Friday, June 01, 2012

Emmy Award for CMOS Sensor Advances

I've received an email on investigation of "Improvements to CMOS Imagers for Use in High-Definition Broadcast Video Cameras" for possible Emmy-award recognition. It might make sense to open the discussion, so everybody can express her/his opinion, anonymous or not. One of the major questions the investigators are interested is "whether the improvements to HD CMOS sensors for use in broadcast video cameras "materially have affected television."

2012 Investigative Subcommittees
Subcommittee Investigating Technology 28
“Improvements to CMOS Imagers for Use in High-Definition Broadcast Video Cameras”
Chair: Mark Schubin, Co-Chair: David O’Kelly

2012 May 30

Dear colleague:

We seek your help on the above-mentioned investigation. Here is the scope of these Emmy Awards:

“An award to an individual, a company, or a scientific or technical organization for developments and/or standardization involved in engineering technologies that either represent so extensive an improvement on existing methods or are so innovative in nature that they materially have affected television.”

Our main committee voted to investigate “Improvements to CMOS Imagers for Use in High-Definition Broadcast Video Cameras;” our subcommittee is conducting the investigation.

We have two questions we need to answer: “Have improvements to CMOS imagers for use in high-definition broadcast video cameras” materially affected television?” and, if so, “Who deserves Emmy recognition for pioneering that material effect?” When considering material affect, please note that the committee includes in “television” video systems that are not necessarily associated with television broadcasting.

There is no reverse time limit or geographic limitation to our work. Whatever we consider, however, must already materially have affected television.

We will present a report to the full committee based on our investigation, and they will then vote on awards. There is theoretically no limit to the number of awards for this technology. There could also be no award.

You may send us as much or as little as you'd like to help us answer those two questions. Please respond no later than July 20 to allow us time to write our report and submit it. The earlier the better, and there's no need for any formality. Feel free to submit partial, incomplete information. Please send everything to both of us (e-mail is fine).

Many thanks!

Mark Schubin
Technology Consultant

David O’Kelly
Canon U.S.A., Inc.


The emails are withheld and available on request, in case somebody wants to send the answer directly.


  1. I would say "No" and "Who cares?"

    That said, it should be fairly easy to find a company willing to send someone to a gala event to pick up an award and then promote to the world their "Emmy-award-winning" technology.

    "No" because HD broadcast is basically a set of resolutions and frame rates that maybe ten years ago seemed like they'd create a challenging data rate to acquire, transmit, and display. But CMOS image sensors were already around, and the progression to HD-spec imagers wasn't limited by some fundamental imaging technology roadblock. This is especially true for studio video cameras, which don't have the same constraints on size, power consumption, and cost that a lot of other systems do. It was certainly clear that with shrinking feature size, one could make imagers with more pixels and higher frame rates, so I don't see HD capability as a breakthrough so much as an obvious progression.

    "Who cares?" reflects that for me the broadcast television model is broken. There is a startling difference between older non-HD video and HD video. However, I only notice it when I've been watching one and then switch to the other. After a few minutes the gee-whiz feeling wears off and I don't remark on the quality.

    Even worse - for the broadcast industry - is that digital broadcast has changed my viewing habits. Where I used to be a regular consumer of broadcast programs and the accompanying commercial content, there is now so much material available digitally that I'm watching almost no TV. It's always there if I want it: start and stop, in whatever amount I want, and the program of my choice; and so for some reason I'm just not very interested in seeing content and advertisements someone else has picked out for me.

  2. HD broadcasts promoted a sea-change in technology across the spectrum. Personally, I always tune to HD channels on my TV, although I try to record most of what I actually watch.

    I think it is terrific that to the extent CMOS image sensor technology has made an impact on the broadcast industry, those companies should be recognized. I don't know what the market share of CMOS is within broadcast TV cameras. Probably Grass Valley, NHK, or one of those companies can comment better.

    Recognition is good for our entire technologist family worldwide.

  3. I think CMOS absolutely has made a material difference to HD TV. The low cost and high quality of CMOS sensors has moved HD video from the hands of professional movie studios into the rest of the world. I don't mean the random public, although they're among those that benefited. Rather, I mean video professionals like news rooms, journalists, commercial makers, and documentary producers. CMOS sensors has enabled an explosion of high quality HD video content from the rest of the video ecosystem. That is what makes Eric's (and my) HD-preferred viewing strategy possible. Even new media models, e.g. YouTube, have benefited. Where would activist journalism and coverage of global events be if we were still shooting NTSC/PAL?

