As technology accelerates, some of the most amazing things are taken in stride as quite normal. Arthur C. Clarke’s famous quote, “Any sufficiently advanced technology is indistinguishable from magic,” is still true, but these days it’s harder to advance far enough to induce those magical feelings.
Design an object and print it out in 3D and no one blinks an eye, all you have to do is design it in some CAD program, render a 3D model and you’re set. Well, maybe instead of designing it, you scan something with a laser scanner and the model appears in the computer, that’s pretty cool, but still, laser scanners are old hat, too, nothing new there. Hmm, … OK, so how about this? You see something you would like to build, you pull out your smartphone, walk around taking a few photos of the object, send the pics up to the cloud and the cloud crunches the numbers and sends back a 3D computer model you can tweak to your heart’s content and then use the model for whatever you wish. Are you impressed yet? You should be. If you’re not, re-read that last sentence.
Autodesk, the AutoCAD folks, have a new service in beta called Photofly where you can do that. It’s not yet 100 percent accurate, but it’s pretty close and knowing how fast these sorts of things are moving, accuracy is only a matter of time. If you’re thinking there may be some issues with this kind of capability, you’re not alone, but the tech is here and it’s pretty impressive.
We’ve seen comments here from some who look at computer concept models and say, forget the computer model, until it’s real it’s not worth talking about, however those commenters should think seriously about technology like this. The ease of movement from real to computer and back to real is becoming easier and faster every day.
Is this magic? No, it’s here now and hands on guys need to stay on top of this technology. Watch the video below for a brief intro to what this is all about.
Link: Autodesk Photofly
Video below:
WillyP says
Truly magic. đŸ˜‰
Now the issue with the 3d model bikes isn’t that they couldn’t be built… The issue, for me at least is that we don’t know whether they will be built. Turning a 3D model into a physical model is cool, but it still is not a real bike that can be ridden, that someone can say, ‘it rides and handles great’ or ‘the ergonomics are a little odd’ or whatever. Taking photos and turning them into a 3d model is a great step forward, maybe someday we will have 3d printers so advanced all you need to do is add gas oil and a battery and off you go. Imagine going to your local print center and printing up the latest creation from some bike designer, in your choice of color, oh, and while your are there, a new toaster, and a dual quad carb setup for your 1967 Chevelle SS that you printed out a few months ago.
Science fiction, or magic?
Paul Crowe - "The Kneeslider" says
I tossed in the 3D ATV above to show what Photofly does, but no one is suggesting printing a complex device, at least not yet. The point is taking even a single component and making a 3D model. It can be adjusted, tweaked and used for any manufacturing purpose. If you have a model, you can machine it, too.
WillyP says
No, but who knows what might be possible in the future? Obviously your not going to produce a running motor from a picture of a motor as the camera only sees the outside of the motor. But could a car be produced from a file, on a 3D printer? Not with the current technology, as the parts produced would not be of the correct material, heat treated or finely machined. But who knows what future developments could produce. A few years ago we would have laughed at the idea of a 3D printer. Now you can watch a real, working adjustable wrench be printed, no assembly even. (search for the video on You-Tube)
Matt Gerard says
The problem with concepts like this is that you aren’t CREATING anything. You are just duplicating something that already exists. That is taking the focus away from using your imaginations to creatively solve problems or coming up with cool designs that no one has ever seen before. As far as originality goes, this is a problem.
On the distribution side, if you take WillyP’s direction on going to the local printer to take delivery of your new ride, that’s a different story. It has little to do with the input method of taking a few pics and duplicating something that someone else took the time and knowledge to create and design, though.
I feel we need more tools to help people get more creative and come up with their own ideas, instead of giving more tools to copy things that already exist.
But its still effing cool…
Matt
GenWaylaid says
Sometimes duplicating something that already exists is enough. Just imagine how much easier it would be to restore obscure or unique vehicles with this capability.
While anyone could grab their digital camera of choice and start collecting images, there’s probably quite a skill curve to producing a set of images that will give a good, complete model. Still, with memory as cheap as it is and all the heavy computing happening in the cloud, there doesn’t seem to be any penalty to overkill.
We won’t be able to duplicate carb bodies, liquid-cooled heads, or anything else with internal passages until someone can figure out a way to integrate x-ray or ultrasound data. Or carve your own in the digital model, as complex as you want. 3D printers can do incredible stuff with internal voids.
