Archive for Karl_Guttag

Lasers – DGL, SHG, and Hybrid

I have no doubt that eventually all projection devices will use laser illumination due to their huge optical advantage.    But there still remain serious technical and business questions as to when they will become a major factor in displays.   A couple of years back we were in the “wild enthusiasm stage” of direct green laser (DGL) development with several companies announcing they had a direct green laser.    But since then, it has become clear that there are still some significant technical and business hurdles to clear for DGL.

The immediate effect of the DGL announcements two years ago was to make the development of second harmonic generation (SHG) and other diode pumped green lasers more difficult as it tended to make the funding and management support of developments in SHG lasers more difficult.      Since then it has become clear that while the DGL developments were very promising, they were still a very long way from being ready for production products.

Simply put, it turns out that the physics for making DGL is very difficult.  All the DGL development are based on around indium gallium nitride (InGaN).  Semiconductor today had a very good article on the subject in 2009  (note UCSB developments led to today’s Soraa).  To get to green they have had to add more indium and to try different crystal plane orientations which are less stable/yieldable.   To have a production product they have to solve simultaneously key attributes including the yield/cost, wavelength/color, stability/lifetime, output power, temperature range, and efficiency.    To date, they can only solve a few of these key attributes at the same time.    I sometimes quip, “I can get everything I want in a green laser but it is spread over 5 different parts.”

It also seems very clear that DGL are at best only going to have enough power output to support low projectors for the next several years.   The big problem is that low lumen projects have low value in the market which means that can’t afford expensive DGL.   This is creating a very big chicken or the egg problem in that there really is no significant market for early expensive, low efficiency DGL.

As reality sets in on DGL progress, there seems to be a resurgence of interest in SHG green technologies.  Getting DGL powerful enough to support the hundreds of lumens for a portable projector and the thousands and tens of thousands of lumens for conference room projector could be more than a decade away.   Companies need the optical advantages of an all laser solution to provide higher efficiency, smaller and less expensive optics, better efficiency, and long lifetimes, than traditional lamps, and they can’t wait on DGLs.

There are several techniques for getting green lasers by pumping crystals with lasers of a different wavelength.   Most commonly this is done by using an infrared laser at 808nm to pump a non-linear crystal (often periodically poled lithium niobate {PPLN}) to generate the second harmonic at 532nm green.   Spectralus   and QD Laser are two small companies with promising developments for relatively efficient SHG green lasers.

One issue with using a SGH and combining it with say direct diode blue and red lasers is that each of the lasers has a different character/beam profile and ages differently which can cause color shifts that would require constant recalibration to achieve accurate colors.   Also the direct red lasers that are available are at about 640nm which while a very deep/saturated red; it is very inefficient in terms of perceived lumens.   A more “ideal” red wavelength is in the 615nm to 620nm range but just like with trying for a 532 green, it has proven difficult to make stable/yielding red lasers in much below about 635nm.   These factors and others have caused the current large laser projector developers, such as Laser Light Engines, to use SHG for red, green, and blue even though direct diode red and blue lasers exist that could produce enough light.

Photodigm  has a more radical approach where all three colors (plus orange) could be generated from a single infrared pump source.   This could be particularly useful and very cost effective in the mid-lumen (say 100 to 3,000 lumen) projectors which use field sequential color such as DLP and LCOS.    Philips at SID 2011 also proposed a single laser only this time a blue laser to pump a crystal to provide red green blue and orange.

Another development pioneered by Casio is the “hybrid” laser projector.  In these projectors they use a red LED and blue laser for blue and use either the same blue laser through a spinning wheel or a second blue laser to drive a green phosphor to get green.    The “green” is not tightly collimated laser light and is not ideal from an optical perspective, but it is the cheapest way to get a very bright and small green light source.  Using blue lasers to stimulate a green phosphor is an admission that DGL, at least bright ones, are a ways off.    There is a so much buzz in the industry about this hybrid green approach that market analyst Insight Media has release a report on it (http://www.insightmedia.info/reports/2012hybriddetails.php)

My conclusion is that projector makers aren’t going to be waiting around for DGL to get going with laser projection.    It is clear that the hybrid/phosphor-green approach is already taking off for 2K to 3K lumens.   I also foresee the hybrid approach migrating down into the 200 to 1000 lumen markets.   But there are bigger overall advantage with smaller optics and microdisplays to be had by having an all laser solution.

Soothsayer 7: Microvision’s Obfuscations Causing a Buzzing Sound

Microvision continues to make thinly veiled accusations against this blog in their May 9th, 2012 Company Displayground Blog (quoting directly with my bold emphasis added):

Our shareholders are following the topic of direct green lasers with avid interest and we get a lot of questions from them on price and availability. They have ridden the green laser wave with MicroVision and are understandably anxious for their investment and patience to pay off. There is another group of MicroVision watchers that take an active interest in direct green lasers and for that matter, all things MicroVision and have quite a bit to say. We welcome such interest as we are really proud of what our patented solution can do and the advancements we have made with the PicoP Gen2 display technology. But it does get tiring to have an open mic of misinformation from parties who only seem to have an interest in not seeing Microvision succeed. We try and ignore this contingent just like the best thing to do when a fly is buzzing around your head is to ignore it. Eventually the fly finds something or someone else to buzz around and the problem resolves itself.

Like most things Microvision writes they raise more questions than they answer.  For example, who is this “group” of watchers that “ have quite a bit to say” about green lasers?  As far as I know, this is the only blog regularly writing about green lasers for small projectors.   I guess they feel like I have them surrounded :-).   I would suggest that the real reason they cannot ignore the “buzzing sound” is that is goes contrary to Microvision’s attempts at obfuscation and eventually their investors and analysts ask questions.

I take the most interest went key specs are missing or when states what is at best a half-truth.  From what I read, Microvision obfuscates, uses straw-men, half-truths,  give meaningless ratios of improvements, and state goals/expectations as if they have been met.  They could put this all to rest if they could be specific about what they consider “misinformation” and give direct clear answers.  

