New Google Glass Design Likely Uses a Transmissive Panel

Google Glass display options scaled to a dime
Kopin Panel, Microvision Optical Engine, Google Prototype CF LCOS, and Transmissive Optical Engine (to approx. the same scale)

As a follow-up to my last post, I though I would show why Google Glass is most likely using a transmissive panel.  It all comes down to size and shape.

Shown to the left is a Kopin transmissive panel more than capable of the resolution shown in the Google Glass videos and the picture I found on-line happen to have a dime in it and I use that dime to scale it to the same size at the Microvision laser beam steering (LBS) with its “dime picture” in the second image.  I roughly scaled a Google Prototype with a color filter LCOS panel to the same scale in the third image.  The Kopin panel is only about 2mm thick but it does require optics so I approximately scaled the figure 8 from Patent 6,747,611 filed by IBM in the year 2000 which shows a near eye transmissive optical engine and gives a Kopin panel as an example.

The Microvision engine is for a projector and does not include the “wave guide” that relays the image out to the eye.  You are also looking at it from the top down but it is about 6mm thick which is similar to the others in thickness.   Part of what makes the Microvision so big is the need to combine and aim the 3 independent lasers at a single mirror as shown (in an older post I showed the combining path).   The engine is on the order of 5 times too big to fit into the space available in the Google Glasses, and that does not include the electronics for Microvision’s LBS which take about as much volume as the optics).

Next we have the color filter LCOS which is much more compact than Microvision’s LBS but has an awkward “T”/”L” shape to it caused by the orientation of the beam splitter with the panel on one side and the LED on the other.  As I wrote in my prior post, this would not fit in the barrel shape of the newer Google Glass design.

Lastly, the IBM patent has a figure that shows a transmissive panel optical engine that looks remarkably similar to the Google Glass that have been seen.  The optical path is straight through and comparatively compact.  There is an adjustment knob (2600) that enables the apparent focus point (according to the patent) to be adjusted from about 18 inches to infinity.   The Google Glass are said to be set for far vision (near “infinity”) and therefore dispense with this adjustment.

Another thing to note is that there is only an LED, panel and a single lens to generate the image plus the beam splitter (doing the function of the thinner wave guide used by Google).   This is a relatively inexpensive device as the LED, a low resolution transmissive panel, and lens combined cost on the order of $10 (and probably less in high volume).

IBM Patent from 2000 to Google Glass ComparisonIn the category of “everything old is new again.” look how closely the Fig. 8 (copied at the left) from the IBM patent filed in 2000 looks like the Google Glass of about 13 years later (left below).  The main difference is that the “computer based device” (today a cell phone) is now wirelessly connected.   A feature shown in the IBM patent is a sliding light shield to support viewing images without the distraction of the background.  Google Glasses would require looking at a black background to clearly see the image in the transmissive wave guide.

Google’s design “cops out” and requires a nose bridge which others including the IBM patent and Golden-i avoid.   The nose is very sensitive to any weight on it particularly over time and it interferes with glasses.  Google has said that the device can be attached to a person’s glass frames but this is very problematic with the variety of frames on the market and the added off-balance weight.

The point I would like to make (again) here is that the display technology has been available to make Google Glasses for over a decade and as my prior post on virtual reality displays pointed out, the limiting factor is the use model (how you use it) and is heavily limited by how you control it.   I don’t see it as practical to have people talking to their devices and looking shifty-eyed and blinking, not to mention looking like somebody who escaped from a lab.

Maybe someday they will add gesture recognition so you can type on a virtual keyboard but I don’t know of anyone that has perfected this technology yet.  Also the images that Google has shown to date are pretty low resolution (on the order of only 320 by 240 pixels) and only fill a small part of one’s vision.   I don’t see people doing a lot of internet browsing with the current Google Glass.   Then we have the privacy issues as in  when someone looks at you shifty eyed through their Google Glass, are they signalling to the computer to look up your information.

