Pico Talk
Cool Uses for Pico Projectors
By Karl Guttag, Syndiant CTO
November 29, 2010
I just got back from my second trip to Asia in the last two months to visit customers. During this period some of what I had seen has turned into product announcements with Philips PicoPix, 3M MP160, and Aiptek T25. There is more coming that I can’t write about yet, but you can expect to see more around the time of CES.
What I wanted to write about today is Syndiant’s Cool Uses for Pico Projector Contest. Back when we first starting looking at doing microdisplays for pico projectors, we thought about how they would be used. What we found is that there are many excellent ideas on how they will be used. Some saw it as a video machine for teens to watch videos and instant sharing of photos, others saw it as something that business people would give presentations with, some like the idea of web browsing, and the list of uses has kept expanding.
I’ve been in the semiconductor industry since the 1977 and have helped create integrated circuits used in the early home computers and video games and I have seen how things change as technology improves and costs come down. I can see this happening with pico projector products; we are just at the beginning and we will soon see HD products, major improvements in efficiency (e.g. Lm/W), reductions in size, and falling costs.
I think it is very realistic that in a few years we will see 40 lumens-per-watt, with HD resolution, less than 3 cc in volume and 6 mm thick for the optical engine and projection electronics, and add less than $25 to the cost of making a product. Additionally we expect to see very “short throw” optics where large images can be projected from short distances. We expect to see “visual input” where cameras support virtual keyboards and mice by looking at gestures. When you factor in so much image capability in such a small space, the possibilities for new uses of pico projector are nearly endless.
We thought the Cool Uses Contest was a way to get people thinking about all the new uses. You don’t have to worry about how the projector will work; you can leave that to Syndiant and our partners. Also you don’t have to be bound by what current products are able to do. What we are looking for is what you think people will do with the pico projectors today and in the near future when you have high resolution low cost devices that fit in your pocket or are embedded in devices the size of a cell phone or small camera.
Get your ideas in by the contest deadline of December 15, 2010!
2010 SID and Projection Summit
By Karl Guttag, Syndiant CTO
June 29, 2010
The last month was very busy for Syndiant, presenting and exhibiting at both SID and Projection Summit. At SID, our CEO Mark Harward presented at the DisplaySearch Business conference and the Cowen & Co. Investors Forum, while I presented an invited paper at SID Symposium titled “High Resolution Microdisplays for Pico-Projectors.” At Projection Summit I presented “Why Resolution Matters.”
As you can tell from the title of my papers, I’m a big believer that pico projectors are going to have ever higher resolution. In the not too distant future we will be seeing 720P and eventually 1080P resolutions in pico projectors. Here at Syndiant we are hard at work on our HD generation of products that we expect to sample this year and go into volume projection in 2011. I plan to write more on HD in future blogs.
While Mark and I received a lot of positive comments about our presentations, the big hit for Syndiant at both conferences had to be our Six-Way Pico Projector Technology Comparison. From some media reports that came out before the show, particularly by Insight Media and PicoProjector-Info.com, we had a number of people seeking us out. In the pre-show reports, we released a photograph showing we were driving all the projectors - including those that supported WVGA (848x480) - with an SVGA (800x600) source, but at both SID and Projection Summit we connected a second computer that output native WVGA for the two WVGA projectors in the comparison. Additionally, we had some competitors come by and we let them verify that the projectors using their devices were in the high brightness and best color modes.
Below is picture I took of the setup at Projection Summit. If you click on the picture below it will take you to a higher resolution version that will let you see more detail. The technology used in each projector is labeled either above or below the projected image. There are no “Photoshop Effects” used to make any of the projectors look better or worse (the image was straightened and cropped from an 18 megapixel original). I did read an on-line comment by someone who had seen the pre-SID picture that Syndiant’s image looked “too good” and must have been “Photoshopped.” This gave me an idea for a marketing slogan, “Image quality so good, you will swear it was Photoshopped.”
(Click on photo to view larger version)
Syndiant’s pico projector technology comparison had a mix of image content from personal to business applications so people could see the differences in the technologies in various applications. Shown above or below each projected image are the approximate electrical power consumption and light output.
Syndiant is an LCOS microdisplay manufacturer and the purpose of this comparison was to show the differences in the various technologies inside the projectors and not the projectors themselves. The projectors represent the most common technologies that are used inside pico projectors today.