$200 Green Laser Clarification:

Microvision’s blog when on to clarify(?) what Microvision’s Lance Even’s was saying green lasers costing nearly $200 that I address in my April 30th blog.  Quoting directly from the Microvision Blog:

The price of direct green lasers is understandably a topic of speculation since the manufacturers of the diodes have not publicly discussed pricing or even the exact timing of commercial availability. We cannot reveal specifics around these issues, but we can and have stated that we expect the prices of direct green lasers to be significantly less than synthetic green lasers which have cost nearly $200.

Microvision, apparently in response to my blog on the “nearly $200” price of DGL clarified that Mr. Evan’s $200 remark was in reference to the synthetic green lasers.   But note it is only an “expectation” (as in some time in the future) that the prices will be lower.

It is also interesting that Microvision is admitting publicly that the lasers were costing them nearly $200 (or maybe more at some point).  You have to wonder why they went to market with the ShowWX using a $200 laser for projector with only 10 to 15 lumens.  Can you imagine what investor reaction would have been back in 2009, 2010, or 2011 if they knew that Microvision was selling a supposedly high volume consumer product with a $200 laser in it?   My speculation is that unless they agreed to pay $200 for the lasers and buy a lot of them, that they would have had no way to build any product and if they couldn’t show (some pun intended) and that would make it impossible to raise money.   To keep investors on the hook, so to speak, they had to use obfuscation about the cost of lasers (some things never seem to change).  The losses on the ShowWx were essentially marketing expenses to raise money from the “shareholders” that were following the company back then.

They seem to have made the calculation that they needed to loose many millions of dollars on the ShowWx product line in order to keep the investor money flowing in.  Now that they consider those losses in the rear view mirror, they are admitting to them.   I wonder when they will come clean on the cost and specs of today’s direct green lasers.

More Half-Truths from Microvision

One more interesting set of half truths from Microvision’s blog:

We are confident that the manufacturers of the direct green lasers will be able to meet the volume requirements as the market demand for direct green lasers grows, and we expect price to fall accordingly.”

The first half-truth is that “manufacturers of the direct green lasers will be able to meet the volume requirements.”  Yes, this is true today, but only because the volumes will be very low.   Among the reasons is that the price of the lasers will be very high and that they are not currently designed into any high volume products.   Secondly, by laws of supply and demand, if the lasers are expensive few will be bought and thus they will meet the volume requirements.  Thirdly, the growth in the market could be very slow so it would be easy to meet what Microvision “expects.”    In short, Microvision like a politician, used a lot of words to give no real information.

What Microvision won’t admit, I suspect because that would be bad for raising money, is that the direct green laser has so far proven difficult to make due to the physics involved.    Certainly very smart and capable people are working on DGL, but they are not ready in the near future for the high volume consumer products.

Augmented Reality and Google Project Glass Part 2

Google Glasses

Since my last post on Augmented Reality (AR) and near eye (head mount) display Google put out some publicity on their Project Glass concept.   Google made it abundantly clear that this was only for the purposes of concept testing and not a real product, but they also said that there would likely be some test products at the end of 2012.

Jason, “The Frugal Dad” wrote saying he had seen my first article on AR and that he has a new “Infographic” that includes Google Glasses as a Future “disruptive technology.”   Unfortunately most predictions about the future turn out to be wrong and nothing I have seen so far in the way of near eye AR, including Google Glasses, I believe will meet the consumer expectations and become pervasive.  I’m not saying it won’t ever happen, but rather there are still many major problems to be solved.

As I wrote before, I think there are many practical issues with near eye displays and augmented reality.   The Google Video “Project Glass: One day…” was obviously a “concept video” and all the images in the display were “fake” as in not what the actual display will look like.

Along these lines, the April 5, 2012 Wired had an article called “Google Glasses Face Serious Hurdles, Augmented-Reality Experts Say” which raises concerns that Google is over-touting the concept.  The Wired article quotes Pranav Mistry, from the MIT Media Lab  and and one of the inventors of the SixthSense wearable computing system, “The small screen seen in the photos cannot give the experience the video is showing.”  Also in the Wired article, Blair MacIntyre, director of the Augmented Environments Lab at Georgia Tech raised concerns that Google is raising expectations too much.   Both Dr. Mistry and Dr. MacIntyre are certainly proponents of AR.  Their concerns, and mine as well, are that that raising expectation too high could backfire on the AR concept in long run.

Dr. Thrun on Charlie Rose Looking Up

Sebastian Thrun, Google Fellow and Stanford professor, was on Charlie Rose on April 25, 2012 wearing a working Google Glasses prototype.   The first 4 and a half minutes of the Charlie Rose video discuss Google Glasses and gives some insight into the issues, not the least of which is whether people are really going to wear something like this.

To see the images in the “glasses” he has to look up, where the Google Concept video suggest the images are right in front of you all the time.  So he can’t see the person he is talking to and the computer image at the same time.   Imagine talking to somebody wearing these when they are clearly looking up while talking to you (particularly notice Dr. Thrun’s eyes in the picture above at 1:11 into the Charles Rose Video).   By instinct, humans are very sensitive to eye behavior, and someone constantly looking away (up) is a distracting behavior.  Now imagine you are walking down the street and searching for something on your glasses and a truck comes by — big oops.

The most insightful comment by Dr. Thrun was “we having yet found this [augmented reality] to be the compelling use case” but he didn’t elaborate as to why.   But this does indicate that Google is still be trying to figure out if AR is really compelling.    Dr. Thrun did say that “the compelling use case is the sharing experience” and commented on sharing pictures at being something they enjoyed — I guess this is Tweeting on steroids where all your friends can see what you are doing as you do it.  In this case the glasses become a hands free video camera.

The Google video has inspired some funny spoofs of it that in their own way make some of the points above:

ADmented Realility — Google Glasses Parody extrapolates on the what could happen with advertising gone wild.

Google Glasses: A New Way to Hurt Yourself and a video shown on Jimmy Kimmel Live demonstrate the dangers of “distractive walking”

The next time on the subject of AR, I plan to talk about more of the technical issues with AR.