WillyP says
No, but custom bike builders aren’t creating from scratch either. So lets say I take some pics of my bike and make a 3d model, I have just saved a ton of work, and now I can edit that model to represent changes I want to make. I could even do things like take the bike apart and photograph the components, or parts even, to help design custom parts, I could explore things like exhaust routing, etc… And further down the road, I could see people sharing models, so for example if I wanted to put a ZRX1200 motor in my ZG1000, I could do it digitally first. I would just pop into the ZRX forum and download a few files.
jar says
There is some value here – depending on how well put together and how accurate the resulting 3D model would be.
Making the leap that at some point the technology can actuallly deliver well developed accurate 3D data sets from collected images, reverse engineering practice immediately seems to become less costly, easier, faster, and more accurate – no more “caliper close” models of parts – snap some images of a sweet connecting rod shape, and pow, there’s the 3D model to start manipulations from. Laser scanners are great tools – but between the hardware and software, for something useful (say Zcorp/Creaform’s Handyscan), a guy is looking at $65k for the hardware, and that again for something like Geomagic or Polyworks on the software side, so call it $100k (white light systems can get even scarier). There are others, Artec scanners and Leios mesh software, that seem promising and drop the $100k to something a bit more manageable, in the realm of $20k, but still spendy.
As for no “creation” here, I would have to disagree. I would think that a lot of guys that pull off a one-of-a-kind design, might do so through skill sets that don’t include “CAD” modelling – they just build it (“just” here intended to generate a smile, not slamming guys who build with their hands, not on a tube). Being able to send photos of their unique creation into a cloud, and in return gain an accurate 3D model gives a guy the ability to take that design any where – patent office, machine shop for multiples, or communicate the thought to a potential investor easily – any where he may want to communicate the thought accurately and quickly.
3D CAD is how the manufacturing world speaks. The ability for anyone, anywhere, with a few photos to acquire such speech in order to entertain desired dialogue is a valuable thing indeed.
We move as the data moves. That simple. No data, no forward progress.
B*A*M*F says
In college, I used to scan fossils for with a system that took photos of an object on an indexed turntable. It then processed the images much like this. The ability to snap photos from a phone without a calibration grid is impressive. The system we had did not handle concavity well, and I don’t see how a photo based system really could. Still, it’s impressive.
Leo Speedwagon says
It’s easy to reproduce anything in 3D, you don’t need a fancy machine. Anything from computer chips, engines even Apple stores can be reproduced without much effort – just send it to China… I’m sure they will reproduce this 3D printer ntitizewithin a short space of time.
akumabito says
I think technology like this could be extremely helpful in restoring vintage bikes. Just try and restore a 1920’s or 1930’s motorcycle from a long-lost and obscure maker by using off-the-shelf parts. Chances are you wouldn’t be able to. Some pieces will just have to be machined or cast from scratch.
Now imagine you can take a bunch of photos of a damaged part, digitize it into a 3D model, then tweak it to original specs. You now have a solid basis to recreate the part. Tweaking a 3D file is a lot easier than designing from scratch. A lot less time consuming as well. So less time learning to use software, more time building and tinkering!
FREEMAN says
I think the process is pretty incredible. I’ve seen a similar thing being done in real-time with a pair of the xbox kinect cameras.
Slacker says
This is very cool tech. I need something like this. It would be invaluable for a bike build. I’m currently going back and forth on a design for a gas tank on my bike, and being able to look at a photorealistic 3D model of it would help me finally make up my mind. Plus, how great would it be to start with just a frame and engine, and slap the parts on in 3D from there?
B50 Jim says
The next step (well, maybe the next 12 steps ahead) is to actually produce the scanned item using photolithography. We’re moving closer to the replicators as used on board the Star Ship Enterprise. Sooner than we think….
Keith says
Makerbot is out there. 8^) Nice concept, NOW print metals and allow me to print the correct metal and you’ll have me in line for one. Not asking for built items…because nothing would be torqued properly. I have a lot of old 2smoke part that need making.
Azzy says
Thanks for the link. Showed to a few buddies and confirmed that this has it’s roots at a company that i started my drafting career at . Very cool, cant wait to play with it (again).
Orlando says
So… Downloaded Photofly and had a go at producing a 3D model with it… The idea is very cool, but there are still A LOT of stuff to fix to make it foolproof. As for making a precision part from one of it’s models… Well that’s still many, many years in the future. There are so many little things that will throw off the precision of the model, that no one would think of developing software like Photofly with precision in mind. You’d need to control lens distortion, film grain (or pixel size), lighting, set precise photo angles etc… And then the mesh you get is way too messy to fix with any kind of precision. For prototyping you’re still better off measuring stuff and making it from scratch, at least you can be sure it’ll be precise. I do 3D work for a living, and think Photofly is very cool for other stuff. Like making a low-res prop for a scene, but I don’t think it’s precise enough for anything that needs animating, or any extensive tweaking.