All the projectors selected are within about 20% of the same power consumption (about 4 to 5 Watts for the whole system) and the distance of the projector to the screen were adjusted so that the areas of the various projected images are about the same. This allowed people to have an idea as to the efficiency and brightness of the various technologies.
The display technologies in the projectors are (top row left to right and bottom row left to right in the picture above):
- DLP® WVGA - The new 848x480 DLP® based projector. Interestingly, the WVGA DLP uses a “diamond” pixel arrangement that causes a number of artifacts including “sharpening halos” and diamond jaggies (particularly noticeable in text) and requires scaling which causes lower effective resolution (see the “Resolution Chart Close-ups” below).
- Syndiant’s SYL2041 – Our projector used the same LED and optical engine as the 640x480 FLCOS projector directly below it, but we replaced the FLCOS device with our VueG8 microdisplay. Syndiant’s VueG8-LCOS is able to demonstrate about a 1.4X improvement in the lumens per Watt and a 1.56x improvement in resolution over the FLCOS device it replaced. The SYL2041 has 1.18X more pixels than the WVGA-DLP and laser beam scanning projectors.
- DLP® HVGA – This is demonstrating the DLP at 480x320 with a roughly 7.6 micron “square” pixels. This projector gave a graphic demonstration of why resolution is important as this projector has about 1/3rd the pixels of the SYL2041. While a newer nHD (640 x 360 pixels) DLP has been announced (but was not available), it will only increase the resolution 1.5X still leaving less than half the pixels of the SYL2041 and less than a quarter the pixels the HD resolutions that Syndiant will be sampling later in 2010.
- Laser Beam Scanning (LBS) – This technology uses 3 fast switching lasers beams and a scanning/steering mirror to direct the beams. The image quality issues people may notice include speckle (noise) particularly in bright/white area, image distortion (“bow tie effect”), flicker/flashing lines (particularly on the outside of the image). It was also a surprise to many people how much power the LBS consumes and how hot it gets with less light output than the LED based technologies. LBS scanning is likely to be the most expensive of the technologies shown due to the comparatively high cost of fast switching lasers as well as and the electronics to control the beam scanning mirror and the lasers.
- FLCOS - The 640x480 (VGA) Ferroelectric LCOS (FLCOS) was the first color sequential LCOS pico projector on the market. Using the same LEDs and roughly the same power and optics as this FLCOS projector, Syndiant’s SYL2041 delivers about 1.4X the lumens per Watt using the same optics and 1.56X more pixels in the same space.
- Color Filter LCOS (CF-LCOS) – This 640X480 (VGA) projector demonstrates some of the color issues with Color Filter LCOS. Most of the earliest so called “LCOS pico projectors” used this technology which had comparatively poor color due to blending between the adjacent color filtered sub-pixels. Unfortunately this has caused some people to associate “LCOS” with “poor color.” This technology comparison will let people see that Syndiant’s LCOS technology has well saturated colors.
Resolution Chart Close Ups
Shown below are close-up pictures of a resolution chart from Syndiant SYL2041, the DLP® WVGA with Diamond pixels and the Laser Beam Scanning projector that I think dramatically shows the difference between claimed and effective resolution. We drove all three projectors with their claimed native resolution.
The SYL2041 has no problems with either horizontal or vertical 1 pixel wide lines. The lines are equally well resolved in either direction. The image is clean with no noise or jaggy artifacts. Not only does the SYL2041 have nearly 20% more pixels, it has “better pixels” as the side by side comparison demonstrates.
The Diamond Pixel DLP is clearly doing different processing horizontally and vertically to convert the “diamond” pixels to square pixels. The processing tends to blur the 1 pixel wide vertical lines together while the horizontal lines have severe jaggy artifacts that are different for every line. Also note on the top of the image where a group of 4 horizontal and 4 vertical lines cross, the processing results in back dots where they should be white.
The Laser Beam Scanning projector demonstrated much lower resolution on horizontal lines than vertical lines. As can be seen in the picture, the vertical lines tend to blur together. Additionally there is a color moiré pattern in the vertical lines as well as some pretty severe speckle/noise in the image.
(Click on photo to view larger version)
Fun on a Garage Door
By Karl Guttag, Syndiant CTO
October 7, 2009
(Click on photo to view larger version)
I was playing with one of our new engineering prototypes around the house and it was fun to project large images on walls from a small hand held device. I wanted to give you an idea of the size of the images and I hit on using my garage door which is a standard 8 feet wide (or about 120 inches diagonal) to show an image filling the door to full width. For added measure I decided to stand in the picture to give an idea of the scale. There is a streetlight that shines reddish light on our garage which is why you can see the brick and frame of the garage door in the pictures.
