It’s the year 2260 and you’re being beamed from your starship to the planet below. Being a descendant of present day 3D printers, the transporter prints you out, slowly making one layer before moving on to the next, going from the ground up. The you-that-was hopes nothing spills out before you’re done. But what if you could print every atom in your body at the same time? If those transporters are descendant’s of Daqri’s holographic 3D printing technology then that’s just what will happen.
Daqri’s process is akin to SLA (stereolithography) and SLA/DLP (digital light processing). In SLA, a laser beam is shone onto a pool of resin, hardening the resin at the beam’s point. The laser scans across the resin’s surface, drawing one layer. More resin is added and then the next layer is drawn. In SLA/DLP, the light for an entire layer is projected onto the surface at once. While both methods involve stereolithography, the acronym SLA by itself is commonly used to refer to the laser approach.
Daqri’s process however, uses a holographic chip of their own making to project the light for all the layers at the same time into the material, a light-activated monomer. Their chip is a silicon wafer containing a grid of tunable crystals. Those crystals control the magnitude and phase of light reflected down into the monomer, creating a 3D volume of interference patterns. The brief description of the process says that a laser is used to shine light onto the crystals, so there’s probably still some scanning going on. However, in the video, all of the object being printed appears illuminated at the same time so the scanning is likely very fast, similar to how a laser in a light show seemingly paints what appears to be a 2D shape on the side of a building, even though it’s really just a rapidly moving point. There’s also the possibility that the beam’s point is large enough to encapsulate all of the chip at once. You can see a demonstration of it in the video below.
In the video, Daqri prints a small paperclip, and the reason for that small size is likely due to the size of their chip. But if it’s scaled up then heating may become an issue. The “hardening” that goes on, called polymerization, involves the formation of long, tangled polymers from monomers and is exothermic, meaning it gives off heat. If, like us, you’ve worked with resin before then you’ve probably noticed how hardening a large volume of resin produces more heat than hardening a small volume. That heating can be enough to melt and deform the object itself.
There’s no word on when this process will escape the lab and appear in our workshops, but for a future Star Trek transporter, it’s a step in the right direction. In the meantime, for an SLA/DLP 3D printer, have a look at the open source RooBee One.
Our thanks to Sascho for tipping us off about this.
If it’s holographic it probably ues one non-scanning laser which covers whole chip. Holograms use coherent nature of laser to do their magic.
Agreed. No scanning. Very unlikely as mechanical movement and holography exclude each other due to vibration problems.
That was actually pretty cool! I was genuinely surprised by the speed. It’ll be interesting to see where this goes. I’d be curious about occlusion limits, and whether there’s… I don’t even know what you’d call it… residual polymerization in the volume of the resin where the light isn’t at it’s highest intensity, but still passing through.
Hard to know with so little information but given its holographic laser nature. A decent guess is that waves cancel out. Remember laser is not like focus light.
In this instance, the holographic image they are projecting into the resin is just like focused light – the hologram acts like a very very complex lens, focussing light at different distances so that it appears to come from an object (in this case a paperclip). For that reason there will probably be some issues with residual polymerisation (it will probably manifest as ‘fuzziness’ round the object so affect the dimensional tolerances).
I think it’s actually a 2D hologram or close to one. There is a reason why it’s a paperclip and not an owl.
Makes sense, since if they had a way of projecting computer-controlled 3D holograms, using this wafer full of tunable crystals, they could sell that alone as a new form of TV set, and make billions. People have been researching R2D2-style holograms for decades now.
just what I was thinking, also as the part develops it’ll start blocking light below it.
There is no limitation inherent in this method that would prevent a true 3D object being printed all at once, even though the object they print in the video has only a small amount of depth. What you do find however is that it becomes a lot more computationally challenging to generate the holographic pattern necessary to form the image if it is 3D rather than 2D, and for a deeper image you will end up with more noise and therefore a much dirtier less precise print for the same resolution in the hologram.
Even if there’s occlusion, maybe it can increase the print speed or even simplify the mechanics of the printer. Just pipeline the whole 3d image over time with new parts starting as old parts cure and start to occlude. If the hologram can pass the top of the resin well enough then the image can be drawn inside the volume and all the unreliable mechanics involved with moving the interface layer can be ignored.