    As for whom to recognize, that's a tougher question. How do you recognize the multitude of incremental developments by an entire industry? Each sensor vendor did their own development of HD-capable devices, as did the image processor vendors. Each camera vendor developed their own cameras. I don't think there's a single developer or product one can point to that represents the entire community. The effect is the combined effort of everyone. Which snowflake was responsible for the avalanche?

  4. My unscientific first cut involved a Google search on "professional HD camera".

    That and similar searches suggest, Sony, Panasonic, Ikegami, JVC, and Canon.

    1. Ikegami does not make sensors. Neither does JVC.

      Regarding the other contenders, one needs to answer a question "whether the improvements to HD CMOS sensors for use in broadcast video cameras "materially have affected television."

    2. Ikegami uses CMOS sensors from AltaSens.

    3. Actually, most of Ikegami broadcast cameras are based on CCDs. Just few of them use Altasens CMOS sensors.

  5. How about Sony back-illuminated CMOS image sensor technology as implemented in products like this:

    1. I've tried this. The response was that BSI has been around for many years in CCDs. Then, how the CMOS BSI has "materially affected television"?

  6. The question is - what differences between CCD sensors and CMOS sensors have affected television for the better? I can point out some diferences bt I have to leave the effects on broadcast TV to others.

    Differences I can identify (not all better for CMOS):

    1 - CMOS sensors use less power
    2 - CMOS sensors accommodate on-chip circuitry
    3 - ROI scanning is easier with CMOS sensors
    4 - CCDs have long had global shuttering - rolling shutters are a step back - even global CMOS shutters are problematic because they generally have less efficiency than the FIT sensors common in broadcast
    5 - On-chip binning is easier in CMOS than in CCDs but it is usually not charge binning so the noise effects are relatively worse
    6 - CMOS dark current can be much less than CCD dark current at the same temperature.

    That is all that immediately strikes me.

    One possible avenue to an answer is "What equipment that is now common was not possible to build without CMOS imagers?"

    1. The industry-changing product was the Canon 5D Mark II DSLR that incorporated 1080p recording (interestingly only added as an after-thought made possible by the live view function) from a full-frame 35mm (36mm x 24mm) sensor.

      This camera has been widely on TV shows, and famously was used to shoot the entire finale of the 2010 season of "House".

      It also put significant price pressure on broadcast and ENG cameras, as it's hard to justify a >$20k camera using 3 x 1/3-inch sensors, or >$80k for one using 2/3-inch sensors, when for $2.5k customers can buy a full-frame 35mm camera (giving the director much greater creative control over depth of field).

      Had the DSLR industry stayed with CCDs, this video convergence would never have happened. The power requirements would not have been compatible with the battery in a DSLR.

      Ironically, the Canon sensor does not even include many of the CMOS innovations used elsewhere in our field. It uses analog outputs (which are fed in to the ADCs on Canon's Digic ASIC)! So many of the improvement that might come to mind aren't even used in product!

    2. Albert TheuwissenJune 2, 2012 at 7:13 PM

      To Dave Gilblom, nice list of characteristics you mention. I do agree with 1), 2), 3). But in 4) you refer to FIT imagers. I think that most broadcast cameras these days use FT or IL imagers. FITs are not that much used anymore, they are too specific and too expensive. 5) Binning in the charge domain is pretty easy for CCDs when it comes down to monochrome devices, and these are the ones that are used in broadcast cameras. 6) I do not agree with this item, CCDs still hold the world record in dark current, although I am happy to admit that CMOS is a good follower.

    3. CMOS image sensors these days have lower read noise at HDTV pixel rates. I wonder if generally they have higher dynamic range for the same pixel size but AT will correct me immediately if I am wrong!
      I also wonder about PRNU these days, at a given pixel size. Anyone know?

  7. The only sensor that would really deserve this marketing-gimmick-prize would be the Panasonic GH2, which beats the CaNikon-Fullframes all along in terms of resolution (check out, if you doubt it). Apart from that, the nearest deserving candidate would probably be either the coming up Panasonic GH3, the Nokia 808 with it's 41-MP-superflexible-scaling-sensor, or the new Black Magic 2K-RAW-Cam, pre-selling@3k$.

  8. Just to let you know, the Emmy committee members are reading your comments.

  9. All I've noticed is more and more reality tv garbage, technology can't improve programing.

  10. Agreed.

    I wish there was a knob on the TV to turn up the intelligence. There's a knob called "brightness", but it doesn't work. -- Gallagher

  11. In Europe we also have a knob called "ON/OFF" ....


All comments are moderated to avoid spam and personal attacks.