has even gotten PDAF to work so well with such poor dual-pixel data....

Earlier this year, at Electronic Imaging 2021, I presented Programmable Liquid Crystal Apertures and Filters for Photographic Lenses . I had hoped color LCD panels would be useful as programmable filters, but my experiments say nope. However, it would be pretty easy to use a pair of LCLV to do shutter-synchronized left/right apertures without moving parts (e.g., using the flash trigger signal from the camera to detect the shutter firing or having an external controller control both the camera and LCLV). I found about 9 stop usable DR for the LCLV, so you'd still get some ghosting with high-end cameras (e.g., Sony A7 series) that have 13-15 stop DR, but straight encoding of JPEGs only codes about 9-10 stops DR... so non-tone-mapped (e.g., Sony DRO off) images shouldn't show ghosting. BTW, EI2021 hasn't posted the above paper yet, but my Arduino driver for LCLV is at https://github.com/aggregate/LCLV .

3D Gunner wrote:

So, I played today doing some experiments with the "described system".

Without calculations , I made a cover with red and cyan filters placed on holes with a diameter of 5mm, located at a distance of 15mm between the centers of the holes, mounted on a 50mm lens with f:1.8 (on Sony a6300 camera). The 3D effect is reasonable up to 70-100cm away, being strong at short distances, 20cm for example. But the distance between the holes is too big, it should probably be reduced to 10mm between the centers of the holes to distribute the colors more evenly throughout the frame. At a distance of 15mm, a quarter of the frame is dominated by red on one side and a quarter of the frame by cyan on the other. The problem is observed even after a severe cut (as in the attached examples).

Sides being single-color means your lens vignettes badly.

Basically, vignetting doesn't just make things darker, but actually clips rays. So, you get only rays from one side and no stereo. The solution is to:

1. Use a lens that doesn't vignette. FF lenses are a better bet for APS-C than ones designed to just barely cover APS-C, but lenses vary a lot.
2. Make the pair of apertures fit within the smaller non-vignetting aperture size.

So the "system" works only at short distances, in the close-up/macro area. At least it was fun.

Subjects are ~60cm away from the camera , strong crop.

Subjects are ~15cm away from the camera, moderate crop.

Calcite crystals are ~60cm away from the camera, moderate crop.

Calcite crystals are ~20cm away from the camera, moderate crop.

Obviously, the effect is stronger with a short baseline if you're close to your subject, but with appropriate lens choice, you don't need to be all that close. It's pretty easy to find lenses that will handle full-body portraits with reasonable stereo effect. Think of it this way: the anaglyph effect is limited to fringes the diameter of the OOF PSF (a non-vignetted "bokeh ball").

Still, your shots don't look bad. 

3D Gunner wrote:

ProfHankD wrote:

1. Use a lens that doesn't vignette. FF lenses are a better bet for APS-C than ones designed to just barely cover APS-C, but lenses vary a lot.
2. Make the pair of apertures fit within the smaller non-vignetting aperture size.

Still, your shots don't look bad.

1. The lens used is a very good one, for FF cameras.

2. I will also do some tests with 5mm holes 4mm apart (9mm separation).

I currently use a device with two lenses with an inter-axial distance of 9mm, for 3D macro, half a frame for each eye.

A typical 50mm f/1.4 will vignette badly until somewhere between f/2.0 and f/5.6. If it happens to be one of those f/5.6 ones, 50/5.6 is just 8.9mm, so your 5mm holes 9mm apart would not work. You need 19mm clear, and that would require no significant vignetting by f/2.6. Being a FF lens on APS-C helps your odds, and a medium format lens typically will do even better, but it's entirely dependent on the lens. Vignetting is usually not the quality metric people look for, but it is pretty much THE metric here.

BTW, longer focal length lenses work better because the same f/number gives a larger diameter for your two stops to fit within.

Using this procedure, I expect to be able to obtain a better image quality in terms of resolution, but with severe limitation to the anaglyphic mode and a lack of uniformity in terms of horizontal color distribution.

Lack of color uniformity => vignetting.

Anaglyph capture effectively interleaves the left/right image pixels, so arguably the resolution is the same as side-by-side half-sensor shots. Biggest advantage is you get the native aspect ratio and effective focal length, and you get to use favorite lenses. It also wins in that the rig doesn't add significant size to the lens, unlike the mirror adapters and dual-camera rigs.

3D Gunner wrote:

ProfHankD wrote:

1. A typical 50mm f/1.4 will vignette badly until somewhere between f/2.0 and f/5.6. If it happens to be one of those f/5.6 ones, 50/5.6 is just 8.9mm, so your 5mm holes 9mm apart would not work. You need 19mm clear, and that would require no significant vignetting by f/2.6. Being a FF lens on APS-C helps your odds, and a medium format lens typically will do even better, but it's entirely dependent on the lens. Vignetting is usually not the quality metric people look for, but it is pretty much THE metric here.