 

 

Soothsayer 6: Microvision’s Lance Evans “Green lasers alone are $200 each now”

First, sorry for being away so long.  Some family and other matters took me a way and I fell out of the habit of posting.  I am going to try to have at least couple of posts up a week.

My “Soothsayer” series started back in December 2011 when Microvision released an 8-K a very trasparent response to this blog.  In Microvision’s Dec. 19th 2011 8-K they stated ““In the coming weeks we intend to provide a series of posts that discuss direct green lasers in more detail, as well as other business updates. Stay tuned!”

Well, it has been 19 weeks since Microvision’s 8K and there has been pretty much silence from them on “more details.”  Microvision has had plenty of opportunities to add “more details” in the last 5 months but has chosen not to.  They say a “slip in Washington is when you accidentally tell the truth” well maybe Microvision did that today.

In a way, Microvision finally broke the silence with an article today in an April 30th Technology Review article with Lance Evans (a director of business development at Microvision) stating, “Green lasers alone are $200 each now.” Remember this is probably for a green laser that supports on the order of a 15 lumen projector and has a wavelength that is too short (too blue) to be used in a typical projector and it not very efficient in terms of lumens per Watt. These are lasers that are a best useable in a car HUD display and not a batter powered cell phone or hand held projector.

Back in December 2011 Microvision’s 8K stated, “DGLs will be much cheaper than synthetic green lasers at introduction.”  How does this fit with DGL’s costing $200 today? 

The article does day that “Evans expects that costs should fall to a tenth of current levels by the end of this year,” but note the word “expects” is corporate speak for “believe, wish, hope, or dream” because you generally can’t hold them legally accountable for an “expectation.”  What other than a wish causes the direct green laser cost road map to drop from $200 now to one-tenth that cost or $20 in less than 7 months?  Are there people lined up to buy lasers first at $200, then $100, then $50 say to support a 15 lumen projector?  None of this makes any business sense.

Even if Mr. Evans “expectations” come true, and DGL drop by 10x in less than 7 months to $20 by the end of the year (really for 2013 production), is this really a viable business?   By comparison, there are 200 lumen LED base projectors on the market today and the cost of the red, green, and blue LEDs combined cost less than $20 today.   For a 15 lumen LED projector today, the LEDs are more like $3 for the RGB set.  For a pico projector to make it into a cell phone, the cost to the cell phone maker of the whole projector including electronics has to be on the order of $25 or less (just ask any cell phone maker, I have asked many).  You can even come close with even a $10 DGL, no less 20, when you factor in the cost of everything it takes to make a projector including optics and electronics.

I’m a long term believer that eventually all projectors and in fact a vast number of other products will be using lasers.   It is just going to take more than 2 year for the brilliant people at the laser manufactures to figure out how to make the direct green lasers at a cost point that will lead to mass adoption.

Augmented Reality / Head Mounted Displays (Part 1 Real or Not?)

Augmented Reality (AR) Head Mounted Displays (HMD) [aka Near Eye Displays,  Wearable Computing, and other many other names] has gotten a big boost in the public mindset with the Feb. 22, 2012 New York Times (NYT) article/leak about Google Glasses with Android.    The NYT article result in a flurry of commentary on the web and television (Google search “Google Glasses Augmented Reality“).  Reportedly Google Glasses will be available in very limited/test market release later in 2012 with a price between $250 and $600 (US).

Augmented Reality (AR) is the concept of combining computer information with the real world.   Head Mounted Displays (HMD) is any display device that in some way attached to the head (including the eye, such as a contact lens).   You can have AR on say a cell phone with a camera where computer information is put on top of the video you see without a HMD.  Similarly you can have an HMD that is only a display device without any AR capability.   But often AR and HMD are combined together and this series of articles is mostly going to be talking about the combined use of AR and HMD.

Some History

Augmented reality/HMDs have found their way into many films as a plot element and this has to some degree already primed the public’s interest.   It turns out it is much easier to make it work in the movies than in real life.  Attempts at augmented reality go back at least as far as the 1960’s.   Below is a montage of just a few of the over 100 attempts at making a head mount display which range from lab experiments to many failed product in the market (they failed so badly that most people don’t even know they existed).

The Airplane Test

So far HMDs have failed what I call the “I don’t see them on airplanes test.”  If there is anyplace you should see HMDs today, it would be by people sitting on airplanes, but have you ever seen someone using one on an airplane?  Why I consider this as a “metric” is that the people who regularly fly are typically middle to upper middle class, are more into small electronic gadgets (just look at what they sell in the on-board catalogs), and the environment stilling on an airplane is one that you would think would be ideal for a HMD.

Back when the iPad came out, you could tell that they were taking off just by the number iPads you saw people using on airplanes (mostly to watch movies).  Interestingly, I have seen HMDs sold in Best Buy vending machines at airports, but I have never seen one “in the wild” on an airplane.   The other place I would have expected to see HMDs is on the Tokyo subways and trains, but I have not see them there either.  One has to conclude that the “use model” for the iPad works in a way that it does not for a HMD.

Augmented Reality (AR)Topics

There are so many topics/issues with Augment Reality Glasses (or which every name you prefer) that there is too much to cover in just one blog post.   In terms of implementation are the technical issues with the display devices and optics, the physical human factor issues like size and weight (and does it cause nausea), and the user interface or use-model issues and feature set (including wireless connectivity).   Then there are a whole number of social/political/legal issues  such as privacy, safety (distractive driving/walking) user tracking, advertisements, etc.   AR is a very BIG topic.

Making a practical Augmented Reality HMD is deceptively difficult.  It is a daunting task to make an HMD device that fits, is light enough to wear, small enough to go with you, produces an acceptable image, and doesn’t cost too much.  And making a display that works and is cost effective is really only the starting point, the rest of the problem is making one that is useful for a variety of applications from watching movies to real time head-up displays.

There are a number of user interface problems related to the fact that a HMD is in some way strapped to your head/eye that make is “not work right” (acts unnaturally)  in terms of human interfaces.    These human interface issues are probably going to be a bigger problem than the physical design of the display devices themselves.   Making a HMD  that “works” and is cost effective is only the starting point.

Will Google Glasses succeed where others failed?

