Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Magic Leap One Video – Diffractive Waveguides Confirmed
At Recode’s Code Media conference on Feb. 13, 2018 (a few hours ago as I write this), they live streamed an interview with Ronny Abovitz, CEO of Magic Leap and NBA commissioner Adam Silver. The interview included a prerecorded video with Shaquille O’Neal, but the big star for me at least, what that Shaquille was wearing what looked like a real Magic Leap One headset. I thought I would do some “instant analysis” while the information is fresh.
Blocks About 85% of the Light – More VR than AR
This article is following up on my prior articles related to Magic Leap’s Rolling Stone reveal. In the crops pictures below from the video frames that follow, I have included the time in the video where the frame was captured (gray time code) with my annotations in red.
The first and most obvious things to notice is the darkness of the glasses. There is not even a hint of Shaquille’s eye visible. I was able to find a point (indicated by the arrow) where you can see through the lens to the background. I reversed out the “gamma correction” to convert to a linear value to make an approximate calculation of how much light the glasses transmitted, and it came out to only 15%. These would be like wearing dark sunglasses, the type you would wear on a sunny day. Only this headset is meant to be worn indoors. While the glasses looked dark in the Rolling Stone article reveal, they were doctored.
One must ask is 15% transmissive really “AR” or is it 85% VR. The common number I hear for being “reasonable transparent” is at least 80-85% transparent. You also have to consider that the peripheral vision is almost completely blocked by the large frames around the lens opening.
Confirmation of Diffraction Grating
As I have been writing for over a year, what Magic Leap hypes as “Photonic Chips” are just “diffractive waveguides.” In the Shaquille video, there is clear evidence of diffractive waveguides as he moves his head around, you the exit port diffraction grating catch the light (see left and right).
For reference, I have included below pictures of Vuzix’s Blade and Hololens exhibiting the same light catching and reflecting effect. The coloration of the light is caused by the diffraction grating effect. Both Vuzix and Hololens are transmitting about five (5) times the real world light Magic Leap One. Note how you can see the person’s eyes (Paul Travers’, CEO of Vuzix) through the Vuzix Blade glasses.
Also for comparison, the Lumus waveguides (left), which use a stack of partial mirrors in their waveguides rather than diffractive waveguides. The Lumus waveguides appear to be are even more transparent. They also don’t exhibit this “catching and coloring” of light the way the Magic Leap, Vuzix, and Hololens diffractive waveguides do.
You will also notice how much Lumus, Hololens, and Vuzix allow a person to use their peripheral vision. Additionally, Lumus [correction] and Hololens have enough eye relief that a person can easily wear most normal glasses.
The Audio Was Marketing Hype and Buzzword Salad
The rest of the video made my ears bleed. It was the usual Rony doing his hype and buzzword salad trying to fit in as much hype in as he could without really saying anything. Rony talks about still being able to see the real world, but what he leaves out is that it is going to be about 85% darker. As for having the NBA Commissioner endorse Magic Leap and a testimonial by Shaquille, I would no more go to them to give an opinion on AR glasses as I would have them do open heart surgery.
The one exception to just buzzwords and hype was when Rony talked about the eventual price. Rony said that the eventually he expected to have a range or tiers of products and he said that “low end someday” will be about as expensive as a high-end cell phone (ala, iPhone X). This would seem to suggest that the early enthusiast version is going to be in the $2,000 or perhaps more range.
Great article. BTW regarding transparency have you take into account the fact that Shaq has a much darker skin color than the others who are white guys ? This coupled with the fact that the ML glasses seem to hug the eye socket blocking out outside light much more than Vuzix, making it seem darker. I am saying this because if you look closely at the image you have captured at 6:52 there is a tiny sliver of the wall behind Shaq visible through the glasses near his right eye which has a luminosity almost on par with the rest of the exposed parts of the wall. Also if you look at the old pics of Rony holding up his “photonic chip” / waveguide thingy its almost completely transparent (unless he was fooling us by holding up a piece of broken glass). Also I can show you several pics of ODG and Vuzix glasses looking very dark in certain lighting conditions.
And yes its sad these look like diffractive waveguides… Rony needs to put out specs fast or else he is simply losing the marketing battle. The optics is beginning to look bad – no pun intended.
I think he better show the damn glasses to YOU! ASAP !
First, the white parts of people’s eyes are pretty much the same, and you can’t see the whites in Shaq’s eyes.
I looked at what I thought were comparable regions. I checked several frames of the video and the frame at 0:51 in the article is what I used. The method is admittedly very approximate as you have to guess the “gamma curve” of the video capture device to back out the linear value (I used a gamma of 2.2).
The Waveguide itself (Rony’s “Photonic Chip”) probably is relatively clear. I would assume the added darkening is to block out the real-world light so you can see the virtual image. Most diffractive waveguides are very inefficient, so they are not very bright unless you have a very bright display source. Going to a very bright display source means adding a lot of heat and heat management as well as potentially more stray light problems.
