Is a lens designed for a crop sensor camera essentially a better choice than a FF lens

Discussion in 'Canon EOS' started by victor_oyarzun, Sep 16, 2011.

  1. Several times I have read in forums the opinion (by experienced photographers) that the lens designed for APS-C cameras are a better choice than the FF ones (given same optics quality) but, I have never seen that someone actually explains the reasons. For example the Canon 17-55mm is a better choice than the 24 – 105L mm or the 24 – 70L mm. or the 16 – 35L mm.
    I understand that 17mm is a wider angle than 24 mm. and some would like to have it but, normally the discussion is about image quality not focal length. Common sense says that a component specifically designed for another would perform better but, is this the case?
    Actually, one opinion that contradicts that thinking is that a crop sensor camera will see only the center of the FF lens where it is sharper.
    So, the question is: Is a lens designed for a crop sensor camera essentially better (better image quality) than a FF lens with same optics quality and focal length?
     
  2. The four lenses you mention all have excellent image quality. So that's not really a reason to choose any of them over the other.
     
  3. Short answer: Yes and no -- sort of -- but mostly yes.
    Longer answer: I wrote about it in an article comparing FF and APS-C formats. The section addressing FF vs. crop lenses is here...
    http://www.graphic-fusion.com/fullframe.htm#resolution
    ... but other relevant points are discussed in the following two sections as well.
     
  4. If you ever plan on moving to a FF sensor body, then I would avoid purchasing a crop lens.

    Your investment in lenses can be many times the cost of the body. If you spend $1000 today for a crop sensor body and $4000 on crop sensor lenses, then 2 years from now decided to go a full frame sensor, you get to ebay all your crop sensor lenses and start over.
     
  5. Thanks Sarah, good writing I learned from it.
    Tudor: Yes, I see your point but, my question is more a theoretical than practical. I just want it to understand if the fact that the lens is designed for a crop camera has advantages over the FF in the same camera.
     
  6. SCL

    SCL

    IMHO if you are using a crop body, a crop lens MAY be a better choice (assuming the optical formula is the same as one for a full frame) if weight is an issue and you do not plan to use it on full frame bodies. Otherwise I see little reason to go with a crop lens if a full frame equivalent is available, AND you have full frame bodies....which I do!
     
  7. I don't have Sarah's level of expertise, but I would answer your question in a different way. I would suggest that whatever the answer to the general question in principle, which is what her article describes, it is often not very helpful in practice to think of the choice in such general terms. It's not as though you often face a choice between two lenses of identical quality and identical specs except that one is designed for a crop and the other isn't. As someone who shoots only crop, I find it more useful to look at the range of lenses with the characteristics I want (length, zoom factor, speed) and read reviews about how they perform. I have three lenses designed for FFformat, and all function superbly on my crop. I also have EF-S lenses that I purchased when those specific lenses seemed to be the best among relevant choices. I regret none of these choices.
     
  8. In general no. In theory for short focal length lenses a EF-S lens might have an advantage since the backfocus distance is shorter and to the retrofocal constraints are less severe, but if probably doesn't make much difference. Of course if you want a 10mm rectilinear lens for your APS-C body, the point is moot since you have no choice.
    Note that the shorter backfocus distance applies ONLY to EF-S lenses. It does not apply to crop sensor lenses from Sigma, Tamron or Tokina (or any other 3rd party lens), since they have the same backfocus distance as full frame lenses.
    [Note Backfocus = distance from the rearmost element of the lens to the sensor]
    At longer focal lengths (say >35mm) the FF lens is likely to show lower edge/corner softening and lower levels of distortion and vignetting. The price you pay is having a slightly larger and heavier lens.
    It is possible that an EF-S lens could be slightly sharper in the center than an EF lens, because it's easier to design a sharper lens with a smaller circle of coverage. However I think this is mostly a theoretical advantage when comparing EF and EF-S lenses. The difference in image circle size is small. If you were comparing a EF lens with a lens designed to cover 8x10, then the EF lens would certainly be capable of being sharper.
    Bottom line - don't worry about sharpness, just buy the lens with the focal length range you need and the price you can afford.
     
  9. ...$4000 on crop sensor lenses...​
    That's scarcely even possible; for that money you can buy them all.
    Realistically this discussion is only about one lens - the general purpose wide-to-tele zoom. For telephoto there's generally no problem sharing crop/full frame lenses; for ultrawide you pretty much have to get a dedicated crop lens.
     