The answer is likely that they will not have a big success with their first device even if it is a big improvement on past efforts.  Even the rumors state it is a “test market” type device meaning that even Google is looking more to learn from the experience than sell a lot of units.   I’m sure Google has many smart people on the device, but sometimes the problem is bigger than even the smartest people can solve.

The idea of a display device that can appear/disappear at will is in compelling to many people and why it keeps being tried both in the movies and television as a plot element and by companies trying to build products.  My sense is that we are still at least a few more technology turns of the screw away from the concept becoming an everyday device.   In future articles I plan on discussing both the technical and user-interface challenges with HMDs.

Soothsayer 5: Thanks for the “Shout Out” on the Microvision Conference Call!

It’s nice to know that a CEO of a public traded company is following my blog.   On today’s Microvision Conference Call (available for a few days) at 34:16 in the call, Microvision CEO, Alexander Tokman, in his closing remarks gave a “shout out” of sorts with,  “The direct green laser is becoming a reality this year and not 2014 as some led you to believe.

As anyone that has followed Microvision or my blog and my comments on direct green lasers the last few months should know, by “some” he means “me.”   First, he has mischaracterized what what has been written.    What I have written is that direct green lasers will not be practical for high volume applications until 2014 or beyond.

As I wrote Microvision’s “Soothsayer(?)” for their “Number One Question” in  response to me in Microvision 8K’s “false soothsayer” comments:

The real question is whether these lasers will be available at a price point, with a wavelength, and an efficiency that is practical.   I don’t doubt that most if not all the companies will be in production with a green laser in 2012, but what constitutes “mass production” is a different matter.   Nichia already has a 510nm green laser in production, for example, and it might be possible to build a heads up display for an automobile with it (albeit a bit expensive for the purpose), but that is clearly impractical in terms of wavelength, efficiency, and cost for building a high volume battery powered projector.   I also question whether they will be bright enough for a volume product other than a HUD.

Instead of addressing the real issues, Microvision has seen fit to play silly word games with the definition of word “commercial.”   “Commercial” simply means that companies will sell a product with a given spec.  It does not mean the that product, in this case a Direct Green Laser will meet the cost and performance requirements for a high volume application.  The reality is that the “commercial” direct green lasers are going to be available in 2012 (and 2013) are going to be way too expensive and with key performance limitations  for any kind of significant volume consumer products.

Now lets look at the math in Microvision’s timeline.  At 15:50 into the conference call Alexander Tokman talks about their commercialization timeline for a consumer product.    He said they hope to have a company committed by “mid year 2012” and that companies will generally take “6 to 18 months” to develop a product.   If you follow Microvision at all, you will note that Mr. Tokman timelines are usually wildly optimistic, but just taken him at his word with a “start” in mid 2012 and more realistically 12 to 18 months to get a product designed and ready for production adds up to products STARTING to be manufactured (not even high volume) in late 2013 to mid 2014 if (and this is a big if) everything goes according to Mr. Tokman’s hopes.

But given Mr. Tokman’s history of over optimism, this would seem to say that he is really talking 2014 or later; oh but wait, he said in the conference call that it was “not 2014 as some led you to believe.”     So which does his mean?

Soothsayer 4: Questions for Microvision’s Conference Call

I have been traveling for most of the last 2 weeks with a number of business meetings.   I thought I would get a quick blog out today ahead of Microvision’s  investor conference call on Monday Feb 27th (tomorrow as I write this).  I’ve listen to a few Microvision conference calls in the past and they usually say they have made great progress to some vague “goals” and then get a few softball/easy questions from the “financial analysts” that are allowed to ask questions and then quickly end the call.   You can listen to if “live” at 8:30AM Easter (5:30AM Pacific) or for a few days recorded at http://phx.corporate-ir.net/phoenix.zhtml?p=irol-eventDetails&c=114723&eventID=4727742.

It does not take some deep dark secret (i.e. confidential) information to figure out that laser beam steering ala Microvision has serious problems.   It just takes some engineering and business knowledge.  Below are a few questions I would like to see asked along with some information.

Microvision “720p” Optics and Driver Board

1) Isn’t the combination of the optics and drive electronics for Microvision’s 720p (see picture)  too large to be embedded into any major brand cell phone?    Right now you have 2 large ASICs and an FPGA to control the lasers and the mirror where the competition has only 1 much smaller ASIC.    According to the technical specifications this optics and electronics module is 35mm × 65mm × 6 .1 mm; how small do you have to get the module to meet the requirements of the major cell phone companies and when will you meet this requirement?

2) According to your technical specifications for the new 720p module, at “27% video” the power dissipation is “approximately 2 Watts.”   If this is 27% of 15 lumens (or even 25 lumens), it would suggest that at full brightness the power dissipation is on the order of 4 to 5 Watts.   You have stated in the past that the goal to meet the cell phone requirements about 1 Watt.   This would seem to be much worst power dissipation than the competing technologies.  How is any cell phone company going to embedded something like this that consumes so much power?     How long do you expect to take to get the power dissipation for the projector module down to your stated goal of about 1 Watt?

3)  According to your technical specifications for the new 720p module, at 25 lumens the module is a class 3R laser product and at 15 lumens it is class 2.   A) at what lumens between 15 and 25 does the product cross over into class 3R?  B) Isn’t even laser safety class 2 a serious problem for consumer product including cell phone companies?   C) What are the issues with trying to sell a class 3R product in the market, particularly when the competing technologies such as DLP and LCOS don’t have this issue?

4)  In your December 19th 2011 8K you stated“So what do I mean by “commercial version” direct green laser? It’s a laser that has passed through intense qualification by the component manufacturer to insure that it meets all of its intended performance specifications, with confirmed reliability and manufacturability necessary for mass production.”  But commercialization only means that they will be available for sale at some price, and it says nothing about them being viable for making profitable product.  Doesn’t all this talk about the definition of “commercial” just avoid the real issue of whether the lasers you need for laser beam steering will be PRACTICAL in terms of cost, power efficiency, lifetime, and wavelength (color) in 2012 and 2013 for embedded products like cell phones?

