1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Full frame sensors

Discussion in 'Weekly Poll' started by Damien_Demolder, Jul 16, 2007.


Iwould like to be suspended by my

  1. Thumbs

    0 vote(s)
  2. Nipples

    0 vote(s)
  3. Love nuts

    0 vote(s)
  4. Neck

    0 vote(s)
  5. Thong

    0 vote(s)
  1. Damien_Demolder

    Damien_Demolder Well-Known Member

    The latest poll is about full frame sensors. Do you think they produce better quality images than APS-C and 4/3 sized sensors? Have your say on the Home Page

    The results of the last poll are in, and it seems you are an independent lot who like to think for yourselves. 52% of you have subject based exposure modes and choose not to use them, while 22% have cameras that don't feature these modes at all. 8% of those that responded use these modes 'often' to help with their exposure and/or processing settings and a further 18% use them occasionally. That's a 74% don't use, 26% do use split. I guess that reflects the camera expertise level of the AP readership in general.

    You can follow the scene mode discussions here

    Thanks again for taking part.

  2. photogeek

    photogeek Well-Known Member

    Not sure about this question - I think it very much depends on the camera the sensor is fitted in, certainly Leica do not think a full frame sensor would have worked in the M8, but then the camera was not designed from scratch to be digital and is a compromise so that M-Mount glass can be used.

    The early Kodak cameras were full frame and supported the nikon F-mount, but would I say the quality is better than a D200 or D2x - definately not.

    Are the Canon Full Frame sensors better than there APS-C / H cameras - probably but on then they tend to be generrally more expensive cameras.

    In theory the bigger the sensor the better the image, provided the lens attached does not vignette and to my mind this means that the physical size of the sensor is only one component of the equation.
  3. parisian

    parisian Well-Known Member

    I suspect that the jury will be out on this one until there are manufacturers other than Canon to compare with. The frequent comments on the disappointing quality of images via the Canon sensor with some of their 'L' glass indicates that simply replacing the film with a like sized sensor is perhaps not such an easy transition.
    The potential cost of re-jigging a whole raft of lenses will be prohibitive and I think that in the race to be first and therefore capture more of the professional and high end amateur market Canon have (very unusually for them) got it slightly wrong.
  4. Lounge Lizard

    Lounge Lizard Well-Known Member

    I'd suggest 'tolerably wrong' as a better phrase. ;)
  5. Burgy

    Burgy In the Stop Bath

    Which ones were they, as far as I can recall the Kodak/Nikon Digital Bodies were the same as the Kodak/Canon Bodies for example the the DCS 3 (1.5 crop) DCS 520/620 (1.5 crop)

    And the earliest Kodak DSLR I can find the DCS-100 which was based on a Nikon F3 body with a motor drive had an imaging chip 1280x1024 pixels and 14 x 9.3 mm sensor

    In fact the first full frame dig SLR is not until 2002 with the Kodak DCS Pro 14n which is a long way later than the 1991 DCS-100
  6. Benchista

    Benchista Which Tyler

    And in any case they have released an updated version of the 16-35mm for this very purpose.

    But the only reason there is any debate on this part of the issue is that the quality is in general so very good. Great resolution, and superb high ISO performance make a very powerful case.
  7. Malcolm_Stewart

    Malcolm_Stewart Well-Known Member

    ...plus, in the case of the 5D a lower pixel density allowing better use of the higher contrast portion of the lens MTF curve. Whether raising the pixel count (in a future model?) will negate these advantages we'll have to wait and see. I'm not including the 1Ds in this comparison as it's aimed at a different market, and may have (?,probably has) different in-built processing.
  8. huwevans

    huwevans Not Really Here

    That's an interesting point of view, Peter - I'm not sure how many Canon 5D and 1DS owners are going to agree with it, though! ;-)

    FWIW, when I tried out the Kodak Pro14n (the first version of that camera) some years ago, it was with a perfectly ordinary Nikon zoom, and there was no evidence of any image quality problems that could be attributed to the lack of a specially redesigned optics. Mind you, the Kodak did not have microlenses on its sensor, which meant no particular reason for problems in the corners, although it did mean less light gathering power, which could only have added to the noise problems.