One last thing, believe it or not I’m not trying to be negative to about Google Glasses, I’m just trying to relate my experience and knowledge of near eye displays.   I think even some people associate with Google Glasses are playing it down a bit trying to get people to understand that they are still looking for how people will use it.  Maybe someday they will have a  high-resolution color display that fits in a contact lens, selectively blocks out the real world, and picks up brain waves to control it, but it looks to me that that day is a ways off in the future.

Karl Guttag
Karl Guttag
Articles: 244


  1. Another great post, Karl. Peeling back the onion only added to my piqued curiosity! Any thoughts on the following?:

    1) Why might the Google Glass prototype be using color filter instead of field sequential?

    2) If I read you right, it appears that Microvision’s solutions are lagging in miniaturization, no?

    3) Any reason you haven’t mentioned the other LCOS vendors who can at least match what Google Glass is likely to achieve in its first iteration?

    Thanks for all the great knowledge!!

    • Thanks Mark, those are some good questions that I will try and answer:
      1) Why might the Google Glass prototype be using color filter instead of field sequential?
      A) I think it is because they were aiming at a small size with low resolution and very low power. Spatial color takes 3 sub-pixels which can impact resolution. But Google Glass is only putting up a small low resolution in part of the field of view so high resolution was not a requirement. With spatial color the panel power is very low since there is not a lot of “activity” changing the pixels the way field sequential color does. Since the brightness for near eye is VERY low (sub 1 lumen) the panel power is more important than the light throughput losses. With spatial color they can use a single and very cheap “white” (really blue stimulating a phosphor) LED. Then you have the size and cost advantages of not having to combine 3 colors into one. The transmissive panel (as I suspect they are now using) further simplifies the optical path.

      Note, however, all the tradeoffs change if the resolution were higher or with higher brightness. Color filter devices have a hard time keeping up with resolution due to the 3 sub-pixels and the color filters and the structure around each pixel block a lot of light. But in the case of a low resolution near eye device color filter can be the most cost, size, and power effective solution.

      2) If I read you right, it appears that Microvision’s solutions are lagging in miniaturization, no?
      A) Laser beam steering has a lot of issues and they have particular problems with respect to near eye. A fundamental size problem is that they need a lot of space to start with 3 lasers, each in their own package, shape the laser beams and combine them, and aim the 3 beams on the mirror. Then there needs to be a path in and out of the mirror which takes space. The optical path required for color filter devices is much smaller.

      3) Any reason you haven’t mentioned the other LCOS vendors who can at least match what Google Glass is likely to achieve in its first iteration?
      A) The point of the recent posts was to identify what was in Google Glass. I figured out that the WERE using the Himax panel in their prototype but I also realized that this would not fit in the newer device they are now showing. I don’t think they are going to use reflective LCOS, even color filtered, due to size and the “T” shaped optical path. If in future version Google goes to filling the whole field of view which would require high resolution then this might change. As it is right now, I think they are going to say with transmissive panels which seems to point to Kopin, Epson, and other that have transmissive color filter devices.

  2. The LBS discussion is irrelevant. Nobody is using a Class 2 engine for a handheld projector because of safety concerns, so it’s unthinkable that LBS would end up in a near-eye product.

    By the way, Microvision are now claiming 35 lumens for its projectors without stating the safety class – likely because it is Class 3.

    • “The LBS discussion is irrelevant. Nobody is using a Class 2 engine for a handheld projector because of safety concerns, so it’s unthinkable that LBS would end up in a near-eye product.”
      I could agree that it is “nearly irrelevant” as it is a ridiculous to use it in terms of size and cost (plus a number of other big problems), it has been proposed as a solution for near eye for over a decade and there are some Microvision stock pumpers trying to insinuate that it might be in Google Glass. I was trying to put a stake through this rumor’s heart as it physically couldn’t fit.

      In thinking about the post, I realize that I forgot to mention OLED displays as it would also at least fit. I think OLED is unlikely because it would be “overkill” in terms of performance and cost.

