NEW YORK – With most flat panel TV vendors expecting to see new growth opportunities blossom from 3-D TV, the likelihood grows that 3-D image/video capturing capabilities will become a nice-to-have feature for digital video camera suppliers in the next several years.
Anticipating the trend, Ambarella, Inc., a Santa Clara, Calif.-based chip vendor that specializes in low-power, high-definition video compression and image processing solutions, unveiled Monday (Nov. 22nd) a 3-D video pre-processor, called S3D. Designed to work with Ambarella’s camera SoCs such as A5 (enables 1080P30 resolution) or A7 (enables 1080P60 resolution) – announced last month, the combination will enable full HD 1080p 3-D video recording and high-resolution 3-D photography, according to the company.
The goal of the combined S3D with Ambarella’s A5 or A7 camera SoCs is to enable a consumer-friendly 3-D digital video camera – priced at less than $200 – capable of shooting 3-D video “that looks really good on 3-D TV,” said Chris Day, vice president, marketing and business development at Ambarella.
Ambarella’s 3-D camera solution also supports the display of 3D video and still images on the camera’s 3D LCD display.
Ambarella’s 3-D video pre-processor, which connects to two standard CMOS image sensors, will combine two video streams into one. Then, the video stream will be processed by the A5 or A7 camera SoCs.
S3D, consisting of many video logics, contains no CPU of any sort. Instead, the video pre-processor is under the control of the camera SoC, according to Ambarella’s Day.
S3D is designed to perform a lot of “clever scaling,” said Day. Declining to detail what makes S3D superior to others, Day simply noted that quality is the biggest challenge in such a 3-D video pre-processor. “Obviously, when you combine two video streams into one, something has to give. Preserving the video quality was our number one priority.”
Low power consumption is also touted as a hallmark of Amberalla’s digital camera solutions. Combined with S3D vide pre-processor with A5 and DDR3 memory, power consumption is approximately 1W, according to Day. The combination of S3D, A7 and DDR3 memory consumes about 1.5 W, he added.
Thus far, Ambarella’s focus has been on the digital video camera market, rather than digital still cameras (DSC) whose market Zoran dominates. As more digital still cameras add HD video recording capabilities, the boundaries between the two have already begun blurring, creating an opportunity for Ambarella to enter the DSC market. But that said, Zoran is a formidable competitor. Zoran, who claims its camera SoC solutions are installed in hundreds of millions of digital cameras, announced last month that the new COACH 12 family of digital video processors is “powering a host of new HD camcorders under well-known brands, including wireless Internet connected and 3-D video camera models.”
How big a market will 3-D video cameras become?
However, still remaining to be seen is exactly how big a “nice-to-have” feature 3-D video/image capturing might become on the digital video camera market of the future.
Chris Chute, research manager at IDC, estimated that 250,000 digital cameras featuring 3-D (with two lens assemblies for capturing stereoscopic images and video) will have shipped in 2010. That number “will grow to around 6 million units in 2014, or five percent of the digital camera market,” he projected. IDC’s numbers only cover cameras -- not camcorders, according to Chute.
He added, “Right now, Fujifilm is the only vendor supporting this segment, although there are others in the camcorder space.” Those offering 3D-equipped camcorders include Panasonic, Aiptek of Taiwan and Viewsonic, according to Michelle Abraham, principal analyst at In-Stat.
Asked about the most important elements for 3D cameras, IDC’s Chute pointed out “the ability to view content instantly on the device via auto stereoscopic LCD. This feature is an absolute requirement.” He then added, “Also important is the ability to view on other displays.” However, he called it also “an inhibitor,” since 3D TV shipments remain quite small and the requirement to wear glasses is also an inhibitor.
In-Stat’s Abraham, on the other hand, noted that “price will be important as it always is in consumer products.” She added, “Video quality will be as well [important] especially with some consumers saying 3D gives them headaches or nausea. The better the quality the less likely that is to happen.”
Ambarella today has a working chip of S3D, whose volume production starts in the first quarter of 2011.