The ODG R8 and R9 are terribly dark. They only let through about 5-10% of the real world light. They are starting with an OLED that does not put out much light and then have a very inefficient optical system. I would not consider them acceptable. Vuzix is another matter, they claim to be more than 80% transparent, and my measurements suggest they are in that range and thus you can see the person’s eye’s in the picture. Lumus (using a non-diffractive waveguide) is even more transparent, but they have a very bright light illuminating an LCOS panel.
Everything in the patents have indicated that Magic Leap was using “diffractive waveguides.” They will also have all the known problems with them.
I would love to test them out and objectively report on them. I have every expectation they will look like Hololens but with a bit wider FOV.
I don’t mean to be an ass, but I don’t think you can make any kind of accurate predictions of a surface that has a combination of a reflective surface and opaqueness without much more information. It’s like trying to judge the opacity of a 1 way mirror from the wrong side of the glass. You’re missing a vital point of data.
You are wrong. It may not be perfectly accurate, but it is not that far off. You are looking for a part of the image that is not specularly lit and you can see both through the lens and the background behind it. It would be much harder (and in some cases impossible) with mirrored glasses but these appear to be only tinted. This is part of why I put up some comparison photos of other people wearing glasses. You can definitely tell the Magic Leap glasses are dark.
if you look at the left image of shaq in Confirmation of Diffraction Grating section you can clearly see the back wall in the left lens. Looks way more than 15%. closer to 50% I’d say
I measured it and backed out the “gamma” (assuming a typical 2.2 gamma) and it came out to about 15%. You can also compare it to the Vuzix, Hololens, and Lumus glasses. Magic Leap’s glasses are much much darker.
Karl what do u think r the chances that they won’t have eye tracking with ml1…..if we go by the patent with the open back instead of a headband ……it looks like it doesn’t have it
plus the fov is not big and it’s demanding especially on mobile hardware where the important thing is battery …..and it would add additional cost for something that can be done just with gaze like with the hololens
Those lenses are guaranteed to be polarized. If the camera setup wasn’t right, it would show up much darker than it would appear in real life. As others have said here, there’s not really a reliable way to gauge transparency from a video.
What are you saying is polarized and how would that affect things? I have been working with polarized light displays for about 20 years and taking pictures of them.
If the camera taking the video was polarized, the darkness of the lenses would vary all over the place as the angle of the lenses relative to the camera changed. If the camera is not polarized then it is light blocked by the display lenses.
The method is not ideal, particularly if you cannot control the setup. And you have to pick where you can see a moderately bright background through the lens. It is good enough to get a good idea (plus or minus about 30%). Have you tried it? Why is it that you can’t see the whites of Shaq’s eyes if they are not blocking a lot of light?
Finally, VR and AR will be the same thing: rendering 3D objects in the 3D scene. The separation of these terms is just because our display devices are technologically limited.
Very very interesting article, as always. So, basically, their photonic bla bla bla tech is just very similar to the one of HoloLens, but maybe even worse. But I’m still asking… if a lot of investors are pouring money into it… there should be a reason… have they something in their R&D department that is so special?
All I can say is that big companies make big mistakes all the time. You just know a lot more about their successes. I have it from many sources that several groups within Google thought/think that Magic Leap is/was a scam but that the VC guys invested anyway. Once “Google” invested, then the others fell like dominoes.
I like to say that this is a bit like looking at stocks. You can look at the charts or look at the fundamentals. I’m looking at the technology fundamentals via Magic Leap’s patent trail and other information available. It’s possible, but not likely, that they filed patent applications on the stuff they are not doing and didn’t the stuff they are doing. I can’t see them risking their NDA’s and contracts with so many people would keep things from leaking out for so long so as not to file patent applications on “the really good stuff.”
Do you have opinions on which glasses that you have tried that seem to be the most promising utilization of the tech? and are any of them public?
So the Vuzix Blade comes in three shades, clear, dark and transition. Is your 85% see through based on the clear shade or the dark shade?
I think that this is based on the most transparent option.
It seems the smarter move were the investors in Vuzix 245M with a much better value vs the Magic Leap investors at 5.5B. why would so much money go into Magic Leap when companies such as Vuzix or Lumus are around. It seems the first big money was chased by other big money and the blind were leading the blind into magic Leap. the original investors have not put in anymore money since.
[…] at Sarah’s work plus the second of the Magic Leap Recode video interview that I analyzed, I found I liked her overall approach but it was missing some details, and I think, based on the […]
[…] Magic Leap One Video – Diffractive Waveguides Confirmed – Karl Guttag on Technology […]
[…] also redid my earlier calculation from the Shaq video and found the light transmission of the real-world to still be in the 15% range. So nothing new […]
[…] ML1 blocks most of the real-world light and block much of a person’s peripheral vision. Back in February 2018 based the video with Shaq, I was able to estimate that the ML1 was blocking about 85% …. This estimate agrees with instrument measurements that it blocks about 83% of the light at the top […]