what happened with Canons APS-H sensor with 120megapixel?

[gkaefer]

source: http://cpn.canon-europe.com/content/news/120_megapixel_apsh_format_cmos_sensor.do

with respect to the current megapixel rumored EOS bodies...
is it likely to get such working experiment from 2010 finally in a production-ready EOS?

Georg

 

[Chris_prophotographic]

they sit on things until a market can support mass production.

 

[Mt Spokane Photography]

they sit on things until a market can support mass production.

Exactly right. No different than a prototype of a new car. Handmade test products will mostly never be in production, but you learn from pushing the limits even if its only a custom 1 off item.
Making a 120mp sensor is easy when compared to making a camera that can use them, or processing the images in a home computer.

 

[bbasiaga]

Maybe it was so depressed over its lack of DR compared to Nikon that it OD'd on sleeping pills. LOL

Seriously though....there are a number of things that make 120mp cameras impractical for home use. Start doing some layer editing on the 22mp images from a 5dIII and you can get 500mb working files no problem. Now multiply that times 6. Ouch. Now your terabyte hard drive only holds a few hundred files. Extreme example, I know, but the point is made. It would take a lot of processing power and ram in both the camera and PC o make that manageable.

=Brian

 

[TAF]

Another possibility is that the Japanese were sending a "message"; I suspect the real use is this (certainly for the large full wafer sensor from the same article):

http://www.spaceflightnow.com/h2a/f19/

 

[wickidwombat]

all the h8t0rs made it go away because they said APS-H wasnt as good as APS-C

 

[jrista]

source: http://cpn.canon-europe.com/content/news/120_megapixel_apsh_format_cmos_sensor.do

with respect to the current megapixel rumored EOS bodies...
is it likely to get such working experiment from 2010 finally in a production-ready EOS?

Georg

It was just a PR stunt to show that Canon can make a 120MP sensor, there are other manufacturers with 65nm equipment tech (Canon has 180nm) and If Omnivision, Aptina, Sony and others will showcase a large 24x36 mm sensor with a single pixel area of 1.1 micron they can do it. And NOW we are talking resolution but the manufacturing costs would be very high for such sensor

If you are talking about R&D costs...well, Canon's already done that, and proven the research by actually manufacturing prototypes. Amortizing that cost over a lifetime of 120mp APS-H DSLR sales wouldn't be any different than amortizing the cost of the 1D X sensor.

Why would manufacturing costs be particularly high? The cost is dependent yeild, which boils down to sensors per wafer. With APS-H, you get more sensors per waver than FF, so cost for a commercial 120mp APS-H sensor should fall somewhere between FF and APS-C costs today. Since FF costs seem to have come down considerably, a 120mp APS-H DSLR should be fairly reasonable in the grand scheme of things.

As for pixel size, a 120mp APS-H sensor would have around 2.1 micron pixels, about half that of Canon's current 18mp APS-C sensor. From what I gathered from the original press release, the 120mp prototype had a very high readout rate of 9.5fps thanks to some kind of onboard parallel readout and pixel processing (which at the very least indicates some kind of on-die hyperparallel ADC). One way or another, Canon certainly seems to have the ability to develop a complex image sensor with a lot of on-die logic, much like Sony with their CP-ADC. Whether that includes additional noise reduction circuitry per pixel I can't say, but 2.1 microns is 2100 nanometers, or almost 12 times the size of a 180nm transistor. Throw in gapless microlenses, and I highly doubt Canon's current fabrication capabilities would make it difficult for them to manufacture a commercial 120mp sensor.

 

[tnargs]

Maybe it was so depressed over its lack of DR compared to Nikon that it OD'd on sleeping pills. LOL....

Make that 'compared to Sony'.