5)  The blog KGOnTech has posted several articles stating the laser beam scanning process will result in lower than your claimed 720p resolution.  Can you actually demonstrate that you can actually fully resolve a 1280 by 720 pixel test pattern?

6)  Lately you seem to have put a lot on emphasis on automotive heads up displays (HUD).    The power and size requirements are much more relaxed for this market.    Is this in some ways an admission that the lasers are not going to be ready for any embedded products in the near future.    

7)  Realistically, how large is the market going to be for an after market automotive HUD?   Are the laser cost, availability, brightness, and other specification acceptable today for the automotive market?

Please feel free to add questions or ask me to clarify my questions in the comments.

Karl

TI DLP® “Diamond” Pixel

Fig. 1 DLP Diamond Pixel (above) compared to test pattern (below)

TI’s DLP® WVGA (848×480) and WXGA (1280×800) microdisplays use what are commonly known as “diamond” pixels.  This article will explain what these pixels look like, their effect on image quality, and the reason behind diamond pixel.

As a side note: all the pictures of projected images were taken on a Qumi 1280×800 projector using the HDMI (digital) input driven at the “native” 1280×800 resolution to try and give the best possible source image.

Diamond Pixel Organization

Fig. 2 Diamond Pixel Column and Row Numbering

Fig. 2 Show how the “diamond pixels” are organized.  The pixels themselves are square, but they are rotated 45 degrees and fit like tiles in a zig-zag arrangement.  The Columns and row numbering are based on how the memory bit (signified by the red dot in Fig. 2) for each pixel is address.  You should notice that the columns numbering is more spread out than the row numbering.  For a WVGA (848×480) there are 608 columns [roughly 848/sqrt(2)] and 684 rows [roughly 480xsqrt(2)].  Note that if you multiple 608 x 864 you will get slightly more than 848×480 so there are roughly the same number “pixels.”  Details on this organization for the WVGA can be found in the DLP® 0.3 WVGA Series 220 DMD data sheet.   For the WXGA (1280×800) there are about 910 columns by 1136 rows (the datasheet is not publicly available).

Image Re-sampling Effect of Diamond Pixels

Fig. 3 Simplified Diamond Pixel Re-sampling

The image quality issues with diamond pixels comes from trying to “re-sample” (remap) the pixels from a normal square grid to the “diamond grid.  Fig. 3 Show shows a simplified example of the problem.  There are two black pixels labeled “A” and “B” shown.  On the left side of the figure, the green grid shows the normal/square array of pixels and it is overlaid with the the diamond grid in red.  As can be seen pixel A straddles/touches pixels 1 through 5 in the diamond grid and pixel B straddles pixels 5 though 11.   There is no really good way to map the pixels.  If you you just mapped to the nearest pixel on the diamond grid, it would cause severe jaggies, so it is best to re-sample/filter the pixels onto the diamond grid.

On the right hand side of Fig. 3, I did a simple weighting of the areas of overlap to remap the pixels.   Notice how the pixels have blurred out.  Also notice that the two black pixels end up mapping into different shapes in on the diamond grid because their relative alignment between the square and diamond grid is different.

TI’s DLP uses more complex algorithms than those used in Fig 3 that attempt to reduce the blurring by sharping (particularly in the horizontal direction) or allowing more jagged artifacts (in the vertical direction).  Any algorithm/filtering/re-sampling by necessity will be a compromise between jagged pixel artifacts and blurring. 

Fig. 4 below shows a close-up of a series of simple test patterns of a checkerboard, horizontal lines, and vertical lines with a picture of the projected image from a test pattern.    The checkerboard is pretty well obliterated in the DLP’s re-sampling process and you see what is known as a classic under-sampling problem where you see the “difference frequency” or a low resolution version of the repeating pattern.  It is also interesting is that the artifacts are different in the horizontal and vertical direction because they use different re-sampling algorithms horizontally and vertically.

Fig. 4 Closeup of Checkerboard, Horizontal Lines, and Vertical Lines compared to test pattern

Fig. 5 below shows more more of the test pattern (click on image to see all the detail – at the end of this article, I have included the whole the test image used so you can duplicate this test).   The longer lines in the test pattern are suppose to be 4 alternating black and white lines.  Where the 4 line pairs cross there are some strange effects (one of which is circled).   On vertical lines or groups of lines there is the occasional “ghost line” (some of which are pointed to with red arrows); these are as a result of the horizontal filtering/re-sampling process.   Vertical lines generally look a little better but they have that ghosts lines (“sharpening halos”) as a result of trying to keep from loosing sharpness.

The vertical re-sampling process (which shows up in horizontal lines) is much simpler and results in more jagged artifacts and less effective resolution.   You can also see in the dot in the letter “i” in the “8 Point Arial” how a single pixel blurs out over multiple pixels in the diamond grid of pixels.  Notice how all the dots in the “i’s” vary (circled in red) and in the Arial 10 point font the dots in the letters “i” even point in different directions.

Fig. 5 Diamond Pixel Projected Image of Center of Test Pattern

Why DLP Uses Diamond Pixel

The “marketing spin” I have heard for the diamond pixel organization is that it can give perceptively better resolution.  But this could only be true if the source image is much higher resolution than the displayed image.  As the pictures above demonstrate, about the worst case image is to drive the Diamond Pixel with its supposed “native” resolution.

At least one real reason that the TI DLP® uses the diamond arrangement is to try and reduce the projector’s thickness.   DLP’s require “off axis illumination” which means the light comes into the mirror array non-perpendicular to the surface.  When the mirror is tilted “on,” the angle of the mirror directs the light toward the projection lens. The current DLP mirrors tilt approximately +/- 12 degrees and they tilt diagonally (at 45 degrees).

Fig. 6 Light Paths for “Square” versus “Diamond” Pixel Arrangement

With normal/square mirror arrays (left half of FIG. 6) this would mean that for light to go out of the projector the light must come from above or below projection lens.  But in order to have the illumination light come from above or below the projection lens would make the projector much thicker.   This was not an issue for DLP in RPTVs or larger data projectors, but with Pico Projectors, customers care a lot about thickness.   The “Diamond Mirror Solution” was to rotate the mirrors so that the light could be in the same plane (right half of Fig. 6) as the projected image.