Hmmm. It makes me ask what does “hologram” mean? This doesn’t use a hologram formed in the resin. It looks like a hologram is used to direct the light. For each print you have to first create a hologram to use as the “projector”, which is pretty much what the headline says.
I foresee great job loss in the paperclip industry.
Agreed. It’s not hologram based at all. Apparently we have no word for this. Simply put it’s a Voxel Laser Printer. There must be something sexier than VLP, right? Beside that, it’s very promising, very exiting stuff.
This method is absolutely hologram based – they are generating a holographic diffraction pattern and displaying it on a Spatial Light Modulator (of some novel type of their own design), which is then being illuminated by a laser beam. The diffraction pattern is redirecting and focussing the light into the shape of the paperclip (the holographic image).
Is this the same technique as how bubblegrams are made? Q-switching a focused point.
” Being a descendant of present day 3D printers, the transporter prints you out, slowly making one layer before moving on to the next, going from the ground up. ”
I was thinking something more akin to TRON™.
Well that’s the next level up.
This method was how I always imagined a rapid prototyping machine to work those last few pre-internet years when they were a mythical oddity hidden maybe at MIT and Area-51, some sort of photo-cure and scanning or hologram in a tank of special resin.
Forget the printing. This chip … it could be also programmed to generate a length of linear field of additive interference of coherent laser light along its axis of symmetry … a.k.a. a lightsaber blade !
Forget the lightsaber. This chip has to be used for actual, real, 3D holograms!
Cool demonstration, although it looks like it was a fixed pattern. The holographic projection does get rid of the need of galvos or a regular 2D projector, but at the expense of needing a flat field laser projector and keeping things relatively still during exposure. Not sure how they would do a second layer and what the z-dimension resolution would be. A demo printing multiple paper clips at different z-spaces would be impressive. The ability to render a hologram on the fly from some programmable projector would be even more impressive. This is a real difficult technical problem, as I helped on a holographic optical memory device about 20+ years ago. Difficult problems make great companies when someone finally figures out how to solve them!
And the patents make millions. ;-)
The reasons for it being a fixed pattern are likely to be computational, which has always been the bottleneck thus far for rapidly updating holographic imagery. The concept of generating the object in ‘layers’ is in this case slightly misleading because a holographic image is not formed in layers, it is formed all at once from a single (usually) 2D image in the diffraction plane (the hologram) illuminated by a plane wave of coherent light (usually an expanded laser beam). I worked on an updatable full color true holographic display in the early 00’s for Ford Motor Company, and the technology has been around for a while but the computational requirements are probably only now beginning to catch up with easily available processing power.
But holograms can’t be dynamically created, can they? When I took a physics of light class back it the 80’s we generated a hologram, requiring a sand table, a beam splitter, a target object, and a piece of film. Does there now exist a way to do holographics without all that?
google for CGH – Computer Generated Holograms. Eazy math, but a lot of.
As for holographic display itself – take a cheap picoprojector, and you will get a hologram capable very high resolution display.
Or you can print calculated hologram on HiDPI laser printer.
For the beginning you can read something like that: http://corticalcafe.com/prog_CGHmaker.htm
What, you mean “Help me Obi-Wan Kenobi you’re my only hope” holograms? Moving images floating in the air in front of you, that you can put your hand through, and are real 3D, that you can look at from every side?
Far as I know they’re impossible. Sega did a game called Time Traveller that gave a floating image, except it was only flat 2D. It was a parabolic mirror and a monitor, the mirror reflected the monitor’s display into your eyes so it appeared to float in the air.
The page you link to doesn’t seem to be able to generate live, dynamic holograms. Just some sort of calculated diffraction thing, or is it a lens?
> What, you mean “Help me Obi-Wan Kenobi you’re my only hope” holograms?
Something near that, but they are monochrome, and the image will not appear as visible from any angle. And you will need a lot of GPU to make them real-time. But in common, as “live”+”moving”+”real hologram” – yes, it’s possible, even in DIY version.
> The page you link to doesn’t seem to be able to generate live, dynamic holograms. Just some sort of calculated diffraction thing, or is it a lens?
Hologram itsels is just diffraction grating. The page I link just shows the way, and it must be clear to anyone, that with the word “Java” nothing could be “dynamic” or even “fast”. However, the same thing done in CUDA or OpenCL could render 3D objects into holograms in real-time. Just google it.