      “By the way, Microvision are now claiming 35 lumens for its projectors without stating the safety class – likely because it is Class 3.”
      Microvision has a history of saying things that turn out to be irrelevant to a real product. My understanding is that they are Class 3 at 25 lumens. But then what difference does it makes, they made all these pronouncements at CES last year for their so call “720p” (“so called 720P” because nobody has really measured it with a test pattern) that has never seen the light of day in a product.

      It is my opinion that the main “product” of Microvision is printing shares of stock to sell people. LBS is one of those technologies that may seem like a good idea on the surface but has a lot of underlying problems. There may have been a time when those at Microvision didn’t realize the issues and limitations and I can’t fault them for that in their earlier days. But as my blog pointed out in late 2011, Microvision has been at the least implying a much rosier picture that was rational.

      This is particularly true with the issues surrounding green lasers. Just about everybody in the industry, particularly the laser makers and their potential customers, knew that direct green lasers were going to be way too expensive to put into any kind of volume product in 2012 or 2013 (or 2014 or 2015 for that matter) but Microvision tried to get people to believe otherwise and had the audacity to call me a “False Soothsayer” for calling them out on it. Still it looked liked they were able to convince enough suckers to keep buying shares of stock. Well, here we are about 1 year and 4 months later and where are the low cost green lasers?

      And by the way, even if direct green lasers were cheap, there are so many drawbacks to LBS that it still wouldn’t be a viable/competitive product for the reasons outlined in my prior posts.

  3. Hi Karl,

    I’ve provided more info here (since the comments section doesn’t allow for screenshots):

    In short, I had another lengthy discussion with HIMX today. They exuded a good deal of confidence as it pertains to their ability to deliver the capabilities you believe Google Glasses will require (while being very careful not to discuss specific customers and prospects).

    Let me know if there’s anything you’d like me to ask them. I’ve rather enjoyed digging around for information to confirm or refute my hypothesis…and you’ve certainly kept me digging!

    Thanks again for your expertise (and an entertaining week!),

    Mark G.

  4. Hi Karl,

    Overnight, Digitimes reported that Himax has won a sole-source component deal on the Google Glass project.

    The article doesn’t cite a source though. It certainly wasn’t Himax — they declined to comment:

    “Himax Technologies reportedly has landed orders for LCoS (liquid crystal on silicon) solutions from Google for its Google Glass devices, and shipments of LCoS solutions are expected to start gaining momentum in the third quarter of 2013 as Google intends to move forward the release of Google Glass to the end of 2013 instead of early 2014 as originally planned, according to industry sources.

    Himax will be the sole supplier of LCoS solutions for Google Glass at present, and the shipments will ramp up revenue growth at Himax in the second half of 2013 as well as in 2014, the sources indicated.

    Himax declined to comment on market reports. But the company revealed at its investors conference held in February 2013 that it has been developing headset-type pico projection technology with clients and that revenues generated from devices other than LCD driver ICs will serve as growth drivers in the year.

    Himax outsources the production of LCoS chips to its subsidiary Himax Display, while subcontracting IC backend testing and packaging services to ChipMOS technologies, the sources noted.”

    Another layer of the mystery unfolds. Fun stuff.

    Kindest Regards,

    Mark G

    • Mark,

      I believe that the Digitimes article is most likely simply an Echo of your posting on Seeking Alpha. In my experience Digitimes seems to string together a bunch of information mixing a bit of news/rumors (in this case your S.A. post) with information they already know (old information on Himax Display and comments on when Google Glass will ramp up in this case).

      Anything is possible, and it could be that your article sparked them to investigate more, but it is at least as likely that they just ran with what you have been writing on S.A. and the followup instablogs.


      • No doubt. They certainly didn’t go out of their way to name a source. The wording is too detailed to be an echo of my article, but to your point, the article surely sent them searching for a story of their own. Source notwithstanding, their article pretty detailed though:

        “shipments of LCoS solutions are expected to start gaining momentum in the third quarter of 2013 as Google intends to move forward the release of Google Glass to the end of 2013”


        “Himax will be the sole supplier of LCoS solutions for Google Glass at present”

        Nonetheless, I’m with you. My response (in queue with SeekingAlpha) states that the Digitimes article should be taken for what it is until a press release or SEC filing provide official clarity.