Conclusions

While the diamond pixels address the thickness issue, they definitely hurt resolution and cause artifacts and a loss of resolution.  The obvious problem is that a single pixel on a rectangular grid will cover part of at least 4 or more pixels on the diamond grid.  So there has to be some scaling/filtering to map the pixels from the rectangular grid onto the diamond grid.   The DLP ASIC tries to combat this blurring in the horizontal direction by filtering and sharpening.  The vertical processing is simpler and results in more jaggies.

Many of the problems with DLP’s diamond pixel are closely related to the resolution problems of Laser beam scanning that I wrote about in a previous article (see: http://www.kguttag.com/2012/01/09/cynics-guild-to-ces-measuring-resolution/).  Namely, it is the problem of trying to re-sample the original image to fit a grid or scanning process that does not match the original image.  Even if the number of “pixels” is the same, the re-sampling process hurts resolution and causes unwanted artifacts (for those engineers in the audience, it is a classic Nyquist sampling issue).

In optical terms DLP optics are “off-axis” which means the illumination of the display is not perpendicular to the imager.  Off-Axis optics are always bigger and their light path takes up more volume.  The diamond rotation keeps this volume from increasing thickness, but still the volume of a DLP optical engine tends to be bigger for the same F-number and Focal Length optics.   In general, DLP “pico projectors tend to have longer throw ratios (longer focal length lenses) and this may be due to trying to keep the optics from becoming larger.

Appendix

Below is the test pattern used to create the images above on a DLP 1280×800 projector.  To use this correctly, you need to drive the HDMI input of the projector at 1280×800.

I have included a similar test pattern for WVGA (848×480) that can be used with the WVGA DLP projectors like the PK201 or  Microvision’s ShowWX/+/HDMI projectors.  Additionally, I have included a 720p (1280×720) version of the test patter in case you come across a 720p projector.

I would like to thank Paul Anderson for taking the picture of the test pattern on his 1280×800 DLP Qumi Projector that I used in this article.

1280×800 Test Pattern:

848×480 Test Pattern:

720P Test Pattern:

Answering Questions on Laser with LCOS and LBS

Syndiant 720p LCOS Panel

I got asked a series of questions by reader “me_wwwing” after the article Laser with LCOS is Focus Free — Yes Really! that I am answering in this post because I thought they would be of general interest.  I did have to edit a few of the questions for clarity but tried to keep the intent as best I could and I have re-ordered the questions to put what I think are the more interesting questions first.   For background for the reader I need to add that from prior questions and comments, I know me_wwwing to be a fan of Microvision’s laser beam scanning (LBS) so that may color some of the questions and the answers.

Q1. Does Syndiant need to use the same lasers as MVIS [Microvision laser beam scanning (LBS)]?

The simple answer is that LCOS panels can use any of the lasers that LBS can use and can use lasers that LBS cannot use.   LBS puts constraints on both Diode Pump Solid State Lasers (DPSS also known as “synthetic” or “frequency doubled) and for direct diode lasers (DGL).

Answer Part 1: DPSS Green Lasers

First, Syndiant makes the LCOS microdisplays and not the entire optical engine.     A panel (LCOS or DLP) optical engine has a wider selection of lasers that it can use both for direct green lasers as well as DPSS green lasers.   So the quick answer is that it can use any of the lasers that a LBS system can use plus it can use laser types and variation that LBS cannot use.

With DPSS green lasers, panels could use the less expensive to make and much more electrically efficient (generally about 2X the efficiency) slower switching lasers.  The slower switching frequency doubled lasers are capable of greater than 12% wall plug efficiency (WPE) with the desirable 532nm wavelength green.   About the best the fast switching frequency double green lasers ever got to was about 6% WPE.

The slower switching DPSS greens are capable of going to very high brightness; they can be over 10 times brighter than today’s Direct Green Diode Lasers (DGL).   Their electrical to lumen efficiency is over 3X the best DGLs today.   And their cost is lower than DGLs.   But the drawbacks to DPSS greens include the size and the small spectral bandwidth that causes higher speckle.

Answer Part 2: Direct Green Diode Lasers (DGL)

It turns out that there are different kinds of direct-diode lasers as well.  Most notably there are “single mode” and “multi-mode” lasers.

LCOS or DLP can use either the single-mode or multi-mode lasers, but lasers LBS can only use single-mode lasers.   Multi-mode lasers change wavelength, phase, and/or optical polarity somewhat randomly.  The rate of mode changing would look like noise in a LBS scanning process so it is unusable in an LBS system, but in a panel (LCOS or DLP) system the hopping simply gets average out as it is faster than the eye can detect.

Since changing wavelength and phase reduces speckle, multimode lasers have less speckle.  It turns out that lasers naturally want to mode hop, so it is easier (cheaper) to make multimode lasers, multimode lasers are more electrically efficient and multi-mode lasers can be made much brighter/more powerful (in fact, as they make the laser cavity bigger to make lasers more powerful it is hard to keep them from mode hopping).  So panel based projectors have a significant advantage in being able to use multi-mode lasers.

Historically, traditional laser uses such as telecom and interferometry, needed single mode lasers.   But the same coherent light for these applications is what causes speckle, so for panel based projectors you want “poor quality” lasers with less coherency and thus speckle.

Q2. How big is Syndiant Controller?

The Syndiant’s 720p ASIC is currently 9mm X 9mm in its current package, but could be put in a significantly smaller package by reducing the pin count for an embedded cell phone.  The current package has a lot of extra pins for supporting a mix of applications that make it bigger.

Q3. What are the dimensions of Syndiant’s PCB?  The new one or the old one to have a starting point.

Syndiant sells the panel and the driver ASIC and not a PCB per say.  Single SYA1231 ASIC for their 720p is currently 9mm x 9mm includes frame buffer memory and ARM CPU and can be put in a smaller package for embedded applications.    So there really isn’t much in the way of board space for the new Syndiant 720p controller.  Basically, it is just one small chip 9mm X 9mm chip and not a “PCB” per say.