I wonder where Nikon's latest announcement is over THEIR latest sensor development? ;D Oh, here it is, an article on nikon.com on the development of their D3 sensor! And guess what, they make a big fuss over the superior IQ from developing the sensor in-house: "...we took advantage of the fact that the sensor was being developed in-house to investigate how best to match the CMOS sensor with the capabilities of the NIKKOR lens, and were thus able to bring out the full capabilities of the photographic lens. In this way we succeeded in obtaining a high level of image quality. Further, with the sensor being fitted to our flagship camera model, we have been able to incorporate functions and technologies that are effective for producing a high level of image quality without any compromise. "

LOLOLOL

 

[jrista]

source: http://cpn.canon-europe.com/content/news/120_megapixel_apsh_format_cmos_sensor.do

with respect to the current megapixel rumored EOS bodies...
is it likely to get such working experiment from 2010 finally in a production-ready EOS?

Georg

It was just a PR stunt to show that Canon can make a 120MP sensor, there are other manufacturers with 65nm equipment tech (Canon has 180nm) and If Omnivision, Aptina, Sony and others will showcase a large 24x36 mm sensor with a single pixel area of 1.1 micron they can do it. And NOW we are talking resolution but the manufacturing costs would be very high for such sensor

If you are talking about R&D costs...well, Canon's already done that, and proven the research by actually manufacturing prototypes. Amortizing that cost over a lifetime of 120mp APS-H DSLR sales wouldn't be any different than amortizing the cost of the 1D X sensor.

Why would manufacturing costs be particularly high? The cost is dependent yeild, which boils down to sensors per wafer. With APS-H, you get more sensors per waver than FF, so cost for a commercial 120mp APS-H sensor should fall somewhere between FF and APS-C costs today. Since FF costs seem to have come down considerably, a 120mp APS-H DSLR should be fairly reasonable in the grand scheme of things.

As for pixel size, a 120mp APS-H sensor would have around 2.1 micron pixels, about half that of Canon's current 18mp APS-C sensor. From what I gathered from the original press release, the 120mp prototype had a very high readout rate of 9.5fps thanks to some kind of onboard parallel readout and pixel processing (which at the very least indicates some kind of on-die hyperparallel ADC). One way or another, Canon certainly seems to have the ability to develop a complex image sensor with a lot of on-die logic, much like Sony with their CP-ADC. Whether that includes additional noise reduction circuitry per pixel I can't say, but 2.1 microns is 2100 nanometers, or almost 12 times the size of a 180nm transistor. Throw in gapless microlenses, and I highly doubt Canon's current fabrication capabilities would make it difficult for them to manufacture a commercial 120mp sensor.

Then I think you are wrong
Canon made a 50Mp sensor 2007 and nothing happend with that one.

About 120Mp I will qoute Albert Theuwissen one of the worlds best sensor experts http://www.harvestimaging.com/index.php?id=1

Albert TheuwissenAugust 24, 2010 1:39 PM

Did the 50 MP sensor ever landed in a product ? As far as I know I did not. So I am wondering whether this 120 MP sensor is intended for any photographic product, I do have my doubts.//end quote

My comments is Time will tell

Whether Canon WILL make a commercial product out of it and whether canon CAN make a commercial product out of it are two very different things. It is quite obvious Canon CAN make these sensors...they DID. Whether they think such a sensor is even viable for commercial use is an entirely different topic, and one not affected by their ability...thats marketing, legal, cost/benefit and all that, processes that can be extremely complex and often completely illogical. Not to mention the fact that all of the original press releases from Canon indicated they believed the sensor was more viable to the scientific and medical communities, there was little mention of the commercial DSLR community.

Either way, Canon does have the technology to produce an extremely high resolution, high readout rate sensor. Whether they put that technology to use in an actual retail 120mp sensor, or whether they use it to progressively improve their general sensor technology, it doesn't change the fact that Canon has already designed and developed the technology. That original 50mp sensor that Canon produced a few years back...well, 46.1mp isn't that far of a cry from 50mp...sometimes it just takes TIME to translate research into an actual commercial product...and usually not because it's "too hard to do." They will use it as and where they see fit.

It doesn't really matter what Albert Theuwissen thinks, or what Eric Fossum things, or what any of the worlds well-respected patent holders think...all just talking heads, same as us.

