Thursday, October 06, 2011

Samsung Presents 1/3-inch QVGA ToF Sensor

Samsung published Youtube video presenting the new 1/3-inch QVGA ToF Sensor - the S5K32D:

Samsung S5K32D QVGA 1/3" TOF Sensor is said to have high resolution and high speed for Z-axis User Interface. The text abstract also says that it's:
  1. World's first QVGA resolution TOF Sensor
  2. Supporting high speed frame rate
  3. One Chip solution for depth image

Update 2012/02/20: The updated Samsung web site lists 3D ToF part as S5K32A. The pixel uses BSI technology and is of 14um size.


  1. Vladimir, please don't put such propagand video in this blog.

  2. What's wrong with the video? Seemed ok to me.

  3. "propaganda" seems a bit harsh. Everyone understands it company marketing material - does not mean looking at is is without value.

  4. Propaganda or not, this is the only material available on the web that talks about the new sensor and gives its part number.

  5. Are there any nice, recent, tutorial papers on ToF sensors?


  7. There really is not that much new in this video and certainly almost no technical information. Some early sensor data was presented at 2011 IISW.

    But, the technology was transferred from the Semiconductor R&D Center to the Systems LSI product group and now it has a life of its own.
    Frankly, it took us a while to catch up to the performance of pioneering TOF work done by many groups around the world since a cold start when I joined Samsung in April 2008. Now I think we are competitive and our IP unique enough.

    The new target is RGBZ as previously announced by Samsung. Maybe a couple of the RGBZ papers submitted by Samsung to ISSCC will be accepted...

    Back to the video. What I like are the illustrations of the consumer applications enabled by ranging sensors. I think this is just the beginning and I hope some systems people are further inspired.

  8. I am very skeptic about this type of consumer applications. Who does really want waving his hands in front of the tv?

    Why don't we see any performance data if it is true that Samsung has actually caught up with the pioneering TOF companies? There is no data around at all, ... but good marketing, indeed.

  9. This kind of marketing propagand is really a pitty for technical community. All the scenes in this video are fake except one where the guy moves so slowly his hand in front of the TV ...

  10. The message I got from this video is that Samsung is not creative and does not pay attention to details. In these regards, the video is an unmitigated disaster, not a source of inspiration.

    Here are some of the selling points, verbatim:

    "You don't need to search for the remote control to change the channel anymore."

    While I sometimes have to look for the remote control to turn the TV on in the first place (because my kids move it) I don't have difficulty finding it to change the channel. Even the first search is little more than a minor inconvenience.

    My impression is that it's a case of a problem (a non-problem, really) being forced to fit a solution, rather than a solution solving a real problem.

    Not creative.

    "Furthermore, this device can make you a magician in driving and gaming"

    So what? The video of the "magician" didn't show any compelling advantage to simulated driving with no controller over simulated driving with a fake wheel, wired or not.

    To me it seems like this is appealing to an audience of young technology enthusiasts who absolutely must have the latest gadget, whatever it happens to be.

    "You can design your set more smaller and fancier"

    Again, so what? And, why does Samsung tolerate grammar errors that my son's third-grade teacher would not?

    "Samsung S5K32D, QVGA 1/3" TOF sensor, will help create excellent gesture recognition User Interfaces!!"

    This is the closing catchphrase. It has none of the clever simplicity of the most basic advertising jingle.

    "You can make natural movement at high speed video"


    This translation is rather unfortunate, because "make natural movement" is an awkward way of saying "defecate" in American English. If it weren't so funny, I would have to admonish Vladimir for publishing culturally insensitive and insulting material.

  11. I agree, it's crappy video..
    .. but I'm not sure if people understand how important this announcement can be. If the price will be acceptable, PrimeSense (or Kinect) will have its first real competitor (finally!).
    Honestly, who cares about next IS with larger resolution and (maybe) lower noise? The market is driven by short-sighted people.. There are really few "different" IS in mass production. IS community does not see all possibilities, applications of computer vision and computational photography; on the other hand CV community does not understand how difficult it is to design and built any IS. So I really appreciate this quite unique announcement.

  12. But Caeleste has designed a VGA resolution tof, no?

  13. It would be bad business plan for a boutique image sensor company to compete against Samsung or Panasonic etc. in this market. ASPs will drop like a rock. On the other hand, 3D camera companies like MESA will be able to build less expensive cameras and help drive an elastic market, or retreat to very high performance, expensive systems.
    Maybe Samsung will drop out though. You never know.

  14. Right. Plus with TOF pixel design and process technology in hand scaling to larger arrays is fairly trivial and is just market driven.

  15. QVGA resolution is nice but not a world record. At least CIF resolution is accomplished.

  16. I expect it is a global fitness tech or at least one little possibility to force Americans for some hand fitness :)

  17. I will try to keep this short because my wife is making me dress to go to church this morning instead of sleeping in, but...

    To whom it may concern (you know who you are):

    For me, an application for this chip that is much more creative than replacing a TV or video game controller would be a system that lets blind people see.

    Specifically, I would be very impressed at this chip in a hand-held camera, with the output used to servo a hand-held tactile image display. The ToF depth information would drive height changes in the display, allowing a blind user to feel on it what the camera sees.

    Even better, there should be an option to map color and contrast information to the tactile image display - by height, tactile pixel vibration, or temperature (for instance).

    Then the marketing video could be someone who is blind asked to pick up the Samsung baseball cap and put it on, and using the system to reach past the delicately balanced glassware to get the right hat, pick it up, and place it on her head without touching anything else.

  18. Interesting. I had a very similar idea many years ago (with a tactile MEMS display) but with just visible image sensors. I had not thought of doing it with 3D TOF sensors.


All comments are moderated to avoid spam and personal attacks.