Syndiant’s single small ASIC compares extremely favorably to the Microvision 720p board which has 2 custom ASICS (one about ~11mm x 11mm and the other ~10mm x 10mm)  and an Altera FPGA (~6mm x 6mm)  on it.  There is also a 4th I.C. on their board which is an Intersil laser driver (5mm x 5mm).   The  picture below is from www.technogytell.com with my notes added it:

Microvision "720p" Optics and Driver Board

Totalling up the area of Microvision’s two ASICs and FGPA, Syndiant’s current control ASIC takes about 1/3rd the area and 1/3rd the ICs of Microvision’s 3 chip controller.   Also just looking at all the power conversion circuitry required in the Microvision “720p” board suggests that it needs a lot of different voltages with some significant power requirements and all this adds cost, board space, and power.

Q4. What size diagonal lens (encased) does Syndiant’s displays need to cover the diagonal of the display?  A length too would be nice. old specs are ok to start with [assuming laser illumination]

The diagonal of the “active display” of the older SYL2010 was 0.21”.  It had a 5.4 micron (10-6 meter)  pixel such that an 800 pixel high display was 4.3mm.   The lens would need to be a bit thicker than this diagonal.  For a small high volume product the lens would be cut to a more rectangular than circular shape to reduce height.   So the lens could be about 6 mm thick.   The SSTDC SEE100 prototype engine http://www.picoprojector-info.com/laseno-see100-module-photo  engine’s lens with its barrel was non-optimized (circular) and was about 8mm thick and 8mm long.  An optimize rectangular-cut lens could have been about 5mm tall.

There is nothing keeping LCOS with laser illumination from going below 5 micron pitch (they will eventually get to around 3 micron and perhaps less) for its pixels.   At about 5 microns the active display for a 720p would be about 3.6mm tall.  A rectangular cut lens would then be about 4.5mm to 5mm tall or a round lens would be about 7mm tall.

Q5. Does the dichroic color combiner lens need to cover the surface of the Vibrating Despeckle unit [the answer addresses the broader question of the light combining system for LCOS and LBS]?

No, the light is not significantly spread out before the dichroic color combiners in an LCOS system.  Also the alignment of the dichroic mirrors/filters and the lasers is non-critical in an LCOS or DLP optical engine.

In should note, however that the alignment of the lasers and the dichroic combiners is very critical in an LBS design.  You should note in this teardown picture (picture taken from a laserpointerforums) that I have labeled the two ball shaped optics (pointed to in magenta) used to aligned the red and blue lasers and then glued into place.   The need to critically align the lasers adds cost and quality issues into to the manufacturing process.

Microvision ShowWX (WVGA) Optics

 

Q6. What is the power to run the PCB and the light engine on a Syndiant’s Displays? (SVGA and WVGA).

That really is a complex and involved question and depends on a lot of factors including the optical engine design and the target brightness.  I am also assuming that this is with LED illumination.   With LEDs the lumens per Watt of power tends to go down as the projector gets brighter, that is one of the advantages with lasers, namely that efficiency does not go down with power. 

It is helpful to break the power into two part, the panel and controller electronics and then the illumination and optics.   For low lumen (less than 30 lumen) projectors, the power of the display and its control is a significant part of the overall power (and less so at higher brightnesses).

The Syndiant LCOS display and ASIC for the WVGA or SVGA will usually consumed about 0.4W.   Due to a number of design improvements, Syndiant’s 720p panel and ASIC consume less power while have a higher color field rate and better light throughput than the older WVGA and SVGA devices while having about 3X the pixels.

In the low lumen area of 10 to 15 lumens with a small, 0.21” WVGA or SVGA panel optical companies were able to get about 7 to 10 lumens per LED Watt.    With some power conversion overhead and including the panel and ASIC this mean that you could get with LEDs about 12 to 15 lumens for 2W of power.  While this is still not good enough for the very high volume cell phone applications, it compares very favorably to the ShowWX which takes 4+ Watts for about 15 lumens.

With lasers the efficiency depends heavily on the lasers used and is dominated by the green laser efficiency which is pretty poor with today’s DGLs.   With today’s DGL efficiencies, LEDs will produce more lumens per Watt, but this will change in the future as lasers improve.   I fully expect to see eventually over 30 lumens per Watt with DGL and LCOS because today with the current DGL, they would be doing well to get 3 to 5 lumens per Watt .

Q7. What would be the lumen output from Syndiant’s displays using the new lasers that Microvision will use [this question was edited for clarity].

This is not a simple question as the spec’s on the lasers and how they can be driven have not been finalized.   I’m pretty sure the lasers that Microvision was using “lab prototype” lasers at CES that were being “over driven.”  The best I can answer right now is to give some insight into the lumens per watt that can be expected as the lasers are perfected.

To begin with, the optical throughput for a laser/LCOS engine should be in the 30% to 40% range where for LBS the optical throughput is reported to be in the 55% to 60% range.   But note, this is the part of the system were LBS looks best in terms of efficiency, but this does not tell the whole story.

Where LBS looses in terms of efficiency is in the drive and control of the laser and the MEM’s mirror.    The Microvision MEM’s mirror alone (not including the ASICs and FPGA’s) at WVGA has been reported to take about 0.4 Watts.   Syndiant’s new 720p LCOS Microdisplay and ASIC combined will take less than that.   Then you have all the power of the 2 ASIC’s and the FPGA that the Microvision 720p board requires.    So LCOS starts with a big lead in terms of power just in power of the display and control.

Then we have biggest power wasters for LBS, that of having to analog modulate the laser drive power.   First you have the fact that each pixel in the laser beam scanning process must be analog modulated at very high speed and high speed analog modulation wastes power.

Additionally, most people are not aware that the laser beam also has to be modulated due to the varying speed of the laser sweep.   If you think about it the laser beam horizontal sweep has to accelerate from zero to the maximum speed at the center of the screen and then decelerate to stop at the far side before returning.  To make a solid image appear uniform, the laser drive has to be constantly varying (for a “solid white” image the drive approximates a sine wave).  See such as the Microvision White Paper (a figure from which is copied below) and the excerpt (copied below) from the Microvision patent application”Apparatus and Method for Interpolating the Intensities of Scanned Pixels“:

From US Patent Application 20090213040

Q7. What is the power to run the PCB and the light engine on a Syndiant’s Displays? (SVGA and WVGA).