        Thanks Yet Again & Kindest Regards,


        p.s. You may have noticed a number of detractors speaking out against each of us. I’m unsure of how often you’ve received investor attention, but pay it no mind. I’ve been doing this for decades and it’s always been this way.

        Your knowledge and integrity is unquestionable and I’ve made that clear as many ways as is possible. Barring the ignorant 1%, you can be sure that events like this only raise ones visibility in a benevolent manner. Whether or not it is invited is another matter entirely…

      • My knowledge and integrity seem to get questioned a lot by people that don’t know very much about the given subject. It happens when they have a believe or a desired outcome and I try and prevent the facts that I think prove the contrary.

        Certainly the Microvision investors don’t like me much. I can’t help it, but they don’t seem to have even a clue as to what it takes to have a viable product and what it will take to have cost effective green lasers. I think the only profitable business Microvision has is in printing stock.

        On the internet, I’m somewhat famous/infamous (I was mentioned in George Gilder’s book “The Silicon Eye”) for my comments on Foveon back around 2002 on (for one of my first post on the subject see that Foveon was making claims for cost and light sensitivity that I said were untrue (and history has backed me up on this). This led to some heated exchanges with all kinds of false statements and accusations about me including saying that I was secretly still working for TI when I had be gone for 4 years. I tried to make reasoned technical arguments based on my semiconductor background, but the Foveon fans would have none of it. The Foveon fans assumed I had a motive and even invented motives when the couldn’t find any. In fact, I (and others) figured out that a good bit of what Foveon was claiming was untrue and/or very unlikely. BTW, from what I can tell, Foveon was about to go under when what was left of Foveon was bought by Sigma. It is now use in Sigma’s very low volume “vanity camera products” that I doubt make any money and are probably being supported by Sigma’s lens business.

  5. I read your uninformed commentary on HIMX wherein you state that HIMX doesn’t produce LCOS.
    You sir, are a fraud. Did you bother to read the CEO’s earning call transcript ? You didn’t so you are a fraud. He stated; “We are working with several top tier customers to develop new head-mounted display products using our LCOS panels.” … Can you read english ?
    I cannot stand jerks like you who give idiotic opinions. You state you were relying on information from a year ago.. STICK TO WHAT YOU KNOW.. JERK..

    • Michael,

      I take from your slightly incoherent message that your interest in just in seeing Himax’s stock go up. You obviously have not read very carefully what I have written.

      When the Seeking Alpha story broke, there were those accusing me of pumping the stock. Then when I made sure that people were aware that in BOTH of my posts which appeared days before the Seeking Alpha article that while I was sure that the earlier prototypes used Himax panels, I was also sure that the latest prototypes did not use the same panels and likely they were using a transmissive panel which is a considerably different technology since it requires making transistors on glass versus the more common silicon.

      Just because a Himax CEO says they are working with “top tier customers” does not mean they have a great high volume opportunity with those customers. As the saying goes, “if I had a dime for every time a company said they were working with top tier companies,” I would be very rich. Clearly Himax Panels were being used by Google, but that does not mean that they will be used in a high volume Google Product. It is my technical opinion that the Google Glass units with the resolution they appear to be supporting would be smaller, lighter and more cost effective using a transmissive panel and NOT the panel they were using the early prototypes and that it is unlikely that the newer Google Glass device being shown used LCOS I believe my opinion would be supported by the most technical people in the field. LCOS’s advantage is in higher resolution and/or higher brightness applications (which could include future Google Glasses type devices but NOT the device they are currently showing).

      I know starting in 2009, Himax was having a lot of problems with their higher resolution field sequential panels. The work I heard is that they were retrenching on LCOS and in 2010 had bought Spatial Photonics as an alternative (link). You may note that Syndiant won the 3M designs away from Himax’s color filter LCOS.