Also, you should know, that there are 2 types of hologram – one is “planar”, where diffraction grating is just one 2D layer, and another is “volume” where grate is made in the 3D volume of some media. “planar” are monochromatic and they need a laser to reconstruct image, second could be viewed with ordinary illumination, but the image will appear only “behind” media.
Since there are no any devices with the ability to create gratings in volume available, only the “planar” holograms could be interesting for DIY. DLP matrices and LCD’s from picoprojectors is good candidates to play with CGH at home.
What you’re saying in terms of the type of holographic image depicted in Star Wars being impossible is sort of true but sort of not. It’s currently not possible to form a real image in mid-air that’s viewable from any angle without some diffusing medium being present (even if that’s the air itself, having been altered in some way to make it diffusing), so there are a lot of volumetric displays out there that move screens rapidly to sweep out a volume that can display volumetric images of the sort depicted in Star Wars. It is also possible to create a true holographic image that _appears_ to float in mid air, but the range of viewing angles will be restricted. Your example of the illusion Sega used for Time Traveller is similar to this, but the parabolic mirror is forming a real image of whatever is on a screen so that it appears to float above the device. This illusion even works for 3D objects and is used in fun illusion products like this one: http://ezphysics.nchu.edu.tw/demo/6_6.htm where a pig appears to float above the device. The pig is actually sitting inside the device, so when you go to touch it, it’s not there because what you’re seeing is only an image of the pig formed in the space just above the device. You’ll find that even in this case however the viewing angle is restricted – as soon as you move down low enough so that you can’t see through the hole, the pig disappears. You also can’t use an object that’s larger than the diameter of the hole in the top otherwise you’ll only see part of the object.
The idea of using mirrors to form a real image using a hologram has been done (I know because I worked on one!) but the limitation comes from the fact that the wider the viewing angle you desire, the larger the mirror you need to form the image. You can trade off image volume for viewing angle for the same sized mirror, but it’s a trade off. We got to the point where we could not find anyone capable of making a mirror large enough to achieve a decent viewing angle and image size combination for a reasonable cost (this might have changed now, this was early 00’s). So in the case of R2D2, the mirror in his ‘projector’ would have needed to be approximately the size of Tatooine to project the image he did in reality, not the size of a small camera lens as is depicted in the movie :(
As others have suggested, Computer Generated Holograms have been around for a few decades now, and the devices necessary to generate them dynamically (Spatial Light Modulators) have also been around for quite some time – you will know them as LCDs. Now you can get a holographic image out of a low resolution LCD, but the lower the resolution the grater the error in your final image, and this generally manifests as random noise in the generated holographic image. This means that the greater the number of pixels you can use to generate your CGH pattern the better, because it will result in a more noise-free and accurate final reproduction. The other factor that affects final image fidelity is the mode that is used to modulate the light that illuminates the SLM. Light can be modulated in amplitude, phase, or both by an SLM. Modulating amplitude or phase alone will get you an image, but a noisier one. What most people don’t know is that LCDs can modulate the phase or the amplitude of light, even though we normally see them in displays where they are being used to modulate amplitude (to vary the brightness of a pixel). To modulate the phase, you just remove one of the polarizers! If you’re really smart, you can figure out how to modulate phase and amplitude simultaneously with an LCD in a maximally error-free way to get the best image you can possibly get with the least amount of noise.
It sounds to me that what Daqri are doing is modulating the refractive index of a medium in a spatial way, which is effectively altering the speed that light travels through each part of the medium. My guess is, they are doing this with acoustic waves in some manner, generating an interference pattern in a solid crystalline medium using arrays of transducers along the edges. This then creates high and low pressure regions within the crystalline medium which the laser light passes through, which will diffract the light and produce your holographic image.
So there’s devices now, that can produce real moving holograms, from an LCD screen? Or something a lot like an LCD screen. Any links? I’d really like to see video of it. Playing it back as video means you can render it all beforehand, so you wouldn’t need as much processing power. I didn’t know real moving holograms existed!
Even if it doesn’t have a full 360 degree viewing range, as long as it has some, as long as you see a different image from different angles, that counts in my book.
Looks awfully flimsy for a cured UV ink.
Shaky, out of focus, nothing is exactly seen in the video. Looks like there is only one reason why the video was shot like this. That shit does not work.
I will believe it when they print a calibration cube or benchy
This isn’t 3D printing. It’s basically a CNC machine.