Augmented Reality Welding Mask

This welding mask uses HDR video to ensure you see just how sloppy that weld was.

The creators of the MannVis augmented reality welding helmet have come up with a set of complex algorithms which allow real-time processing of HDR videos at 120 frames per second — in stereoscopic 3D, no less. The process creates an effective contrast ratio of 100,000,000:1 which is far beyond natural human capability. This incredible contrast ratio is used in an augmented reality view to show the welder inhuman detail as they weld.

The MannVis Augmented Reality Welding Helmet

(Thanks, John Courte!)


  1. Wow, this is potentially huge. I used to work for our family business, which involved TIG welding lots of stainless steel. My dad always had a hard time training employees to TIG weld, and the welder flash was one reason why. As you can see in the video, even with the mask, you can’t really see what you’re doing because of the bright light. It was one of those things you just had to instinctively figure out, because you had no idea how good the weld would be until you were done.

    Another use would be in the construction of nuclear power plants. I had a shop teacher explain to me years ago that the thick walls were welded by stainless steel arc welding. They X-ray each weld to check for bubbles or other imperfections. If there are any flaws at all, the piece is thrown away and they have to make a new one. Technology like this could cut down on those errors and save a fortune.

    But I do wonder how well that modified helmet protects the welder’s eyes. Still, it looks badass.

      1. It’d be hard NOT to be better, since they can justt have a huge plate of steel in front of their face. :)

  2. Holy smokes. I’m not sure it’s apparent to people who haven’t welded for a living how revolutionary that is. Normally through a welder’s helmet, you can only clearly see the arc or flame itself. You can dimly see the puddle of melted metal, and you can see maybe a one inch square of the workpiece by light reflected from the arc. Without the arc actually lit, you can see nothing at all, which is why you see welders set everything up with their helmets out of the way, and then “nod” the protection into place when they’re ready to go.

    Having that kind of view of an active welding operation would make it a whole different skill. The binocular headmount 3d is nice, but I’d be more than happy to weld with my hands under a metal shield, while watching the cameras’ output on a screen over elsewhere, if only I could get that image of the bead being formed. If this can be done for the price of four webcams and a laptop, I think a lot of weld-manufactured products are about to get a lot better quite soon.

      1. They’re all-or-nothing. Once you strike the arc, they darken, and your FOV is just as restricted as using straight shaded glass.  All they’re good for is saving you the trouble of opening/closing your hood as you work.  

        (Which is nice, don’t get me wrong.  But all things considered, it doesn’t make as much of a difference as you might think.)

  3. Hydraulophone? Oh, hydraulophone!

    Also, the last section of the video shows that humans are no longer needed for welding.

  4. I’m not familiar with welding, so I wonder: 

    with these visuals, can’t (human assisted) robots do a better job?

    1. Robots are already used for lots of different types of production welding. The main place where robots don’t work is in structural welding (bridges, skyscrapers, etc…). Can you imagine what would be involved in setting up, aligning, etc… a welding robot just to do one joint, then having to move it to do the next one? Any time there isn’t a single, repeatable job that can be done in a factory, robotic welding doesn’t make sense. By that same token, the other place where robots don’t make sense is in custom work.

    2. Where I work we build earth moving equipment and most of it is welded by hand.  It’s not that a machine couldn’t do it better, but it’s the fact that a person (or persons) can weld up an entire section in one place at one time.   Multiple welds (at multiple depths and thicknesses) in varying angles and positions…a robot could be built to do it, but I think the cost would take years to recoup.

      Now we do have automatoed CNC style plasma cutters that take the raw plate steel and turn it into the basic shapes that will be formed and welded up…

  5. I love how they get creative towards the end with driving safety and such.  The potential applications for this are absolutely mind-boggling.  The ideas are coming to mind faster than I could ever hope to write them down.

    1. Also, make the youtube video show a frame from 3:10 or so and I bet everybody will watch the video and not miss out on the awesome.

  6. Wow. Just wow. I am completely unable to weld because my eyes don’t seem to adjust well to the arc even with a good helmet. This is awesome.

  7. Seems like you could home-brew something cheap and almost as useful by rigging a single HDvideo camera with a filter and feed it to a small tablet in front of your face.