BTW, longer focal length lenses work better because the same f/number gives a larger diameter for your two stops to fit within.

Using this procedure, I expect to be able to obtain a better image quality in terms of resolution, but with severe limitation to the anaglyphic mode and a lack of uniformity in terms of horizontal color distribution.

2. Lack of color uniformity => vignetting.

3. Anaglyph capture effectively interleaves the left/right image pixels, so arguably the resolution is the same as side-by-side half-sensor shots. Biggest advantage is you get the native aspect ratio and effective focal length, and you get to use favorite lenses. It also wins in that the rig doesn't add significant size to the lens, unlike the mirror adapters and dual-camera rigs.

2. This phenomenon is not about "vignetting" in the common sense about "lens vignetting" = "Vignetting, also known as “light fall-off” (sometimes spelled “light falloff”) is common in optics and photography, which in simple terms means darkening of image corners when compared to the center."

The "vignetting" we're talking about doesn't diminish as I close the aperture, but on the contrary, it gets more pronounced, so that at f:5.6 the frame is already "split in two", half red and half cyan.

Yes, it's vignetting -- look at my talk slides above.

You should NEVER be closing the native lens aperture when doing these anaglyphs!If your aperture holes are way too far apart, they don't act as apertures at all, but as vignetting elements on the native lens aperture. That's what stopping down the host lens does. Follow the rays.... 

1. As I already mentioned, I used a 50mm f:1.8 FF lens.
The system works as in the images shown, only at maximum lens aperture, i.e. at f:1.8 (5mm aperture with 15mm inter-axial distance).
At an inter-axial distance of 9mm (reduced from 15mm), the system works much better, as anticipated, with much better horizontal color uniformity.

I also tested the version with 5mm apertures and a 9mm inter-axial distance (4mm distance between holes) with a 90mm f:2.8 Macro (FF lens) and the results are as I anticipated, i.e. good with reasonable lateral uniformity (only with the aperture open at maximum, i.e. f:2.8).

Again, it's a matter of correctly sizing the anaglyph aperture WRT vignetting....

The results obtained with two cameras are clearly superior, without being limited to short shooting distances.

Well, sort-of. My original anaglyph capture work was NOT to create images to be viewed as anaglyphs, but to enable the same types of reprocessing as people do with lightfield cameras. In fact, 3 color channels are useful for that, and provide more info than two cameras shot normally. In fact, this method offers both higher accuracy and resolution than lightfield cameras, but it is computationally much more demanding.

Now I am preparing a thin black disc (the material in which the holes are made is recommended to be as thin as possible) which will be laser cut, both the disc and the holes.
The disc with the colored filters will be mounted in place of glass from a UV protection filter, in an attempt to see how much the image quality improves.

In my experience, a programmable paper cutter can actually do better than a laser. It's not a huge difference, but the pivoting knife in a paper cutter gives an edge only CNC milling of rigid materials can match. Lasers basically degrade the material they are cutting, giving a slightly damaged and rough edge; they also leave an angled kerf in thicker material. Anyway, if you're cutting thin material, it will not make much difference.

3. The two lenses in the macro system I mentioned do not have the optical quality comparable to that of quality photo lenses, but still offer separate full color images at reasonable quality.

Expected, although it isn't hard to find good lenses suitable for that approach from either C/CS/D mount lenses or old rangefinder/point-and-shoot cameras. Basically, I'd 3D-print a TLR-like focusing dual mount, probably rail driven, if I were going that route.

Incidentally, for non-macro stereo capture, I've often done things with pairs of cameras, especially CHDK-supported Canon PowerShots. That makes it easy to exactly match the typical human baseline... as I did in:

3D-printed stereo rig using CHDK sync of two Canon PowerShots

3D Gunner wrote:

ProfHankD wrote:

Incidentally, for non-macro stereo capture, I've often done things with pairs of cameras, especially CHDK-supported Canon PowerShots. That makes it easy to exactly match the typical human baseline... as I did in:

3D-printed stereo rig using CHDK sync of two Canon PowerShots

Now I understand what you meant regarding "vignetting". I am not a native English speaker, so I have to follow some of the purely technical details more closely.

I chose to make the diameter of the holes 5mm in the first place to get more depth in the image. The result is quite good in terms of depth of field.

Your 3D-printed stereo rig is very well made. Congratulations!

It's on thingiverse if you want to make one: A4K2: Stereo Capture rig for Canon A4000 .

Really easy print, just a pain to put the battery, regulator, switch, and USB sockets in the handle. I'd use a USB battery if doing it now, but they weren't common in 2014. 

..................................

Right now I'm more interested in making an optical device that mounts the two vertically rotated 3D images on the sensor halves, in Top/Bottom format instead of SBS.
This allows more advantageous storage of "landscape" formats, like the device made a long time ago by Lawrence Heyda Studios .
Do you have an idea about this system?

I don't like flat mirrors for optical systems. Otherwise, this stuff's pretty easy to do.