  10. There are some very good crop lenses. One big advantage is the fact that they can be significantly smaller than a full frame lens. A Tamron 17-50 2.8 (1.2 lbs) is a lot smaller than a Canon 24-70 (2.1 lbs), and the Tokina 50-135 (1.8 lbs) is a whole lot smaller than a Canon 70-200 2.8 (3.2 lbs).
    An APS-C camera and the two EF-S lenses weigh in at just over 4 lbs. (1.1 lbs for a rebel body + 3 lbs of lenses) An equivalent 5d Mk ii setup weighs in at 7 lbs. You also save a big chunk of change.
     
  11. Having written what I wrote, I do have to agree with others here. I shoot both FF and APS-C. I use almost all FF lenses, even on the crop body, because it's simply more economical and because my FF lenses are pretty good (mostly L optics). However, when size and weight are at a premium, I also carry around the humble little 18-55IS on my crop body. (The two will stuff into my "purse" -- a Crumpler 4 Million Dollar Home -- and carry easily, so that I can have a "just in case" camera with me.) I've taken some nice pics with that combo.
    If I were absolutely certain I'd have no interest in owning a FF body, I'd probably opt for EF-S optics, simply on the basis of size, weight, and cost. However, if I were uncertain how long I might want to own my lenses, I might opt for "L" lenses simply for their excellent resale value. All of my L lenses are worth more now, used, than I paid for them new, while my consumer lenses have depreciated or simply maintained their value. (This is due to the shrinking US dollar, of course.) There are many factors to consider in selecting lenses, other than native format advantages.
     
  12. It really depends on whether you usually shoot wide open or stopped down to, say, f8 or so. Adding to, and recapitulating a little, to what Bob Atkins said, the main difference is greater light fall off in APS-C lenses vs FF ones. But this difference is mitigated to a great extent upon stopping down. So an f2.8 FF lens can be shot at f2.8 and have far less light (and corner sharpness) falloff than its APS-C equivalent. But if you're shooting at f8 it's unlikely you'll see any difference.
     
  13. I agree with those that say "yes"
    you can easily see this with the MTF resolution data on crop models at sites like photozone
    www.photozone.de or lens comparison sites like
    http://www.the-digital-picture.com/Reviews/ISO-12233-Sample-Crops.aspx?Lens=100&Camera=474&FLI=0&API=0&LensComp=398&CameraComp=474&SampleComp=0&FLIComp=0&APIComp=1
    EF-S type lens have a narrower light cone (using all the light, not just the center!) which matches the smaller sensor size and this helps resolution ----- but there are other factors that contribute to a lens characteristics !
     
  14. Brett,
    I don't follow your light cone comment. Why would the light cone be different? The fact that a FF lens would have a larger cone is negated by the fact that the crop sensor only sees part of it, exactly the same part that it would see through a crop lens. The density of photons is even (comparatively) across the total light cone, it doesn't matter what percentage of that light cone you sample at the same registry distance. Which leads back to Bob Atkins point about Canons EF-s reduced registry distances.
    On this subject it would take a brave and very well versed lens specialist to contradict his thought process. And as he so very eloquently summarised, just buy a lens you need and use it.
    Personally I found achieving potential resolution advantages in crop cameras to be unrealistic in real world imaging, I could tell you what tripod I used from some of the files from some cameras when extreme pixel peeping. Differences that small will never show up in a print or in electronic reproduction, and are never achievable without optimum technique and subject conditions.
    Technique, subject contrast and lighting conditions will be the limiting factors for resolution for all DSLR's sold new today. It is fun to theorise but the differences are academic.
     
  15. The larger light cone of a FF lens will give more margin for error to avoid vignetting and softness in the corners. The larger light cone of a TS-E lens (larger than a FF lens in order to allow for the tilt and shift) is one big reason why their IQ is so good when just used as a prime lens.
    That said, theoretically a well designed EF-S lens can equal the same FF lens and be lighter and more compact. If there's such a thing as an EF-S with absolutely no vignetting or softness outside of the center, then there'd be no advantage in going to a EF lens designed for FF. This is more likely to happen with a prime lens. With most zooms, particularly wide zooms, the EF-L lenses for FF bodies are likely to be superior to the EF-S lens of the same zoom range, but not necessarily always.
     
