It seems like the world is ready for a true 3D display. We’ve seen them in sci-fi for decades now, with the ability to view a scene from any angle and inspect it up close. They’ve remained elusive, but that might just be changing thanks to this open-source persistence-of-vision volumetric display.
If the VVD, as it has been named by its creator [Madaeon], looks somewhat familiar, perhaps it’s because editor-in-chief [Mike Szczys] ran into it back in 2019 at Maker Faire Rome. It looks like it has progressed quite a bit since then, but the basic idea is still the same. A thin, flexible membrane, which is stretched across a frame, is attached to articulated arms. The membrane can move up and down rapidly, fast enough that a 1,000-fps high-speed camera is needed to see it move. That allows you to see the magic in action; a digital light processor (DLP) module projects slices of a 3D image onto the sheet, sending the correct image out for each vertical position of the membrane. Carefully coordinating the images creates the POV illusion of a solid image floating in space, which can be observed from any angle, requires no special glasses, and can even be viewed by groups.
With displays like this, we’re used to issuing the caveat that “it no doubt looks better in person”, but we have to say in the GIFs and videos included the VVD looks pretty darn good. We think this is a natural for inclusion in the 2021 Hackaday Prize, and we’re pleased to see that it made it to the semi-finals of the “Rethink Displays” round.
That is amazing.
Nice work.
Next up: durability over time.
Bruh, don’t be such a downer, this is a really prototype
This. The builder admits the rubber band suspension is a weak point, these obviously being there to provide tension on the flexible membrane.
But is a flexible membrane necessary? I recall my dad’s old Nikon SLR’s having ultra lightweight titanium shutters so extending that idea, could the flex sheet be replaced with a very thin titanium sheet etched with a very fine pattern of holes, and coated on one side with luminous paint. The edges of the plate might be folded a couple millimetres to provide some extra rigidity.
Then the rubber bands and edge supports might be replaced with spring-loaded arms or Kevlar (one of my favourite materials) tethers. Maybe even evacuate the box or fill with helium like some 1970s minicomputer fixed-head drives used to be, to reduce air resistance, but it seems the flap rate is sufficient for the amount of Z-travel.
It’s a sweet machine, would love to see further development on it.
The reason for the rubber bands is not to provide tension for the flexible membrane. I mean, they also do this. But the main reason is that as the arms swings up and down, the distance between the arms ends is not constant. As the length of the membrane is fixed, the rubber bands keep the membrane centered and extend to compensate the increased distance.
The membrane can also be rigid. It has to be thin, and I found this this sheet of material for retro-projection (I tested several fo them). It is quite diffcult to find a thin sheet of material, not fragile, good for retroprojection, and transparent when not projected, and also light weight; But I am still researching and open to suggestions.
Spring loaded arms are a nice suggestion, indeed. Kevlar tethers? Do you have a link for those?
Removing the air is a nice idea too, but then I would have to enclose the display in glass/acrylic, and I would prefer to avoid this as the added reflections make the “hologram” less visible. But it would remove a lot of noise, so maybe worth a try too.
By a Kevlar tether, I don’t mean anything exotic. Kevlar thread is not expensive and a small cotton reel sized spool will last many years:
Example https://www.ebay.com/itm/361780340380
It’s used by people who build custom fishing rods. Get the unwaxed thread rather than waxed.
For a tether, you could make a jig with two pins at the right distance apart then tie knotted loops for each end, draw up taught to the pins, add a few half-hitches and secure with a tiny drop of CA.
One end would be attached to a high quality spring, the other of course to your membrane support. But, make sure it doesn’t move (rotate) on there, as Kevlar under tension slices through things readily – it can even cut through itself.
A single strand ought to be sufficient, Kevlar having a superpower – when I say it’s phenomenally strong, I really really mean it. This stuff is no dental floss or polyester thread. Cut it with sharp scissors, make no attempt to break it with your hands.
I’ve used Kevlar+CA binding for repairing all manner of things since I bought my first spool of it 20 years ago (it was way more expensive then) – RC car chassis, broken gears, broken telescope mounting casting and so on.
I also use it in conjunction with carbon fibre rods. A 1mm CF rod along the edges of your membrane would provide stiffness. Actually you could split a 1mm rod lengthwise (with an Exacto blade) and it would still be stiff enough.
Example https://www.ebay.com/itm/133777199874
I also have some thicker Kevlar thread (more a fine yarn) with which I have lifted myself off the ground by a single strand.
I suggest to anyone here on Hackaday reading this: get yourself a spool of this wonder material, you won’t regret it.