    1.  I think you’re missing the point. A camera has less dynamic range than a human eye. So he is taking the output from several cameras; some looking at the bright parts of the image, others the darker parts, and combining them into an HDR image. I think. Either that, or maybe he’s switching alternate frames though an LCD filter to get wider range, then combining those.

  8. Err… this seems to me to be a (probably very) expensive pain in the arse, and depending on the situation, potentially dangerous. Firstly, it looks heavy. secondly, those cables trailing out of the top WILL get caught on something in conditions other than ideal bench welding. If I was bent at an awkward angle in the nook of some large bronze (as has happened before), or lying on my back in a steel structure, I wouldn’t want all that stuff trailing about. 

    Auto-darkening, adjustable filter visors are totally affordable these days, and frankly, they work incredibly well. If you keep your visor clean, then even a low power tig arc is enough to give you plenty to see with (although I understand how someone such as @Robert might find it useful). 

    1. Don’tya think this might be a prototype? Make it work first, then make it smaller/tougher/prettier. Once you eventually buy one of these, I’m sure the problems you highlight will have been addressed.

    2. I agree with Gyro, it’s a prototype with off the shelf stuff to make it work.  Price and durability I think would be the biggest issues (along with a power source/duration).  If you could get this down to $300 for a full/wide view setup that weighted at most 30% more than your standard auto dimming helmets now you might have a winner.    Given the fact that small LCD screens with the resolution needed for that short of a viewing distance are still expensive I don’t expect to see this on a shelf for many years.

    1. Honestly I think a separate power pack wouldn’t be a bad idea.  You wouldn’t need thick cables or anything that couldn’t be run under your shirt.  But it’s be nice to keep the helmet weight down and wear a couple of li-ion power cells on your belt.

  9. Didn’t read all the info about the article, just went straight to the video, saw the chin and lips of the guy in the video, and said “Hey, that’s Professor Mann!” Then he said “EyeTap”, and I got wierded out because I recognized someone just by their lower jaw :/

  10. Seriously this is some technology. I am a big fan of Augmented Reality, Its like a daily routine for me now to see whats news in Augmented Reality after I experienced an augmented newspaper in India by AdStuck Consulting. Then, it was used mainly as a marketing tool for various brands. But, using AR on a welding helmet is too good. This technology has no limitations, Good Job !

  11. I was always frustrated in metal shop class [circa 1985] because the darkened glass in the hood is too dark to see what you’re doing until you strike an arc in front of you. So you basically have to start out blind.

    Also because the frickin shop teacher was a former wrestling coach and liked to wrestle with guys standing nearby while you were trying to blindly arc weld for the first time. Christ, what an asshole.

  12. Are any of you using a modern auto-darkening Helmet? My Miller Digital Elite isn’t a chore to see through at all? I’m apparently missing something here, maybe I should watch the video again, but this seems like a million dollar answer to a question that hasn’t been asked. Seems like they’re trying to invent a market for a product that’s unnecessary. 

    1.  Once you strike the arc while wearing your auto-darkening helmet, there is no way you see anything near as good as what that video was showing.

      Auto-darkening helmets solve the problem of being able to see better before you strike the arc — not after it. After you strike the arc, an auto-darkening helmet is exactly the same as any other helmet out there. That is, you just see a tiny puddle of light.

      With this thing, it’s like you’re welding without a helmet at all, in nice lighting, and with an arc that is easy to look at.

  13. in the 1980’s a friend of mine, Mike Kan, did this very thing — he had a transmissive LCD in front of a (VHS) camcorder lens, and hardware/software that darkened pixels that were saturated, in realtime. I personally saw it operate, and he used to take it to various hacker events in the San Francisco area.

    I distinctly remember his main demonstration — a newspaper held next to a 100 watt light bulb. Both the newsprint and the lettering on top of the bulb were well within the dynamic range of the sensor.

    I’m sure MannVis is superior, but the technique isn’t new or even novel. Which is absolutely not to disparage it! OMG it’s an amazing device, developing it into a fully functional deployable device is no trivial task.

    > auto-darkening helmet

    yes, but they simply darken the entire viewing area. This is “pixel by pixel” or equiv.

Comments are closed.