The prototype projector generating the image is the small gold object in the bottom of the picture and is about the size of an iPhone®. It is generating an 800x600 pixel image (SVGA) with about 17 lumens of light output using LEDs and is powered by an internal cell phone type battery. I am pointing to the Arial 8 Point font which is clearly readable down to the dots in the letters “i” which are a single pixel. You can also clearly see all the horizontal and vertical gray lines in the spreadsheet which further demonstrates we are doing the full SVGA resolution using our SYL2041 panel. While this is an engineering prototype, we expect that the SYL2041 panel will be used in LED pico projectors with a range of light output from about 20 to over 50 lumens.
Also visible in the picture are the seams, paneling, and door handle “texturing” from the garage door itself as well as a little shadowing on the right side of the image where the bricks block the streetlamp. While not the ideal environment for showing projected pictures, it is fun to see such big images from such a small device (my neighbors were certainly impressed). It’s easy to imagine that soon pico projectors will be used to decorate homes for parties and holidays in addition to watching videos, showing pictures and “serious” applications like business presentations.
Spreadsheets are a simple way to demonstrate resolution, but pictures of people are a better way to demonstrate color. We have been all over the world with our microdisplays and people marvel at our color control particularly for flesh tones. So I also am posting a picture of my kids projected on the same garage door.
(Click on photo to view larger version)
Recent Syndiant Happenings and Pounding Pixels
By Karl Guttag, Syndiant CTO
September 12, 2009
It has been a very busy time here at Syndiant. In the last week we:
- Announced three new microdisplays, the SYL2030 6mm tall 0.21” diagonal 854x480 (WVGA), the SYL2041 854x600 0.37” diagonal (SVGA and WVGA), and SYL2061 1024x600 0.44” diagonal (WSVGA)
- Announced the SYA1012 ASIC controller
- Announced a volume production order of our SYL2010 for the Laseno™ focus free laser projector with SVGA resolution (manufactured by SSTDC in China)
- Demonstrated the new SYL2041 SVGA imager in a prototype LED projector and the Laseno projector at the DisplaySearch’s Emerging Display Technologies Conference in San Jose September 3, 2009.
It looks like my last blog stirred up some discussion and I answered a number of questions put to me by picoprojector-info.com. One of the questions put to me was whether it was true that our microdisplays would work and be focus-free with laser illumination, well; the Laseno projector demonstrates that it works and it is focus-free. I also said that we would soon have higher resolution and smaller microdisplays which our product announcements that we are sampling the world’s first Wide-SVGA 1024x600 SYL2061 microdisplay aimed at higher brightness LED accessory projectors and the 6 mm tall SYL2030 aimed at embedded pico projectors.
While we are proud to enable the Laseno, the first high volume laser pico projector, we expect that the majority of our customers will use LED illumination for the next several years due to the relative cost of LEDs versus lasers. With LCOS and our small pixel digital drive technology we can easily scale our pixels to meet our customer’s requirements for both laser and LED illuminated systems.
Without giving too much away we are continuing to drive down power and size while supporting ever increasing resolutions in even smaller sized devices. In particular, we expect to have 720P devices small and low power enough for embedding in cell phones in 2010. Syndiant will continue its leadership in bringing higher resolution to pico projectors.
Using our all digital LCOS we have a big advantage over other technologies in that we are leveraging high volume CMOS silicon factories to make small pixels. This enables us to affordably build small high resolution devices. While our current smallest pixel is 5.4 microns on a side, we are developing smaller pixels to give more pixels in the same area. At the same time our CMOS design uses relatively low voltage to support low power per pixel. As resolution increases, the “power per pixel” becomes a big factor for embedded applications.
In any discussion of resolution it is important to understand the difference between the “input” resolution and the “effective” or delivered resolution. Most display technologies today such as LCOS and DLP® (and LCD monitors) have pixels on a fixed grid so it is easy to count pixels to tell the resolution. But for scanned beam technologies like CRT’s and Laser Beam steering, the “effective resolution” is often lower due to beam switching speed and beam scanning accuracy among other factors. For example, I have read of companies claiming 854x480 WVGA “resolution” and then say they can (only) display a 10-point (MS-Arial or the like) font. But a “true” WVGA device should be able to display a readable 8-point Arial font. While this may not seem like a big difference, the “effective resolution” of a device when trying to display WVGA that can only clearly display a 10-point font is really only about half that of one that can clearly display an 8-point font.