    On a slightly different tack, and again FWIW, I was recently examining the test images from a multiple DSLR and MF digital back studio shoot. All sorts of high-end kit were involved, from the Nikon D2X and all the Canon full frame cameras, right up to a stack of 22MP backs (nothing higher than that). Of course, given the various different bodies involved the lenses used were similarly varied, and so conclusions have to be guarded. But one DSLR stood out from the others, producing an image which in fact bore reasonably favourable comparison even with the 22MP backs, and it wasn't any of the ones with the full frame sensors! :) Of course, had the test shots been done at ISO1600, say, instead of the base sensitivity, I doubt that that result would still hold.

    Howbeit, certainly you can get outstanding images from an APS-C sensor, just as you can from 'full frame'. What you inevitably can't get is quite the same level of shallow DoF, in like for like situations, or indeed the same degree of choice in wide lenses as is available to the full frame shooters. Likewise, the latter have a harder job getting wide DoF and less scope at the telephoto end. So, you pays your money and you takes your choice. I'm about as happy as I could be with my APS-C kit. :)
  9. Footloose

    Footloose Well-Known Member

    The first full-frame DSLR was the Contax, I don't know if any of the Kodak ones were actually full frame or not. Although 'potentially' the larger sensor area should produce better images, but, [as Canon may have discovered] will a new breed of lenses be needed to overcome the limitations of existing optics?

    Looking at the pixel counts of the various formats, it does not seem to follow that because a sensor is physically bigger, it will be populated with as many pixels per sq mm as very small sensors. Look at the pixel count on sub APS cameras, and then scale these up to say, those on a MF sensor. There certainly seems to be the enthusiasm and technologies being developed (because of the much vaster number of units sold) to increase the count on very small sensors, so is there some optical-mechanical barrier which is preventing similar increases in pixel-power, that the fabrication plants have yet to overcome?

    Maybe the problem is that the optical designs required for a 35mm sized sensor, results in leneses which are of much greater diameter and as a consequence, have a negative impact on camera ergonomics?

    Much of what I am stating is conjecture, but may well have a bearing on how successful a full frame camera may well turn out to be.
  10. JDCB

    JDCB Well-Known Member

    I'm not sure how many elements go together to make an overall image: sensor size seems to me to be only one of a number. The lens through which the light passes and is changed, sensor performance unprocessed, pixel size and density, whether pixels have microlenses or not, the amount of UV and IR light present, the quality of the in-camera processing performed, the quality of the off-camera processing by differing RAW converters, the calibration of the displaying monitor, or the ability of the printer, the size at which the image is output, and so on surely all have a part to play.

    Stripping away or standardising every variable apart from the size of the sensor, then a bigger sensor should in theory produce a better image than a smaller sensor, but real life is not like that.

    I suspect that the whole argument is rather sterile considering that truly great images very often contain technical, visual and physical defects caused by the inadequacies of the film, sensor, lens, camera, or printer used.

  11. parisian

    parisian Well-Known Member

    In an addition to my earlier comment;
    One reviewer (BJP to my memory) argues that; 'the current Canon lenses are not, as yet, meeting the abilities of the sensor.'
    Nick tells us that one of the wide angle zooms has been re-thought already. My original argument stands - slightly wrong it is /forums/images/graemlins/tongue.gif

    Slightly is the key word here. I am not suggesting devastatingly or 'don't touch it with a bargepole' wrong; just a little bit which simply surprised me for Canon.
  12. Footloose

    Footloose Well-Known Member

    If AP tested all sensors using the same optic (On a bellows assembly using a MF Prime - as I am guessing this will give additional distance between the rear element and the sensor.) of irrefutable quality, in a test-rig, so that (Insert manufacturer's name here!) users cannot argue that the lens used was duff, there will always be squabbles about how good the sensors used, actually are!
  13. Barney

