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. Dom_Rivers

    Dom_Rivers Well-Known Member

    It makes me wonder, with programs like DxOptics how long it will be before that level of processing becomes an automatic function of the camera, ie it recognises the lens and corrects the known optical flaws or even straightens the fisheye effect (easier to make than true ultra wides)allowing lens manufacturers to concentrate on producing a sharp or fast lens without having to worry about distortion or vignetting and allowing the lens to be smaller overall.
    I can see a day when you buy a Sigma 6-12mm wideangle and it comes with a CD full of the correct software updates for all the different bodies it may fit.
  2. Benchista

    Benchista Which Tyler

    Some already do, I believe, to an extent.

    Anyway, my fisheye defished isn't as wide as my rectillinear wideangle - you lose a fair bit of FOV in defishing.
  3. beejaybee

    beejaybee Marvin

    Oh come on now. I don't think APS was ever aimed at serious amateurs. Surely the idea was a replacement for "Instamatic" cartridges with a couple of extra features - limited control of proportion plus a built-in neg storage system. Anyway "professional" films never made it into APS cassettes; the choice of anything other than medium speed colour reversal film in APS was always very, very limited.

    Happy snappers (who would surely have accepted the quality that APS was capable of) were simply put off by the premium price charged for APS D&P. DIY amateurs were put off by the need to get special equipment to process APS film, several orders of magnitude more expensive than a 35mm developing tank, together with the realization that the proportion control was only masking of the film aperture. Professionals had too much investment in lenses to think of changing (the same problem that has afflicted the DSLR market). Now if someone had invented a reel for dev tanks that took APS width film strips, and a device for unloading the film strip from the cassette that could be used to get the film strip to the reel, that would have been a start. Sales of APS SLR bodies that took existing lens mount systems might then have triggered a serious market, despite the focal length magnification factor. Specialist films would have become available if there was a perceived market.

    Despite the marketing blunders, APS was at least successful enough to kill off Instamatic cartridge and Kodak disc film. The APS/35mm format battle amongst compact camera buyers might also have helped to some extent to open the market for compact digital cameras. That battle is lost; feature lists and the magic word "digital" beat quality into the dust so far as mass markets are concerned.

    That gets back to the point of the thread. Full frame sensors should be better than C-size, which should in turn be better than four thirds. But, unless you are prepared to jump up in price range by a factor of five, and accept kit which is relatively large, clunky and doesn't compete in terms of "specification" (raw numbers; don't forget to count in all those ridiculous special modes which top-range kit doesn't have, nor need), or if for some reason you have an aversion to Canon kit, then the full-frame "choice" simply doesn't exist.

    Is a EOS-5D better than an EOS-30D? Well, it certainly should be. Given the limitations of "old" (designed-for-film) lenses, is the price difference worth paying for the rather marginal (in most cases) increment in performance? That's up to you.
  4. JDCB

    JDCB Well-Known Member

    Absolutely - pros and cons. That said, while I have taken a couple of BW landscapes on 800 or faster film, it's not something that I would routinely do, and it certainly wouldn't cause me to break out lotsandlotsandlotsanlots of additional beer tokens to have the possibility. For most people, DX sensor plus 12-24 from whichever brand is by a big distance the more cost-effective route to similar quality results for landscape work. For those so inclined, MF or LF with film is probably the ultimate. A panoramic head can be had for £100 upwards, which give the DX shooter additinional options without spending on a specialist DX lens, but with the need to spend time at the PC. In any case, it is in the area of landscape that I feel that the FF DSLR advantage is absolutely negated.

    Take the same kit to a smoky crowded jazz club and the answer probably changes.

  5. beejaybee

    beejaybee Marvin

    Sorry guys. You can correct distortion, even vignetting at the cost of some decrease in sensitivity (or increase in noise level), but not loss of information due to aberrations or deficiencies in focusing.

    If you disagree, I look forward to receiving a copy of your filter which will give infinite depth of field to a 50mm f/1.4 lens used wide open, or (possibly more usefully, given the excessive minimum speed of most sensors, and currently available films) reduce depth of field so that distracting backgrounds can be thrown out of focus even when the lens is stopped well down.
  6. Hotblack

    Hotblack Dead Horse Flogger

    Now who's being silly? I thought the correction of optical flaws (ie barrel and pincushion distortion)was what was being discussed.