      There is an outside chance that there is some incredibly cleaver new optics and Himax redesigned their panel to fit, but it would have been far simpler to use a transmissive panel. According to Mark Gomes, Himax is also working on transmissive panels but this is not widely known.

      I would suggest based on your second to last sentence that you have a self hate problem, because you clearly have idiotic opinions. My blog is about technology and not promoting or hurting any given stock.

      • I would like to echo Mr. Guttag’s sentiments. I have possibly gathered more information on HIMX’s current operations than anyone else. However, I would have had no chance of knowing what to look for without Mr. Guttag’s expertise and guidance on the technical requirements for a device like Google Glass. No chance.

        Thus, in my book, if there’s anyone sticking to what he knows, it’s him. The fact that we came to different conclusions is inconsequential. His opinions have no effect on whether Google has chosen HIMX or not…and that is the main thing that will drive the stock over time.

        His facts are generous and his opinions help to tell you what questions to ask to find truth and conviction. Recipients of those gifts should use them wisely…and graciously. Attacking any individual who donates time and expertise is simply bad form.

        Kindest Regards,

        Mark G.

  6. “The engine is on the order of 5 times too big to fit into the space available in the Google Glasses, and that does not include the electronics for Microvision’s LBS which take about as much volume as the optics).”

    The technical specifications for the PicoP Gen2 Display Engine list the dimensions for the
    Optical Engine as 35 × 20 × 6.1mm Max (W × L × H). This indicates that the optical engine can be smaller than these dimensions.

    • I’m not sure what point you are trying to make, but I would estimate that the outside case (the optics have to fit inside of this) of the optical module on the Google Glass is roughly a 12mm x 12mm around by up to 15mm long. The Microvision module only fits in one of the 3 dimensions (see HERE for views of Google Glass from all directions). If you give up say 2mm (1mm on both sides) for the case plastic, the optical module needs to be about 10mm by 10mm by 15 to fit. The thickness being under 10mm for Microvision can’t be moved to help the other dimensions. So being 35 (long) by 20 (wide) it take up an area of about and the Goolge Glass design has about 10 by 15 or to give so the Microvision module is about 4.6X too big. What makes the Microvision module area so big is the light combining optics (biggest part), 3-lasers and the light pathway in and out of the mirror.

      Then you have to realize that Google has all their own electronics associated with radio receiving, processing/image creation, and the like that they have to fit in. There is no room for the relatively huge electronics associated with Microvision’s beam scanner. If you crunched all of Microvision’s electronics to fit the form factor of the Google Glass, you wouldn’t have room for all of the necessary electronics. A panel would only require one or two very small chips to drive the LED and the panel.

      • Karl,

        What I am stating is that the Microvision module depicted in the photo in your article is the same dimensions as the one that was exhibited in a pico projector at CES 2012. This pico projector projected a medium sized image bright enough to be visible under overhead lighting. (see the link below)

        Are you saying a module of these dimension would be required to project an image onto a small substrate? Are you stating it is impossible to fit a Laser Beam Scanning solution into the dimensions you estimated for Google Glass?

      • “Are you stating it is impossible to fit a Laser Beam Scanning solution into the dimensions you estimated for Google Glass?” Short answer, YES.

        Not a lot of what is the Microvision module shrinks with lower power. You still have to combine and aim 3 lasers onto the mirror and get the light back out and all that takes space/Volume. You certainly can cut some of the bulk in terms of heat sinks and heat spreaders due to lower heat at lower brightness but essentially all of the optical components remain and most of them will be the same size. You still have the two ASICs for controlling the mirror and the laser scanner. The form factor of what it takes to combine the lasers and get them on the mirror are all wrong for fitting in newer Google Glass device.

      • I find it interesting that I asked about Laser Beam Scanning and you focused on Microvision.

        You conveniently ignore prototypes by companies like Btendo that are smaller than any LCOS engine prototype that uses a direct green laser diode.