  16. "I don't follow your light cone comment. Why would the light cone be different?"


    because the diameter of EF-S lens is smaller !
    Consider the small sensor and the small lens of the iPhone, it's matched to give the best resolution. The same principal is similar to EF-S lens. All of the light is used in a lens whether EF or EF-S. The light that strikes the sensor perpendicular and straight is good (optically speaking) but some light in a larger diameter lens (EF) will strike it obliquely at an angle which is not so good
     
  17. Sorry, I didn't word my query well. I understand the light cone of the EF-s lens is smaller, but I don't understand why that is more efficient than using a fraction of a bigger light cone, from a photon per area point of view. If the registry distances are the same, as they are for non Canon crop lenses, where is the advantage purely from a photon per sensor area point of view?
    I don't think the cone size, in these cases, is as important as the glass transmission indices and characteristics. I don't see why using a wide coned f2.8 24mm lens on a ff camera is less efficient, when used on a crop camera giving an effective fov crop to 36mm also at f2.8, the light transmitted to the sensor, per area, is exactly the same. Besides the center of a bigger cone could be just as good as the extremes of a smaller one.
     
  18. Sorry, I didn't word my query well. I understand the light cone of the EF-s lens is smaller, but I don't understand why that is more efficient than using a fraction of a bigger light cone
    touched on here:
    http://www.graphic-fusion.com/fullframe.htm#resolution
    because more light ~photons~ are striking the sensor perpendicular (straighter) than at an angle

     
  19. Brett,
    First off an edit to my previous posts, I didn't, of course, mean registry distance, I meant backfocus.
    But I don't find Sarah's article (your linked explanation) convincing as written, it seems more like conjecture in places and I'd certainly argue the point about vibration. But it still doesn't answer my point. If the lenses have the same backfocus then they have the same cone angle, whatever percentage of that cone you record, is exactly the same.
    In the specific instance of some Canon EF-s lenses, where there is a different backfocus distance on some of the wider EF-s lenses, there might theoretically be an advantage in cone angle, but for every other crop lens, including all Nikons DX range and some Canon EF-s lenses, the cone angle is exactly the same for FF or crop.
    I still can't understand your comment.
    "EF-S type lens have a narrower light cone (using all the light, not just the center!) which matches the smaller sensor size and this helps resolution "
    How does a narrower light cone, in and of itself, help resolution?
     
  20. If we read Sarah's article closely the example she gives is using a lens designed for full frame on a compact P&S camera, not a crop-sensor dslr. Of course, a lens designed for a smaller than postage stamp sized sensor is going to work better with that sensor than a lens designed for full frame.
    I've read several reviews of TS-E lenses saying that one big advantage is the larger image cone of those lenses leading to better edge sharpness when used as prime lenses (no tilt or shift) on FF cameras.
    I can't notice any difference between using my L-series lens on my 5D2 and 7D, but I DO use software (DxO's Optics Pro) that accounts for differences in the bodies and adjusts for vignetting, CA, geometric errors, etc. at each focal length and each aperture.
     
  21. A couple of points about the light cone.
    First, it is obvious that, at least at shortish focal lengths, it is easier to design a lens of given focal length to provide the same absolute resolution across a crop sensor as compared with achieving that resolution across a FF sensor. So you can either have a simpler, lighter, and cheaper lens of the same absolute resolution, or you can invest in making the loens better so as, for example, to achieve the same resolution relative to sensor size.
    Secondly, a point I hve not seen discussed, and I would be interested to know whether it is a significant issue. If you look at the back end of, say, a 135/2, you will see that it has a rectangular baffle with rounded corners, undoubtedly designed to mask some of the light rays that would not strike a FF sensor because the sensor is not circular. Why? Presumably to stop them bouncing around the mirror box and perhaps not being absorbed, ultimately hitting the sensor and reducing contrast. The implication of there being a baffle is that it can make a difference, although actually most lenses do not have it. If you use a FF lens on a crop sensor body, then all the light rays that would hit the FF sensor but not the crop sensor are free to bounce around. It would be interesting to know whether this ever creates a problem in practice, and if it does whether the use of a longer hood can counteract it.
     
  22. Don't worry about it.

    Compelling photos are far more driven by photographer vision and creativity than tiny differences in
    lenses.The energy put into sorting out small lens differences will pay larger dividends when redirected towards thinking about and creating more powerful photos.
     