    IIRC its optical physics that dictates depth of field. No filter can retrieve an OOF background. Some clever photoshoppery can render any background OOF however.

    or are you just deliberately playing the troll tonight?
  7. Benchista

    Benchista Which Tyler

    Actually, it's for landscape work that I find FF absolutely ideal, and I'm not alone - Colin Prior, for instance, has largely abandoned his 617 cameras for a 1Ds Mk II. You say you don't routinely use ISO 800 film for landscapes, and neither did I - that's the point; FF digital opens up new areas that simply weren't practical options before, and that's one of them. It's also just about the only area where I think there is any detectable difference at normal ISO settings - biggish blowups of landscapes show a fair bit more detail than I got with my 6MP APS-C SLR, or can get with my 8MP one. Cost effective? That's a very different issue, but for landscapes and travel, there's nothing better.
  8. Barney

    Barney Well-Known Member

    Ah yes but then Dom wouldn't have his Missus shouting after him going "Oi! Nick! Where has our door stop gone? It's getting ever so stuffy in here!" like you now would he? :D

    Plus he'd still have more money for beer. /forums/images/graemlins/tongue.gif
  9. Benchista

    Benchista Which Tyler

    No, of course not - his name is Dom, so that's what she would be shouting with regard to the doorstop. I hardly ever get shouted at for moving the doorstop, because why on earth would I move the Nikon when I've got so many cameras instead? /forums/images/graemlins/tongue.gif

    Nah, he would've spent it by now. :D
  10. beejaybee

    beejaybee Marvin

    It's not just the size of the sensor elements, or the element count of the whole sensor, that matters. For every lens, and dependent on the aperture set, there is an optimum sensor density derived from the physics of light diffraction. The larger the imaging system, the smaller the diffraction effect. That's why the lenses in compact digital cameras don't stop down below about f/8 - if they did, diffraction would limit performance so much that a pinhole would serve just as well - whereas with medium format kit the optimal aperture / resolution trade-off occurs somewhere around f/16-f/32.

    Making the sensor bigger increases the potential for image resolution as well as increasing sensitivity and/or the signal-to-noise ratio. Whether or not that potential is realized in practice depends critically on the compromises dictated by design and economic manufacture of the lens system. That is fixed for all time, is independent of the technology used for the sensor and is equally true for digital and film cameras.

    My guess is that DX/APS-C sensors are "good enough" for most people but inadequate for others, just as 35mm never completely replaced medium and large format in the film world. Whether four thirds is adequate for the serious amateur market still appears to be an open question.

    The ideal camera would have a huge sensor (let's say 10x8 inches), weigh nothing, require neither power source nor consumables and slip easily into a shirt pocket. In the real world, compromise is inevitable. The working solution is bound to depend on many factors, of which sensor size is only one.

    What really p*ss*s me off is the way that 35mm focal length equivalents are being used instead of actual focal lengths when referring to lenses mounted on cameras which do not use a sensor measuring 36x24 mm. Perhaps we need to dumb down and start referring to lenses as "superwide", "wide", "standard", "tele", "supertele" as appropriate. In fact this makes perfect sense with zooms, where the focal lengths marked are in any case approximate and vary considerably with the focusing distance set.
  11. Barney

    Barney Well-Known Member

    What good would that do? So I put a 35mm lens onto a Nikon F80 it's a wide angle, on a D70s is a standard. Using 35mm equilivent makes sense because these lenses are used on bodies with differing formats. People who shot on 35mm usually have the brains to multiply up to three digits by 1.5 or 1.3 (1.6 might take a little more head scratching) while those who have come to photography during the digital age will know no different and equate the focal length to the image through the view finder.
  12. parisian

    parisian Well-Known Member

    We had this discussion in the lounge last week. The concensus was to leave well alone as regards lens marking.
  13. beejaybee

    beejaybee Marvin

    No. If it covers 36x24 mm it's a wide angle, if it covers only a DX/APS-C sensor it's a standard. If it covers 6x6 it's a superwide...

    Suppose I take a full-frame image and crop to APS-C format. That doesn't change the lens, or its description. It does change the way the lens is being used but that's a different matter. However there is a huge difference between a lens of focal length X designed for DX/APS-C sensors and a lens of focal length X designed for full frame when fitted to a full frame camera, even if the reverse isn't true.

    1.6 is dead easy, multiply by four twice then shift the decimal point. It's 1.3 that's the most awkward for those of us that only learned tables up to twelve times.
  14. Benchista

    Benchista Which Tyler

    Well no - tha angle of view applies to the format in use, not the coverage of a lens, and it would be hugely more confusing to adopt the principle you're suggesting, where two lenses of the same focal length capable of being mounted on the same camera would be described differently on the grounds that the coverage of each was different. You can only describe a lens as a wideangle or whatever with reference to a format - it's not in itself an intrinsic property of the lens.
    As to what happens when you crop, well it DOES change the angle of view of the final image, of course.
  15. beejaybee

    beejaybee Marvin


    Wide angle lens design tends to use a strongly negative front element. Standard lens design doesn't.