3D is indeed getting tiring, but at the same time somewhat exciting since, at the moment, it seems to be the only place for the video industry to go. Higher resolution, frame rate, etc., only do some much on smaller screens. After all, how many of us can actually put the 2 or 3 story screen that can take advantage of those into our homes? On the other hand, a lot of people have problems with 3D since it is not quite real and those differences cause some interesting visual issues, especially when viewed from off-center.
There is japanese brand who has released a 3D camera. You can take 3D photos and see 3D photos on the tiny LCD w/o wearing glass. Personally, the idea is exciting. The product is fascinating; yet, it doesn't provide extra information. I believe the currently 3D is a transition to a real 3D that Sony is developing. http://www.youtube.com/watch?v=lAS55_RngoQ
The problem with high quality optics is that the components which enable them don't go down in price as much as the silicon inside the camcorders as time progresses.
I think Junko mentions $200 or lesser as the price point of these camcorders (which is equivalent to the Kodak Playsports, Sony Bloggies and Flip cams of the present day). Essentially, we should get a free ride to 3D at the same price point sometime in the future ( at least that is what I am gathering from this report :) )
As for higher video quality, longer battery life and higher frame rates, these have remained the bastion of $500 or higher priced camcorders, and the manufacturers probably want to protect their margins. That market is probably low volume compared to the $100 - $200 priced impulse purchase shoot-and-share market.
Ganesh @ AnandTech
Traditionally we think of 3D as deriving from combining simultaneous images from different vantage points - but there are other ways of detecting depth. If we move about, near things move differently from distant things. The focal point also varies depending upon the distance of two objects. If the light source is near the camera, distant objects will also be less well lit. Instinctively we're aware of these effects (you really can close one eye and walk around without bumping into things). It will be interesting to see if 3D simulation cameras become available that utilize computation rather than stereoscopic images to create 3D effects. Such images might also not require special projectors (if they use focus and other effects rather than parallax to create 3D images).
I for one would love to see what the cost is and if reasonable, use it for hiking in the mountains. What a great way to capture the visual impact of views. I have never gotten quite the same feeling from pictures and wonder if 3d video is just the trick? I also could see opportunities for training 3d videos, there is nothing like a perspective when trying to learn something (especially if using the 1st person perspective with narration).
Yes, there is correlation between the two streams. The H264 MVC specifications allow this spatial correlation to result in more efficient encoding (thereby reducing bandwidth requirements / storage space).
An intersting aspect that Junko could probably shed light on is whether Ambarella's S3D + A5s / S3D + A7 combination takes advantage of MVC encoding to reduce storage requirements. Also, what semiconductor solution does the Panasonic 3D camcorder use for this purpose? The FujiFilm 3D camera doesn't shoot videos as far as I know, so that is of secondary interest here.
Ganesh @ AnandTech
Yah, 3D hoopla does get old. But sometimes, the images are just wonderful. But to business: the two-into-one doesn't go bit. When the video streams are combined, surely no one wants to deal with twice the bandwidth. So here's the obvious question, which may have be answered about a thousand times: what is the correlation like between the left and right images in the 3D stream? Surely it has not escaped anyone working in this field that the correlation is very high, and this is always a recipe to keep bandwidth down. BTW, a comment to Duane: much of what we see and do is in 2D effectively. Past 10 meters, binocular accommodation is negligible; those cues are missing. So shooting at someone or landing an airplane on a carrier or something similar is a 2D task. Curiously, at 3D movies, an important cue goes missing: focal accommodations. In a 3D film, your eyes focus on the screen, effectively at infinity. So the brain is presented with something that cannot be right: binocular accommodation, but no focal accommodation. Still, 3D seems to work... I think the young crowd is better than us old guys at this stuff because the younger crowd has the advantage of: youth.
Join our online Radio Show on Friday 11th July starting at 2:00pm Eastern, when EETimes editor of all things fun and interesting, Max Maxfield, and embedded systems expert, Jack Ganssle, will debate as to just what is, and is not, and embedded system.