 

[expo01]

A company like Canon is doing alot of R&D. The magic phrase for prototypes like the 120MP sensor is "proof of concept". You are testing certain aspects. The 120MP sensor has shown that you can do multiple line read outs at a time and produce stunning speeds, which is now famously featured in the 1DX. There you go, a 2010 prototype lead the way for 12 FPS. Another issue they observed with the sensor was, that you are getting diffraction blur at such a wide open aperture that it is currently (2010) not managable to deal with in a finished product.

There is not a shred of doubt that we will someday have 120MP FF or even MFT sensors. When Canon is launching their new 1Dsomething with 120MP you can be assured that it will feature an all newly developed sensor. They won't just take the 2010 prototype and mass produce it.

Another prototype they did is actually now boing used. The ~20cm by ~20cm CMOS is used to map space in an observatory located in Japan. Like most astronomy telescopes and imagers they are one of a kind. Produced to specific uses.

A camera featuring a 20cm by 20cm CMOS and a fixed 43mm (35mm equivalent) lens would be a truely awesome piece of photography gear I would crave for!

Imagine this scenario. You are out at night on the streets with a model. The moon is shining. You found a good spot for some nice fashion portraits. You take your camera out and start doing your "photographer thing", then you say to your assistant "hey dave, go kick out that street light over there, I don't want to use ND filters tonight".

 

[Drizzt321]

A company like Canon is doing alot of R&D. The magic phrase for prototypes like the 120MP sensor is "proof of concept". You are testing certain aspects. The 120MP sensor has shown that you can do multiple line read outs at a time and produce stunning speeds, which is now famously featured in the 1DX. There you go, a 2010 prototype lead the way for 12 FPS. Another issue they observed with the sensor was, that you are getting diffraction blur at such a wide open aperture that it is currently (2010) not managable to deal with in a finished product.

There is not a shred of doubt that we will someday have 120MP FF or even MFT sensors. When Canon is launching their new 1Dsomething with 120MP you can be assured that it will feature an all newly developed sensor. They won't just take the 2010 prototype and mass produce it.

Totally agree, the purpose of making something like that is "Lets see if we can do it, and what we'll learn about lens/cameras at such a high MPx." And guess what. They did learn something! And I'd imagine some of that technology is, as you said, in the 1D-X, and probably more coming in the Studio high MPx that's been rumored. Hopefully with some additional new sensor tech around banding & DR.

Imagine this scenario. You are out at night on the streets with a model. The moon is shining. You found a good spot for some nice fashion portraits. You take your camera out and start doing your "photographer thing", then you say to your assistant "hey dave, go kick out that street light over there, I don't want to use ND filters tonight".

LOL, that'd be pretty amazing. The problem might be your shutter speeds measured in seconds

 

[jrista]

A company like Canon is doing alot of R&D. The magic phrase for prototypes like the 120MP sensor is "proof of concept". You are testing certain aspects. The 120MP sensor has shown that you can do multiple line read outs at a time and produce stunning speeds, which is now famously featured in the 1DX. There you go, a 2010 prototype lead the way for 12 FPS. Another issue they observed with the sensor was, that you are getting diffraction blur at such a wide open aperture that it is currently (2010) not managable to deal with in a finished product.

Even with diffraction at wider apertures, a 120mp sensor is going to resolve far more detail than a lower resolution sensor. Even stopped down to f/8, f/16, etc. the 120mp sensor will still be better...you just get diminishing returns. At NO POINT, EVER, will a 120mp sensor produce WORSE results than a lower resolution sensor. That is a common misunderstanding, but a misunderstanding nevertheless. Diffraction, once it occurs, does not degrade IQ below that of a lower resolution sensor. At worst, you would get the same IQ as a lower resolution sensor. It's an asymptotic relationship.

A 120mp sensor would be entirely viable today, and would probably produce better IQ as well. Not only that, you could ALWAYS scale your images down to reduce noise, and still print large. Or simply keep it at native size and print HUGE. No matter how you slice it, diffraction is not the thing keeping a 120mp sensor off the market.