    Take an extreme case - a 135 mm lens designed to cover 10x8 inch format. The lens design and construction clearly has to be typical of wide angle lenses. It will clearly function as a 135 mm lens when mounted on a APS-C SLR, giving a strong telephoto effect. But the design and construction does not change from being a wide-angle... Similarly fitting a 135 mm telephoto lens designed for full-frame 35mm to a 10x8 field camera will definitely not give a wide angle effect.

    What I'm saying is that we should NEVER use 35 mm equivalent focal length; ALWAYS use the raw value. After all, we never used to have to refer to an 80mm focal length lens on a 12-on-120 Hasselblad as "41mm full-frame equivalent", or a 50 mm lens on a full-frame SLR as "98mm Hasselblad equivalent". (Figures based on the lengths of the respective frame diagonals.)

    My suggestion to switch to categories was based on the fact that this battle appears to have been lost so far as digital cameras are concerned. The discrepancy with compact digicams is of course even more marked than it is with DX/APS-C format DSLRs.

    I've said enough about this. If you still disagree, that's your right.
  16. El_Sid

    El_Sid Well-Known Member

    Why would we? An 80mm Hasselblad lens doesn't generally get used on a 35mm camera and thats the point here... Lenses designed for a 35mm frame can and are being used on smaller format equipment therefore a working knowledge of their equivalence is useful. 35mm has become so universal that most users, especially the less able/experienced ones, are comfortable visualising a lens' particular AoV in terms of what they know. The longer APS-c continues as a format the easier it will become to visualise the effects of lenses according to their real focal lengths rather than their equivalences.
  17. Benchista

    Benchista Which Tyler

    The definition of a wideangle lens is simple; one that is shorter than the diagonal of the format in question. That's the only possible definition; in itself, there is absolutely no connection whatsoever to construction of the lens.

    It would were the coverage adequate. It makes no difference to the angle of view if the lens is constructed as a telephoto or retrofocus design - if the coverage is sufficient (and clearly it won't be in your second example, but if it was) there would be exactly the same field of view. That you wouldn't get what you term "widengle effect" is purely that there's insufficient coverage of the format.

    Agreed, except that many people DID refer to an 80mm lens on 6x6 as "equivalent to a 50mm lens on 35mm", and indeed there are many, many equivalence tables published in books going back years.

    Yes, but your suggestion is simply wrong, and makes things even worse.
  18. Barney

    Barney Well-Known Member

    Do you want to take this end of the stick because I think you've got the wrong end! You've just disagrered with me and then repeated what I said.

    But you were complaining about using 35mm equivilents. So why add to the confusion by adding another description to already over complicated lens naming systems

    Well not really. for 1.3 just add roughly a third, easy to work out and there's no need to be really precise unless you're totally anal. And your method of working out the 1.6 equilivent agrees with my point that it would take a little more head scratching.
  19. beejaybee

    beejaybee Marvin

    And there's the nub of the matter. If you are using a smaller sensor then, in order to view the image at the same size, you have to magnify it more. Lenses designed for 35mm full frame tend to have a circle of confusion (the diameter at the image plane of the bundle of rays which should be a dimensionless point, if all aberrations were fully corrected) of around 1/30 mm (at full aperture, focussed for infinity). Using such lenses on half-frame / DX / APS-C cameras will lead to relatively poor quality images unless the original lens design criteria were more stringent, the lens is used well stopped down to control aberrations or the image is viewed smaller.

    Interpolation and/or sharpening in software can make an over-enlarged image _appear_ less fuzzy but cannot reveal image detail below the capability of the lens to resolve it. (Image processing can, however, and usually does, result in artifacts which can be much smaller than the resolution limit of the lens, but is sometimes confused with real detail.)

    That's one very good reason to use a lens designed to match the format of the camera.

    Probably the only reason you often (usually?) seem to get away with fitting full-frame lenses onto DX / APS-C format DSLRs is that the central part of the field is likely to have a smaller circle of confusion than the extreme corners. Coma in particular tends to increase sharply the closer you get to the edge of the design image circle. And that's why wide angle lens design is different; controlling the off-axis aberrations and field curvature tend to dominate the design, whereas supertele lens designers can almost ignore these effects.

    Another reason for preferring matched lenses is that there is less out-of-image-frame light to baffle out, absorb or allow to spoil the contrast of the image.
  20. Barney

    Barney Well-Known Member

    You're making about as much sense as Rab C Nesbitt chewing a toffee underwater.

    Your point was ...

    But you appear now to be demanding that we have a complete separation between lenses for 36x24 film/sensors and smaller sensors. Which is even more daft! Why should I duplicate lenses just to use one on my F80 and one on my D70s??

Share This Page