The table below shows the most of the common resolutions being consider for pico projectors today and in the future. I’ve also included a column showing the Number of Pixels Relative to WVGA to help with understanding the relative resolutions of the various display sizes. Also shown is the aspect ratio, the ratio of horizontal to vertical pixels.
Our first product, the SYL2010 could perform either WVGA and SVGA resolutions to support wide format video without a loss in resolution for standard definition NTSC television (SDTV) as well having enough horizontal pixels to support web browsing and reading .pdf-type email attachments. It also supports SVGA because this was considered the minimum necessary to support business graphics presentations. One common theme we were told by those in the cell phone industry was that they expected pico projectors to support higher resolution than could be view on a small LCD screen. We were commonly told that WVGA was the minimum resolution that a pico projector was expected to deliver (and thus the reason for the comparison to WVGA in the table above).
The new SYL2030 is more optimized than the SYL2010 for embedded use and only supports WVGA. The SYL2041 supports a bigger pixel size with SVGA resolution to support higher light output with LEDs. The SYL2061 supports even brighter projectors and increases the resolution to Wide-SVGA.
I have been involved in video and graphics for over 30 years and one thing that I have always seen is a desire for more pixels. Part of the drive for resolution is content and user interfaces. You can see this trend in computer monitors that started with less than VGA resolution and now routinely provide 1080P and higher resolution. Just try browsing the internet in a window on your desktop at less than 1024 pixels wide and you will find that most normal websites become hard to use because they are designed assuming higher resolution. Netbooks typically have Wide SVGA resolution while laptops typically have WXGA+ and higher resolution. Televisions and video cameras now routinely have 720P and 1080P resolution, not to mention still camera resolutions which are typically many millions of megapixels.
And then bucking the trend toward ever higher resolution we have nHD which we understand is being heavily pushed by TI’s DLP® for their next microdisplays for pico projectors. “nHD” stands for “one-ninth 1080P HD” since it has 1/3rd the pixels vertically and 1/3rd horizontally (or more simply “not-HD”). As the table above shows, it has just over half the pixels of WVGA and is 1/4th that of 720P. This resolution is lower than even standard definition TV so that even SDTV will have to be scaled down (which blurs the images) to be projected at nHD. nHD is too low for doing typical business presentations and browsing the internet will be difficult at best. One can only assume that DLP is dropping down to nHD, in spite of the market wanting higher resolutions, due their higher “power per pixel,” comparatively large pixel (about twice the area of Syndiant’s SYL2030 pixel), and higher cost per pixel.
In addition to have higher resolution than competing solutions, we have much better/finer control of colors. This results in much more lifelike images, particularly on skin tones. Color depth/control is a whole other topic to discuss in the future.
In conclusion, there are strong and rational reasons that will continue to drive higher resolutions and Syndiant is working on higher resolution devices while others seem to be going backwards in resolution or saying that resolution really doesn’t matter when they can’t demonstrate the resolution they claim to have. Or as one person commented at a show, “we know that with your technology you are going to pound them with pixels,” to which I could only comment, “Yep, that pretty much sums up our strategy!”
Reflections From Asia on Pico Projector Market Take-off
By Karl Guttag, Syndiant CTO
June 26, 2009
Syndiant has developed microdisplay technology to meet the size, cost, power and resolution requirements to embed pico projectors in consumer and business products. We are currently working with more than 70 companies including a number of cell phone and consumer product manufacturers, as well as optical design OEMs and the makers of LED and Laser light sources.
For my first blog I would like to talk a bit about some conclusions I reached based on customer meetings and conferences I have attended in Asia and the United States over the last several months. From my interactions with customers and industry insiders, I am finding that the reality of this market may differ from what you may have read in reports on heavily promoted product concepts.
By November 2008, we were being told by our customers in Asia that the pico projector market is going to be significantly bigger than any of the forecasts from market analysts. Rather than a slow progression from stand alone projectors to embedded pico projectors in cell phones, we found that multiple cell phone companies were gearing up to have pico projectors as the next “must have” feature in a cell phone. But while the size of the market appears to be increasing, the equipment makers were still in need of a small, high resolution and cost effective microdisplay to enable market take off.