    Barney Well-Known Member

    Go on, name it!!!
  14. parisian

    parisian Well-Known Member

    XD2 re-arrange to your liking :D
  15. huwevans

    huwevans Not Really Here

    Yes - that sounds like it! :)

    But in all truth I wasn't trying to plug any particular camera, or at least not on that basis. There are so many variables, not least of which is the lens. And whilst the lens used on the Canons has a fine reputation, I did find myself wondering if might have been a dodgy sample (I'm just assuming it was the same one each time), because nothing in any of the three Canon shots looked quite as crisp as we've come to expect.
  16. tessar

    tessar Well-Known Member

    As a retired lens maker, who knows little of electronics, may I make a plea in this debate.Now that I have a Fuji pro with a 1.5 sensor I need a good prime short focus well corrected lens. A 14mm f6.3 retrofocus design is possble to make that is small,light,and cheap.However Nikon choose to offer only f2.8 an almost impossible design,weighs a ton and costs a fortune.The depth of field with a 14mm lens is so great that for most of us f6.3 would be perfectly adequate.
    My M/F 6/7 outfit uses Linhoff Schneiders all are f5.6 or f8 max apertures and I have no problems.
    I accept that some fast lenses are needed, but dear Nikon can we have some simple ones as well, that I can afford!
  17. Barney

    Barney Well-Known Member

    Hmmm, well there's no funny symbols in there so it can't be a Sony or a Pentax. :D
  18. Almark

    Almark Well-Known Member

    I nearly went for the Canon 5D, the full format appeaeled to me, as someone intersted in landscapes. Instead I went for Nikon. Anyways I have spoken to a few 5d owners who spent so much on the camera that they bought 'kit' lenses. They found problems.

    From what i have been told by such users is that to get the benefit from the 5d you need to invest in 'L' lenses, which are expensive. They learnt from their mistakes, sold off their 'kit' lenses and saved up for 'L' ones.

    If Nikon brought out a full frame I would consider it. In the meantime I will wait for the DX3
  19. Clive

    Clive Well-Known Member

    I don't think it is a surprise that the D2X gave the best image quality at baseline ISO. Fitted with a good lens designed for 35mm size (like the 17-35/2.8) the results across the whole image area, corner to corner, are staggeringly good. But, of course, the corners are not there. They are in the image that falls beyond the limits of the APS sized sensor.

    If I put a Zeiss lens on my Hasselblad 501 I can see the degradation in the corners. Likewise with a summicron 35 on my Leica M6. The very best lenses on film cameras exhibit this degradation. Taking this further, there is vignetting in those corners, something which is often exacerbated by digital sensors.

    Put a lens designed for an APS sensor on the D2X and the effect is similar to the problems with a full frame sensor. Perhaps the best results would be for lenses to be designed for a sensor which was 1.5x the size of the full frame sensor.

    But noise at higher ISOs is a different story. The best digital SLR I have had for coping with the noise problem was the 6m pixel D100 (on an APS sized sensor)
  20. Malcolm_Stewart

    Malcolm_Stewart Well-Known Member

    I bought a 5D early on, and use it mainly with non-L prime lenses. (24 f2.8, 35 f2, 50 f1.4, f1.8 and f2.5 macro, 85 f1.8) I do have a number of L teles bought over many years, and I now have an L zoom (70-200 f4L). When travelling, I have used an EF 80-200 f4.5-5.6 consumer zoom. With my style of photography (whatever that is!) friends find it difficult to tell which class of lens took which photograph. I guess matters might be different if I took a lot of amateur sports in less than very bright sports halls, where my EF135 f2L might not be good enough!

    Consequently, I'd argue that for much 5D photography, L lenses are very nice, but certainly not essential. A 5D with the EF35 f2 makes a reasonably compact "street camera", which no L lens would allow.

Share This Page