That really is a complex and involved question and depends on a lot of factors including the optical engine design and the target brightness.  I am also assuming that this is with LED illumination.   With LEDs the lumens per Watt of power tends to go down as the projector gets brighter, that is one of the advantages with lasers, namely that efficiency does not go down with power.

The LCOS display and ASIC for the WVGA or SVGA will usually consume less than ess than 0.4W.   Due to a number of design improvements, Syndiant’s 720p panel and ASIC consume less power while have a higher color field rate and better light throughput than the older WVGA and SVGA devices while having about 2X to 3X the pixels.

In the low lumen area of 10 to 15 lumens with a small, 0.21” WVGA or SVGA panel optical companies were able to get about 7 to 10 lumens per LED Watt.    With some power conversion overhead and including the panel and ASIC this mean that you could get with LEDs about 12 to 15 lumens for 2W of power.  While this is still not good enough for the very high volume cell phone applications, it compares very favorably to the ShowWX which takes 4+ Watts for about 15 lumens.

Q8. Which VGA display panel is best suited for the cell phone market?

I don’t know that there is a “best.”  Right now about the only reasonably high volume embedded panel is the color filter LCOS one by Himax that is used in cell phones for the India and China market.   The performance of these engines is too poor to be used in “first world” markets.   I think many of them are also less than VGA resolution.   They are typically about 5 to 10 lumens with pretty poor color and contrast and use very cheap but relatively large optics.

Personally, I don’t see a big “first world” need for a VGA or WVGA display.   Why bother projecting an image that lower in resolution than the cell phone’s display?   I think the big market will be to projector resolutions that are at least 720p

Q9. Does Syndiant plan on building the Light Engine?

I can’t comment on Syndiant’s future plans, but Syndiant’s business model has been to be a panel supplier.    There are very large number of good optical engine companies in the world so I don’t know why Syndiant would want to change.

Q10. Is Syndiant depending on someone else to make the Light Engine and Syndiant only sell the display panel?

Syndiant makes the panel and for use by many companies.   This allows different companies to make different products aimed at different markets.

Q11. Does Syndiant’s displays need a Vibrating Despeckle unit with DGLs?

It depends on the DGLs.  With the multimode DGLs you can drive them in such a way as to induce more mode-hopping to reduce speckle.   I would expect the there will not be despeckling required with volume production DGLs.

Q12. Does the light coming off the Vibrating Despeckle unit need to cover the homoginizer [a reference to the 3 year old SSTDC optical engine that used a vibrating despeckler]?

In the old SSTDC design with a vibrating despeckler, the light was partly spread before going to the despeckling mirror but not as large as the homoginizer.

 

Laser with LCOS is Focus Free — Yes Really!

Focus Free LCOS+Laser Projection (click for larger image)

Yes, when LCOS panels are used with lasers they can be “Focus Free.”   I have found that even very technical people have a hard time believing this as it goes against one’s everyday experience dealing with “normal” light and lenses.    People assume that the main function of a lens is to “focus light.”  After all, people are used to having to focus a camera lens or with a projector using lamps or LED light.

The optical physics of why it is focus free would take a long technical discussion, but it has to do with the laser light being effectively infinite f-number.  It is analogous to stopping a camera down to a high f-number where the depth of focus become very large.

Hopefully, “seeing is believing.”  I have uploaded a couple of still pictures (click on images for larger versions) and a short YouTube video demonstrating the focus free nature of the a Laseno Projector (sold in the U.S. as the AAXA L2).  This projector used a SYL2010 SVGA (800×600) LCOS plane with a 5.4 micron pixel pitch and 0.21″ diagonal.

For the top picture in this article, I projected the image the ceiling and some crown molding in my house.  This ceiling has lots of angle and different depths to it and with the crown molding in the way there is some obvious depth differences. Of course with the Laseno/AAXA L2 projector, there is no focusing necessary.

Projected at and angle to demonstrate focus free

Projected at and angle to demonstrate focus free (click on image)

For the picture on the left, I projected the image at a skewed angle from the side to cause a range of depths to be displayed.  The problem you have is that while the projected image is focus free, when the laser light hits the screen it looses it high f-number characteristics and thus the camera needs to focus.   By projecting the image on a flat piece of paper and shooting the picture straight onto the piece of paper I was able to focus the camera while demonstrating the focus free nature of the projector.

But perhaps the best way to demonstrate the focus free nature of a laser/LCOS projector is with a video.  I shot a short ~1 minute video where I mounted the projector on a little dolly and pulled it back away from the screen.  There was some shaking as I moved the projector so I stopped occasionally as I moved it it back so you could see it was still in focus.   I zoomed with the video camera in so you could see the detail in the video image.   Note that the camera’s exposures was locked/fixed on the starting frame, so as the image gets larger, it becomes darker by the ratio of the area so as the projector pull back the video gets a little dark.

I would recommend watching the video at 720p and in full screen to see how the focus is maintained.

Focus Free Video Demonstration

Other Information on the Images

The ~2 year old Laseno projector I used for these pictures has a fixed focus lens.   The image become well focused about 8″ from the projector to infinity.

The Laseno projection lens is not of high quality and you will see some serious chroma aberrations in picture as well as some spots having some blur due to the quality of the lens.  Additionally the projector has “100% offset” meaning that it projects through only the top half of the projection lens so that the projector will project upward from a flat surface without having a keystone effect.  Because of the offset projection, the image is best at the bottom of the image (which is from the center of the lens) at the chroma aberrations (color separation) become progressively worst toward the time.

You definitely will see laser speckle in the images.  The despeckle design was low cost and done over 3 years ago and it uses frequency doubled green lasers which inherently have a high amount of speckle.  Most people who have seen the AAXA L1/L2 “live and compared it to the ShowWX have said that the speckle with the Laseno/AAXA L1/L2  is less than that of Microvision’s ShowWX.