  23. Is a lens designed for a crop sensor camera essentially better (better image quality) than a FF lens with same optics quality and focal length?
    The image quality is about the same. An EF 16-35/2.8 and an EF-S 17-55/2.8 IS produce similar image quality. In terms of line-pairs per millimeter they are about the same. I'm sure there are those who can measure some differences, but who cares.
    However:
    • Crop lenses do offer more focal length. Why choose a 16-35/2.8 when you can get an extra 20 mm and IS with the 17-55/2.8 IS?
    • Crop lenses are often cheaper. 17-55 is less expensive than 16-35 or 24-70.
    • Crop lenses offer focal lengths not available on FF. There's no 10-22 for FF.
    • Crop lenses offer more useful focal length ranges, compared to FF lenses adapted to crop. 24-70 really is much more fun on a FF cam when it can go wide.
    So in short, crop lenses are often the more practical solution for a crop cam. But disregard anyone who says they're much much sharper. That's simply not true.
     
  24. Arie,
    That is a totally spurious argument, if you are trying to talk equivalence, meaning getting exactly the same images, angle of view and depth of field, from the same position by using different lenses on different formats, between systems you can't compare the EF-s 17-55 f2.8 IS to the 16-35 f2.8 or the 24-70 f2.8 on FF.
    • The true FF equivalent to the EF-s 17-55 IS is the EF 24-105 f4 IS, f4 on full frame gives the same depth of field as f2.8 on a crop camera, 17-55 on 1.6 crop equals 27-88 on FF, a poor range compared to the 24-105. Both have IS, The 25-105 is weather sealed. As for price, well the 17-55 is $1,120 the 24-105 is $999. Any way you look at it the 24-105 on FF is a better buy than a 17-55 on a crop camera.
    • Not when you compare actual equivalent lenses.
    • 10-22 on crop is 16-35 on FF, and it is a variable aperture, the 17-40 is the FF equivalent, giving a wider range, less depth of field, constant aperture and again being cheaper. The EF-s is $820 the 17-40 is $797. What there isn't is a crop camera equivalent of a 16-35 f2.8 on FF, it would need to be a constant aperture 10-22 f2.
    It seems, and is a surprise to me, if you want a camera and good three zoom kit, apart from the body price and weight, there is little to favour the crop camera route. Having said that for the vast majority of users a 7D is a more capable camera than the 5D MkII.
     
  25. Scott the OP asked about using EF vs EF-S lenses on a crop camera. Not whether a FF camera with a 24-105 mounted on it is better or worse than a 7D with 17-55.
     
  26. Thanks all for all the responses, I am glad that generated a discussion that I can not participate because my knowledge is limited. However, they give a better understanding of the matter.
    What I take from the responses is that there is no significant image quality advantages in the lenses designed for crop sensor cameras.
    Regards.
     
  27. I think "significant" is the key word here. This is obviously a controversial subject. There are a few people on this list who are optical engineers and could address this issue better than any of us, but they haven't weighed in. I'd sort of hoped they would. Ah well...
    Scott is correct that I've introduced a few points of of conjecture in some of my discussions. I'm not an optical engineer and don't have an intimate working knowledge of the compromises in optical formulae that must be considered in lens design. However, I do know this: Wide angle lenses are awkward and difficult to design, as compared to "normal" and tele lenses. The wider the lens, the more awkward and clumsy the design. A FF lens on a crop body has a wider/clumsier design than it really needs to have, considering the smaller image circle requirements of a crop sensor. That can't be helpful for image quality. But yes, that's partially conjecture! ;-)
     
  28. I think if you own a crop camera/s alone it makes sense to maximise your lenses for than format. Even if you are of the mindset that you might get a "FF" camera at some point in the future. If you own a dual format kit there is no IQ disadvantage to getting lenses to use across your formats. But as Arie points out, crop camera lenses focal ranges really are better designed for crop cameras and are a better idea than trying to fit the square peg of "FF" lenses into the round hole of the crop format.
    Sarah,
    I can't hold a decent discussion on lens design, but, in the interests of a pleasant conversation, the problems I have with your thoughts on vibration didn't, I believe, take two important factors into account. One, the relationship of sprung to unsprung mass of crop camera mirror and shutter mechanisms when compared to that of their "FF" cousins. And two, the necessity to keep the crop camera more steady for a give enlargement. I can't satisfactorily work the figure out for myself, but as a minimum the crop camera has to move 1.6 times less than a "FF" one to get the same subject blur (assuming a static subject), but that figure might arguably be closer to 2.6 times steadier.
    Just a thought :) Scott.
     