      • There is not a lot of information out about BTendo and I know of your proclivity for Microvision. Btendo was aiming at front projection and not near eye. Their projection engine (the only one I have seen is shown in may be a little smaller Microvision and their electronics seems to be a lot smaller than Microvision’s but still the overall seems bigger than what Google Glass is doing.

        I don’t know if it is religious thing with you, but you should note that laser beam scanning keeps failing in the market. Samsung messed around with it and so did Hitachi among others. The problems with LBS go much deeper than just the cost of the green laser.

  7. Why do you Walter Mitty characters continue to think you know better than the experts?

    I don’t think anybody in the industry has ever considered LBS to be a realistic proposition because of the problems it has.

    The reality is that the LBS technology was completely oversold and you fell for it.

    • The short answer is the one inside the Google Glass, at least in 2 critical dimensions (but the Btendo is likely thinner). Per the link, the Btendo engine you point to about 18mm wide by 20mm long by 5mm thick.

      The key point is that this is not in Google Glass either. As per Mark’s comment, “The reality is that the LBS technology was completely oversold and you fell for it.” The Btendo engine probably cost many times what the LCOS engine costs.

    • Compound Photonics (CP) makes some nice claims, but I have yet to see it or even hear of them showing even a prototype in public. It is pretty hard to compare image quality and speckle on on something you cannot see.

      It should be noticed that to achieve the high light output, CP uses a frequency doubled green laser. This has the pros of being available at higher light output and far less expensive but it likely has narrow bandwidth and high inherent speckle. So my expectation would be that unless there is some extremely good speckle reduction, there may be quite a bit of speckle.

      It is true that it should be Class 1 even at 60 lumens whereas at about 25 lumens LBS will be Class 3R.

      • Karl is correct about the amount of speckle. The company itself acknowledges it in the comments section of the below video.

        Additionally, the Compound Photonics engine is 4.1cm3 whereas the STMicro engine is 1.7cm3.

  8. google glasses are interesting, himax are making the lcos panels which make sense, why people are focused on panels, there are no people talking about the optics, lenses, beam splitters etc.any reason? or there is no big companies making optics, do you think google making their own beam cubes etc? Please comment

    • There are many companies, large and small, that design optics for head mounted displays, you just may not know of them. Some firms just sell design expertise and others make optical engine to incorporate in another company’s final product. Sometimes companies will use their own in-house optics development and sometimes they will buy it from outside companies. The Google Glass design, from what I could tell was nothing special.

      A bigger issue is that head mounted displays, including Google Glass, has not proven to be a very large market to date. It is impossible to support a robust ecosystem with the current size of the head mounted display market. So far the big money is in selling technology to big companies (Occulus for over $2Bto Facebook, Osterhout Design sold patent rights for over $100M) that hope to make a play in the future. Microsoft’s CEO has already indicated that Hololens could be 5 year (or more?) from being a consumer product.

  9. “I аm curious tօ find out whaqt blog platfor you Һappen tߋ be սsing?
    I’m havіng some small security issues ѡith my latest website
    aand ӏ’ԁ likе to finnd sߋmething mоre secure. Do you Һave any recommendations?”

    • The website is now hosted on, not to be confused by (see for example I’m paying for a “business-level” plan so there are no ADs and I can use any plugin. In this way, takes care of the security issue.

      I’m no expert but I kept having a number of different problems with Bluehost causing my website to go down. For the last several years it seemed about twice a year the website would lock up and the “solution” was to keep buying more stuff (storage, better firewall, etc.). Bluehost was requiring me to do all the server management. The final straw was that even simple things could not be addressed by the first-level technical people anymore. I paid extra for a virtual private server which was supposed to give me better service, but the service I go kept getting worse. Each time I had a problem, they were bumping up to a higher level technical person that they said has 48 hours to address the problem. Bluehost “solution” was for me to hire a website manager. Going to meant that the headaches of running a server are managed by them. In the end, it is also going to cost less, but the reason for moving was all the headaches.

Leave a Reply