While there are some interim products based on technology such as color filter LCOS and DLP® that were in a position to get to market earlier, they have major drawbacks in fulfilling the size, cost, power and resolution requirements of the market, particularly for embedded devices. We are already seeing color filter LCOS products being designed out in favor of field sequential LCOS and many former “DLP-only design houses” also have field sequential LCOS designs underway. While laser beam scanning has been heavily marketed, most people we talk to in the industry do not see it as a viable technology for a number of reasons including cost, speckle, and the limit on brightness due to eye safety. For explanations of the various technologies please see here.
As the OEMs and end product companies looked more closely at the requirements for pico projectors, it became increasingly clear that the small pixel, field-sequential-color LCOS device is the best technology to meet all the requirements for embedded products. Furthermore, Syndiant was the only company demonstrating a small pixel LCOS device with the cost, size, resolution and image quality that the equipment makers could use to make very high volume products.
The list below summarizes what I have been hearing relative to competing technologies:
- DLP is going to be too expensive, take too much power per pixel and be too low in resolution to have a long life in the market. It will be designed into some early products due to the established support base of Texas Instruments, but even design companies that used to be exclusive “DLP houses” are now designing for LCOS. Currently, the smallest DLP pixel size is about 2 times bigger than Syndiant’s pixel.
- Laser Beam Steering (LBS) is far too expensive, has electronics that are large and power hungry, has high speckle noise, low color depth, poor bright field uniformity, scaling and distortion that lowers the effective resolution, brightness that is limited to ~10 lumens due to safety regulations, and the list goes on. Beyond well known cost issues, the image quality demonstrated to date also has a lot of issues, not the least of which is speckle. Apparently, its circuitry can’t handle too much white on the screen in a single image as they don’t show typical PowerPoint and MS-Word type images which usually have white/light backgrounds. The PowerPoint and text demos made in public to date only show lower resolution text shown on dark backgrounds. On request they did put up a “white” background, but it had random rippling in it suggesting that something in the circuitry couldn’t handle driving large areas of white.
- Color Filter LCOS is too big, the color saturation is poor and it does not scale to high enough resolutions to be practical in the long run. While today most pico projectors are using this technology due to its early availability, the large size and poor image quality is causing it to be designed out in favor of field sequential color LCOS.
- FLCOS is a near-eye (viewfinder and headset) technology that is being adapted to pico projectors. It uses a type of liquid crystal that, while fast switching, only allows the light to be on for about half the time - a big drawback for brightness in a projector. Additionally, the smallest available pixel sizes are almost 5 times bigger than Syndiant’s current pixel which makes these FLCOS devices far too large to support WVGA resolution in a cell phone.
In addition to keeping in touch with our customers and watching the competition, we are in constant communication with optical design houses and companies developing LEDs and Lasers. While our microdisplay can work with LEDs or lasers, LEDs currently have a very large cost advantage over lasers for pico projector applications.
Leveraging developments in LED lighting, color LEDs continue to show remarkable improvements in light efficiency and continue roughly doubling the light output per Watt every two years. We fully expect there to be pico projectors with greater than 10 lumens consuming less than 1.5 Watts using Syndiant’s microdisplays by early 2010. With the continued improvement in our devices, as well as improvements in LEDs and optics, we expect to have greater than 20 lumens per at 1.5W for embedded products with LEDs by 2012. With the cost advantages of LEDs and the light output efficiencies possible, LEDs combined with our panels will soon enable high unit volume embedded pico projectors.
Laser light has some unique properties that make it interesting as a longer term light source as the cost of lasers comes down. The optical properties of lasers when used with our microdisplays will provide focus free operation and even smaller optics. Lasers have advantages when going for higher brightness and even higher efficiency. Lasers with LCOS can go to higher brightness levels without the eye safety issues of laser beam steering because when used with LCOS the laser beam is spread out over hundreds of thousands of pixels rather than laser beam steering’s use of a single pixel sized beam. It is possible using lasers to reach over 20 lumens per Watt today using our microdisplays and as the green laser improves, we expect this efficiency to grow to around 40 lumens per Watt. To give an idea what this means, a 50-inch HDTV would requires less than 20 Watts of power.
I hope you found this interesting. Be looking for my next blog in which I’ll discuss resolution.