The prototype has survived a lot of travel as I attended several faires. It is still fine! But at the end of each day, the rubber bands must be replaced. I have a bag of them, as I was able to find just the correct size and black color so they are almost non visible, and cheap to replace!
Have you looked into trying a spiral surface? It’s going to decrease the overall area, but increase the durability by much.
Or just a spinning oblique mirror, done over 40 years ago in a BYTE magazine article:
https://archive.org/details/byte-magazine-1978-05-rescan/page/n17/mode/2up
Does it compensate for the bed sheet billow effect? It’s not flat most of the time.
The pillow effect is mostly visible at the beginning and end of each movement, but projection starts a bit later after the start, end ends a bit before the end. A thin frame around the film would totally remove this issue, but also adds weight. I still have to find the best material and shape for it.
Maybe some kind of mesh material to allow air to move through more easily?
I’ll bet it doubles as one heck of a fan!
Noise, quite a bit! But does not move that much air, so during these hot days, it is of no help unfortunately ;)
Why didn’t you feature the new open-frame version?
Volumetric 3D displays have an annoying limitation that there is no hidden line removal, which makes it difficult to make sense out of complex objects. So while cool, I haven’t figured out a good use case except decoration.
Is that really a problem here? I mean, there’s only one possible viewing angle here, so the software could handle that. It wouldn’t be perfect since you could be erasing some lines that one of the eyes should seeing, but it’s better than nothing
Technically doesn’t have to be line art – a bit like those printed paper stack 3d prints you could do full colour fully rendered surfaces (in theory) so only the edges nearest you would really be seen – being brighter than the rear, though would probably need biased reflectors in the moving screen to really help cut out the bleed through from the rear and sides. Or very constrained viewing angle so you can only look at the one face being rendered – in which case its trivial to display only the ‘visible’ lines – its just not telling the display to render them.
The full colour etc is much more complex to figure out, probably needs to be mechanically significantly more advanced too, but in theory you can project the full starwars style hologram where only the side you are looking at is particularly visible, even for line art. I think the most likely method to achieve it convincingly would be a retroreflective moving screen and a collection of projectors – with the light from each projector being bounced back near perfectly the way it comes from it gives you however many projectors you use ‘sides’ to the projection – so your sphere would probably look more like a beach ball with some noticeable vertices (only so many pixels you can calculate and projectors you can afford to fit etc).
Personally I don’t think it needs hidden line removal anyway, as in person the image appears properly 3d, so your stereo vision knows where each line really is (assuming you are not visually impaired of course – though even then the size, shape and scale to your accustomed to your visual defects brain is very likely to interpret the distance correctly – though obviously that isn’t true for everyone).
Great prototype, I think next version should work in vacuum chamber to avoid distortions or get better refresh rate
I remember the amazement I felt when I first saw the Actuality Systems Perspecta around 2002, (it uses a spinning screen instead of a waving screen).
On the surface it looks as the steps taken so far has been small, but we have to remember that this project is (as I understand it) made by one person and that it is fairly portable in comparison.
@jpa has a good point about the limitation and I can’t see any feasible way to come around it.
Any solution would have to depend on the point of view of one person, loosing the point of the volumetric display alltogether.
as for use cases, Viewing almost any point clouds would be one such usecase. Maybe boxing matches or chess games would be interesting to watch on this.
Yes, I developed everything by myself.
I would say that the benefits of the VVD are: Simplified workflow (slicing a 3d object in parallalel planes a lot simpler than slicing for a rotating mirror; Costs (the VVD can be built with less than 2K); Safety (even if the rubber band snaps, the film is so lightweight that it cannot do any damage to people/things; While a rotating mirror has to be encased in a hard plastic dome, because if something goes bad, since it is spinning really high speeds, it would do some damage).
Quite a while ago, I had a thought
“How about I 3D-ify POV display by using a 2D LCD Screen instead of a 1D LED strip?”
“Nah, it’s not doable or at least not practical”
“Wow, someone has actually made it”
Now I’ll wait for few more years when the side wall are removed, device size being reduced to that of a milk carton, output is in color and costs as much as android smartphones.
I also had a thought about using an LCD screen.
The problem is the refresh is not nearly enough to draw POV 3d images.
During my research, the best LCD tech was Blue-phase-mode-LCD, still not a commercial product, and that was still limited under 400Hz.
The DLP I am using has 4000Hz refresh with monochromatic images.
Hard to find something better for that price (7-800€ when I bought it).
Have you thought of creating more colors by using more monochrome projectors?