  29. Using the same focal length of lens on both a FF and a crop-sensor body will result in the same amount of image stabiliization for both. Expressed another way, the adequate shutter speed for a particular image on a FF camera will also be adequate on a crop-sensor camera. The focal length is the same on both cameras and the crop-sensor merely has a narrower angle of view.
     
  30. Thanks, Scott! I'll have to go back, now, and see what I said about vibration. ;-)
     
  31. David,
    If you consider the same size print or screen image size, a crop camera capture has to be enlarged 2.6 times more, that means any movement will be enlarged more too.
    If you use the same focal length but do not crop the FF image (obviously making the fov and image different) the crop camera needs to be at least 1.6-2.6 times more stable for the same camera movement as a percentage of sensor size.
    If you crop the FF image to match the crop image camera movement is equal, you are enlarging the same area of sensor capture and also camera movement.
    If you do what most people do and use the correct focal length to frame your subject the same through both viewfinders and reproduce both images the same size, the crop camera, again, must be kept at least 1.6-2.6 times more stable.
     
  32. Scott,
    To get an image equivalent to a 16mm lens on a FF sensor, you'll use a 10mm lens on a 1.6x crop-sensor, requiring less IS, although the difference will be very small. Going to supertelephoto, to get a print with subject the same size as on 1.6 crop-sensor with a 500mm lens you'll need a 800mm lens on a full frame camera, requiring considerably more IS.
    OTOH, you correctly said:
    If you crop the FF image to match the crop image camera movement is equal, you are enlarging the same area of sensor capture and also camera movement.​
    Hence, for the same subject size in the same size print, IS is equal if the same lens is used.
    Dave
     
  33. Let me try to explain my thought process better.
    Scenario one: You use the same lens on both cameras from the same place. The framing is different but the subject is reproduced on both sensors the same size. In print or on screen if you want the subject the same size the enlargement ratio is the same, though the full frame image in total is larger. Because of that, any camera shake is enlarged the same, the camera systems are equal. In that instance you have just cropped the FF image down to the same size as the crop camera to make just the subjects printed the same size, so all movement is equal. In this instance, say wildlife shooting with your longest lens, the crop camera should show an advantage due to the higher number of pixels landing on the subject. My testing showed it does, in optimal circumstances, though not by anything like the numbers say it should.
    Scenario two: If you use the same lens from the same place with both cameras and make the enlargement of the total capture the same size, ie you make two 8x12 prints of your full captures, you end up with two differently framed images, but the crop camera motion blur is enlarged 1.6-2.6 times as much. Ergo the crop camera needs to have been kept steadier than the FF one. In this instance you may be using your widest wide angle to get as much of a scene as possible in one shot, a ball game and crowd or something, a scene setter.
    Scenario three: You have a 70-200 on both cameras. You shoot a subject for optimal framing, with the crop camera you are at 100mm for the same framed image with the FF camera you have to use 160mm. To make two more 8x12 prints you have to enlarge the crop camera capture by 2.6 as much to get those prints. You do have to use 1/100 of a sec for the crop and 1/160 for the FF, but that is only a 60% difference, not a 260% difference. Again, to my reckoning, you need to keep the crop camera 1.6-2.6 times steadier, minus the 60% allowance for focal length changes, or 100-200% steadier. Very common scenario, portraits, macro etc, anywhere you have the focal length to get the framing and perspective you want.
    The difference between a 500 and an 800 is again 60%, you always need a 60% faster shutterspeed for the FF camera to negate the additional focal length leverage on camera shake, but FF camera equivalence allows you 100% gain, or one full stop, in iso anyway for the same noise levels. So you actually gain another 40% working those two figures in. What you can't get round is that for the same sized print the crop capture has to be enlarged 2.6 times as much.
    This is all just talk though, I am not suggesting for one second that you can't get sharp images out of a crop camera. It was just a logical extension of my thought process, and as I am currently stuck in the middle of the Caribbean Sea I don't have much else to do :)
     
  34. I am currently stuck in the middle of the Caribbean Sea I don't have much else to do :)
    Scott, you are a truly sick person! ;-)
     

Share This Page