How sensor physical sizes are measured? CX 1", 4/3 1.33"

Discussion in 'Open Gear Talk' started by Naveed Akhtar, Jan 14, 2012.

  1. Naveed Akhtar

    Naveed Akhtar SC Regular

    186
    Dec 8, 2011
    London, UK
    :confused:
    Let me first explain, why I reached there. I noticed that Canon G1X has a smaller sensor than normal APS-C but more pixel count than many 4/3 sensors, so maybe the pixel density is higher, so according to the law of physics that says bigger pixel size means better quality, it may end up a noisier sensor.

    Now back to the question, how we calculating it's 1.5" and why is Nikon CX is 1"?

    Lets take the example of CX format as its easier to see 1" as a whole number.

    The physical dimension of CX sensor is: 13.2 x 8.8
    The diagonal length would be: 15.86
    From Maths: Square root (Square of Height + Square of Width)
    Now 15.86mm is 0.62in
    as 25.4mm is 1"

    So if this the sensor diagonal length is not 1" then what is this 1" in Nikon 1?

    Similarly 4/3 means 1.33" while the diagonal length is 0.85" accordingly
    Similarly Canon G1X got 1.5" sensor and the diagonal length 0.92"

    According to the 4/3 standard definition, I found it is the diameter of the 4/3 tube, so is it the image circle? well in that case to get the 1.33" image circle, dont we need 1.33" diagonal length on the sensor?

    I know you guys are Photographers, but I also know you are the only people I got to ask such question, with huge photography experience and knowledge about history and mechanics of numerous camera systems and most importantly a passion to learn more and share/discuss the ideas.

    Thanks for reading!
     
  2. Luckypenguin

    Luckypenguin SC Hall of Famer

    Dec 24, 2010
    Brisbane, Australia
    Real Name:
    Nic
  3. Naveed Akhtar

    Naveed Akhtar SC Regular

    186
    Dec 8, 2011
    London, UK
    Thanks

    Thanks Nic, I been to that page earlier, but couldn't reach upto the bottom. Thanks for your achored link, it was spot on.

    That is why I was wondering why my calculation is consistently with the proportion to bigger format sensors. i.e. 1" vs 1.33" vs 1.5" and calculated 0.62 vs 0.85 vs 0.92. So, it is an age old tradition to report a value which is actually one third less of the diagonal of sensor size :wink:
     
  4. bartjeej

    bartjeej SC Hall of Famer

    Nov 12, 2010
    Real Name:
    bart
    yeah, to me it always made more sense to just describe the sensor size in length and width, or surface area...
     
    • Like Like x 1
  5. olli

    olli Super Moderator Emeritus

    Sep 28, 2010
    Metro Manila
    Real Name:
    olli
    Totally agree with this. The current multiple ways of expressing sensor size are absurd. Length X Width (in mm) and surface area is all you need to make a comparison.
     
  6. Brian

    Brian SC Top Veteran S.C. Charter Member

    638
    Jul 7, 2010
    Schneider Optics

    So, if you were going to replace your sensor with a Vidicon tube- it would be that size.
     
    • Like Like x 1
  7. wt21

    wt21 SC Hall of Famer

    Aug 15, 2010
    So if the Sony is 368mm2 in area, and m43 is 225mm2 in area, then the APS-C is 63% larger? I had always heard something more like 25% larger.

    If FF is 860mm2, then it's 137% larger than APS-C and 282% larger than m43?
     
  8. bartjeej

    bartjeej SC Hall of Famer

    Nov 12, 2010
    Real Name:
    bart
    yup, pretty much...
     
  9. wt21

    wt21 SC Hall of Famer

    Aug 15, 2010
    Goes to reinforce my "I would really like a mirrorless FF camera" even more!
     
  10. Naveed Akhtar

    Naveed Akhtar SC Regular

    186
    Dec 8, 2011
    London, UK
    Yes that's why normally on review sites they wrote the actual sensor size w x h in mm. It was just bit confusing that mostly manufacturers are describing sensors as format type, specially smaller than fullframe. Format type is not actually the diagonal of the sensor but the diameter of the tube (something in old and ancient photography) not very useful these days.

    Still its easier to remember sensors with those ancient values, in my little opinion. It is easier to remember 1" then 0.62" of the actual length. Remembering it in L x W is even more difficult 13.2 x 8.8 mm, sounds like measurement of a box and is two values, or the area 116.16 saure mm sounds like plot area size for your new property. For technical references its all fine and we can go and deal with all these numbers and go even further and find individual pitch of single pixel. Otherwise, Its just a convineance of remembering the number.

