Canon's 50MP CMOS Sensor = 19 x 28mm Prototype

Discussion in 'Canon EOS' started by eos 10 fan, Jun 4, 2007.

  1. "Canon has built a 50 megapixel CMOS monstrosity, which is reportedly almost
    twice the resolution of its nearest competition, and is prepping it as a sort of
    large format surveillance camera for monitoring large, busy areas such as
    parking lots and theme parks, along with detailed work like factory part
    inspections. Despite the sensor's clear industrial-end aims, Canon has managed
    to build its prototype at 19 x 28mm in size, the same dimensions of the sensors
    in its DSLR cameras..."

    http://www.engadget.com/2007/06/04/canon-crams-50-megapixels-into-a-cmos-prototype/

    --
     
  2. Canon: Please,

    1. Make a 50D camera with this sensor.

    2. Make at least one lens that will be able to resolve that resolution.

    3. Make a 64 GB or larger memory card to store that amount of data

    4. Make an optical serial new communication link technology to download to computer, at comparable rate to current fastest technology.

    5. Make an affordable mainframe super computer to process and store that data in my home or office.

    6. Make a higher than Blue Ray capacity recordable media to back up photos.

    7. Then finaly, make our salaries so high, so we could possibly afford to get the items needed, as explained in 1 - through 6.

    Thanks for pushing the boundaries!
     
  3. Well, should I wait until that comes out, or should I get the 1Ds Mark what-ever? Maybe I should check my bank account first.
     
  4. If this hypothetical surveillance camera is video, that's a staggering amount of data that would need to be stored. The only thing that makes sense to me without knowing more about it is some kind of live, realtime remote-control system, or a time-lapse still system.
     
  5. Wish #5 is the key.
     
  6. Since 12MP sensors have been made (and used in cameras) at the 5x7mm size, cramming 50MP into a large sensor really isn't a big deal. It's just technology (getting the yield up). If they just scaled up the digicam sensor lithography they could get around 200MP on a chip that size with image quality and noise no worse than a current digicam. Yield might be low, but I'll bet they could make a few as a demonstration.

    The question is why?

    People are constantly moaning that their current lenses aren't good enough for their current 12MP full frame cameras. There's really very little point in putting a 50MP full frame sensor in a camera, nevermind a 50MP reduced size sensor.

    I can see some possible industrial or scientific applications, but I really don't think it would offer much to photographers.
     
  7. I agree with Bob Atkins. There is such a thing as a benefits to expenditure curve. It was axiomatic in my last business that the last few per cent in increased benefits usually doubled the cost of the project. As an old nemesis used to tell me the best is the enemy of the good. We are, I believe, at that point on the usable megapixel curve.
     
  8. consider that a Fuji F30 sensor scaled up to full frame would be around 120 megapixels. In good light this would be pretty interesting with the highest-resolving lenses. in poor light, shallow DOF, high ISO, slow shutter, handheld, cheap lenses (i.e. >$3000) - big waste of disk space.
     
  9. Really, Bob? If we have hit the ceiling with APS-C, isn't this a good argument to push full frame? I sure could use more than 10MP; 50MP would be great in my book. So what do you think is the most 35mm lenses are capable of resolving?
     
  10. phew... internet discussion saved for a month or two ,-)
     
  11. People are constantly moaning that their current lenses aren't good enough for their current 12MP full frame cameras.
    I've read complaints about light fall-off, but this is common even with a film camera. What other complaints are there?
     
  12. I'd say that much above 24MP for a full frame sensor you'll start hitting a serious wall of diminishing returns.

    There's no end of complaining about edge and corner quality, especially with wideangle lenses.

    Half the people who buy a lens seem to think there's something wrong with it. It's either not that much better then the kit lens, or it won't focus properly or it's not sharp at the edges etc. etc.

    The 16MP 1Ds MkII gives absolutely fantastic quality images. Once we get much past 20MP, it will simply be specsmanship that drives things. Who can make the camera with the most pixels. It will be marketing driven.
     
  13. >I've read complaints about light fall-off, but this is common even with a film camera. What other complaints are there?

    How many lppmm can the finest Canon L lens resolve? Anyone know? Has any of them reach 200 lppmm or beyond yet?
     
  14. Oh, yes, yes, YESSSSSSSS ! ! ! ! ! ! ! ! !

    I want it! I love it! Bigger is always better! Give me 50MP, 100MP, 1,000MP. I can handle that...