    Plus it is easier to refer new standards with previous one, just like we call full-frame to 35mm film, which again is not exactly 35mm. Fourthird; even if its not that valid now a days, becuase we are not dealing with those TV tubes as much, still it gives a good idea that ok its same size as Fourthird standard use to be and so and so. Just to relate!

    In essence, all we want is to compare and see how much one sensor is bigger than the other and for that purpose if they all are multiplying it with one constant value, I think its fair. It could be worse if they all have their own multipliers to confuse us even more. Now all am taking from these values is if CX is 1 unit Canon new APSC is 1.5, that gives me very good idea that 4/3 sits between at 1.33 unit. So I can visualise and relate it to my PC monitor and say, If my previous monitor screen was 10" this new one will be 15" and so on. Just an easy way to relate different sizes.

    There are still some difficulties I find when I was measuring the pixel density. The reported W x H is also inconsistent, Canon and mostly all other major manufacturers Pany, Oly and others report its sensor sizes area with effective exposure to image circle which is effective sensor area, where as Nikon (I don't know why) like to report its sensor's actual dimention (Check D3100, D5100 and D7000 all have different WxH). Now getting the actual pixel density is possible in one system and effetive density in rest of them, but then you can't accurately compare them with each other. Probably thats why DPR has stopped showing pixel density in their camera specs sheet.
     
  11. Brian

    Brian SC Top Veteran S.C. Charter Member

    638
    Jul 7, 2010
    The practice of using the diameter of the Tube to describe the sensor size is useful for defining the image circle of the lens. C-Mount lenses are built to different specs: 1", 2/3", 1/2"- same physical mount, different image circle. When looking for legacy C-mount lenses, best to get one with a 1" coverage.
     
    • Like Like x 1
  12. Naveed Akhtar

    Naveed Akhtar SC Regular

    186
    Dec 8, 2011
    London, UK
    Well Sony sensors just like Nikon have different dimensions on different models. Their effective pixels are not directly reported unless you want to go individual camera and calculate the ration between actual and effective sensor and then apply it to the surface area. But more or less you are right .. Its 225/368 = 61% less the margins it got 64.3% roughly 35.7% less area than Nikon standard APSC
    and 225/332.7 = 67.6% that is roughly 32.4% less area than Canon DSLR APSC
    and 225/261.8 = 85.9% that is roughly 14.1% less area than Canon compact APSC G1-X
    and roughly double of Nikon CX

    If they (Canikon) shrink the size of their smaller FF DSLRs, improve the C-AF atleast upto the origional and slow Pen E-P1, without compromising metering and Focus accuracy and then add a tilting LCD like Sony A580.. then I won't bothor about the mirror that much. Mirrorless FF but without any lenses like NEX system and many others doesnt appeal to me.
     
  13. Naveed Akhtar

    Naveed Akhtar SC Regular

    186
    Dec 8, 2011
    London, UK
    Thanks Brian, was it image circle or the diameter of the tube. As if its the dia of the image circle, why do they need to make a rectangular sensor with diagonal one third smaller than the circle? or is it a practice of the past and not valid anymore. I am thinking in terms of GH1 and GH2 and their multiaspect sensors and their precise coverage in multiple aspects to the image circle!
     
  14. Luckypenguin

    Luckypenguin SC Hall of Famer

    Dec 24, 2010
    Brisbane, Australia
    Real Name:
    Nic
    I guess with any lens there is a difference between the absolute image circle and the usable image circle. I think I'd prefer to see manufacturers use the actual diagonal measurement and/or area to describe their sensors, but tradition is as tradition does.
     
    • Like Like x 1
  15. Naveed Akhtar

    Naveed Akhtar SC Regular

    186
    Dec 8, 2011
    London, UK
    Even more confusing is Fulframe that not only say the actually film (sensor) measurement 35mm but actually tells it lower than actual (opposite of APSC and smaller says). As 35mm is the width, diagonal is roughly 43.3mm! Anyways enough for all these measurements for a day! =]] :biggrin:
     
    • Like Like x 1
  16. Luckypenguin

    Luckypenguin SC Hall of Famer

    Dec 24, 2010
    Brisbane, Australia
    Real Name:
    Nic
    35mm of course being the height measurement to each perforated edge of the film as opposed to an actual measurement of the exposed frame, but as you say, enough measurements for one day. Shoot more, theorize less! :wink:
     
    • Like Like x 1
  17. Brian

    Brian SC Top Veteran S.C. Charter Member

    638
    Jul 7, 2010
    The measurement for the C-Mount Vidicon tube looks like the diameter of the tube. I have two old Vidicon cameras, one C-Mount and the other a Bayonet mount. They are both over 30 years old.
     
  18. Naveed Akhtar

    Naveed Akhtar SC Regular

    186
    Dec 8, 2011
    London, UK
    That what I thought how it should be .. Thanks very much for confirmation Brian!