    I'm dying, I'm fainting, just the thought of that makes me knees feeble and my mind blurring, I'm hot, I need to lie down and cool myself....



    :)



    Happy shooting,
    Yakim.
     
  15. I come from the LF (4 x 5) and MF crowd. I'm very quality conscious, there is no way that any 35mm print (from film) over 5 x 7 satisfied me, even after trying Contax 35. MF color prints from 6 x 6 (Hassy) were borderline for me so I went to 6 x 7 and 4 x 5.

    I was hoping the 1DsII would give me acceptable quality for quick turn commercial work. I bought it almost 2 years ago half expecting to sell it in a month if, after thoroughly testing it and having prints made, I was not happy. When I looked at used pricing I realized I could keep it this long and probably lose less than a rental fee.

    I reshot setups (with 1DsII) that I had shot on LF and MF and had 32 inch prints made. I was shocked at the quality. In one case I was getting picky and thought that maybe the same poster from a 4 x 5 shot had better saturation. I pulled the properly stored poster out and in fact I preferred the 1DsII version.

    Based on nearly 2 years of use and many lightjet prints up to 32 inches I seriously have to wonder why I would need more pixels. Will I buy the 1DsIII when it comes out? Most likely, it will pay for itself anyway, so why not? But if it never comes out I don't really care. We have incredible imaging power available to us at well under 10K, most folk's images are well below the technical quality level available to us now.
     
  16. The lens resolution issue is a bit of a red-herring, because image resolution is a combination of both lens and sensor resolution. And since a 16Mp full-frame sensor can "only" resolve about 67 lppmm, and a good lens can resolve well over 100 lppmm in the centre of the field, we get a combined image resolution of around 45 to 50 lppmm max. Which is well below what a good prime lens is truly capable of.

    So, increasing sensor resolution is not a completely pointless task - far from it.

    IMHO Canon are to be praised for having the technical acumen to realise that large sized and high-resolution sensors are the way forward, and that the four-thirds and smaller route is a dead-end that's already been pushed about as far as it can go.

    As far as storage requirements are concerned; here's a little anecdote. I've just bought an 8 Gigabyte 120x CF card. It cost me about the same as I'd paid for a 256 Megabyte 16x card not-so-long-ago. My USB2.0 card reader can read that 8Gb card in less time than my older USB1.1 reader took to read the 256Mb card. Plus the hard disk in my computer now holds 300Gb, rather than the 40Gb that was common a couple of years ago. So, if this sensor ever makes it into a still camera, where's the problem going to be in handling large image files?
     
  17. So, if this sensor ever makes it into a still camera, where's the problem going to be in handling large image files?
    Agreed - especially since there exist 16 GB Compact Flash cards, and 1 TB Hard Disk Drives, now.
     
  18. William Michael

    William Michael Moderator Staff Member

    >>> Oh, yes, yes, YESSSSSSSS ! [ . . . ] I'm hot, I need to lie down and cool myself.... <<<

    Apparently the boys in the TRD room are working on an EF 16 to 300mmF1.2L also, it is razor sharp and it weighs only 500gms.

    Better have another lie down.
     
  19. Yes, I think I will... :) But what is TRD?

    Happy shooting, Yakim.
     
  20. William Michael

    William Michael Moderator Staff Member

    TRD is yet another TLA (three letter acronym) from: Technological Research and Development, or Technology, Research and Development. TRD or TR&D is a forerunner to R&D (Research and Development) which is more common nowadays. But I am told I am a tired old fart because I still use the former.

    There is another TLA of interest here:

    http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=00LGL5

    And perhaps you could consider HSY

    Cheers

    Bill
     
  21. It never cease to amaze me what great new stuff you learn in photography forums other than photography... :)


    Happy shooting,
    Yakim.
     
  22. "How many lppmm can the finest Canon L lens resolve? Anyone know? Has any of them reach 200 lppmm or beyond yet?"

    You can probably get 400 lp/mm in the center of the aerial imaage of the best lenses at f4.

    However if you think you have trouble focusing now, try accurately focusing on a 400 lp/mm image with a sensor capable of resolving that. You're going to need auto focus bracketing to hit focus, and you'll only get the high resolution over a very small depth of field.

    If you think user complaints are bad now, just wait!
     

Share This Page