Which lenses are good enough for 21MP?

Discussion in 'Canon EOS' started by christian_balslev_van_randwijk, Sep 26, 2008.

  1. Hello again,
    it seems that a 21MP sensor will severely test any lens, so I'm guessing (but don't know) that only L-glass will
    have resolution to match the resolution of such a sensor, and probably only the latest and greatest L-glass, or am I
    mistaken? Would love to hear from any 1ds mark iii users. I'm guessing, if you use a less than perfect lens, on such
    a high resolution sensor, you would get poorer results, than if the same lens was used on a sensor with lesser
    resolution, or am I completely mistaken?
     
  2. It won't. 5D MkII has the same pixel density as an 8MP camera. So if a lens was good enough for your 30D, it should be good enough for MkII.

    What may cause more problem (and it is true for original 5D also) is the corner performance of a lens. Since the extreme corners on APS-C camera are automatically cropped out, the corner performance of lens does not matter as much.
     
  3. That is the 5D MkII has the same pixel density as an 8.2MP 1.6x crop APS-C DSLR(like the 20D and 30D).

    The current EOS camera which most severely tests the resolution of lenses is, in fact, the EOS 50D, not the 5D MkII or EOS 1Ds MkIII. You'd have to have a full frame camera with 38MP or more to exceed the pixel density of the 50D.

    Zafar is absolutely right that it's the edge and corner perfromance of lenses that will be more severely tested with full frame cameras, not center resolution.

    In the center, if it's good enough for your 20D, it's good enough for your 5D MkII
     
  4. According to canon the reason that the 24 1.4 was redone has a lot to do with the fact that some of the older designs could no longer resolve with the new cameras such as the 1Ds3 and 5D2. So while I hear you guys saying the pixel density is the same there must be more to it than just that...
    JC
     
  5. "So while I hear you guys saying the pixel density is the same there must be more to it than just that... JC"

    Yep, gettin' ready for the 50MP 1Ds MK IV...
     
  6. "So while I hear you guys saying the pixel density is the same there must be more to it than just that... JC"

    The 50D can be included with the likes of the 5D2 and the 1Ds3. As others have said, the 50D has very high pixel density and also uses just the centre of full frame lenses so is therefore far more demanding of the optics.
     
  7. "Hello again, it seems that a 21MP sensor will severely test any lens, so I'm guessing (but don't know) that only L-glass will have
    resolution to match the resolution of such a sensor, and probably only the latest and greatest L-glass, or am I mistaken?"

    I believe that there are many non-"L" lenses that are more than sharp enough for a 21mp sensor. They may not be that sharp at
    every f-stop, but neither will all "L" lenses.

    I only own 4 lenses, all non-"L", and the 100 f2.0, and 50 f1.4 certainly out perform my 13mp 5d sensor. Also, the 35 f2.0 when
    stopped down a couple stops is quite sharp in the center as well. My 24mm f2.8? Eh, maybe not...

    My point being that the "L" lenses are not the only very sharp lenses in the Canon inventory. The "L" lenses are fast, and high
    quality zooms, and very fast prime lenses. For those who don't need these expensive features, other Canon glass can be just
    as sharp, but without the features just mentioned.
     
  8. The 50D has a pixel density which would correspond to a 38MP full frame DSLR - the 1Ds MkIV maybe?
     
  9. I can't remember anyone worying about this in the days when film ruled the world. So why are we concerned when the digital sensor takes another step towards the analog continuity of film?
     
  10. I can't remember anyone worying about this in the days when film ruled the world. So why are we concerned when the digital sensor takes another step towards the analog continuity of film?
     
  11. `I can't remember anyone worying about this in the days when film ruled the world`

    I guess it was cheaper to buy better film than a better sensor. plus most only got lab corrected and printed 6x4 tho they were 3 1/2x5 when I started. Not many ever saw images on a screen 100% 50+ x 36+ inches :)
     
  12. Well maybe trannies, but there was a projected viewing distance :)
     
  13. >> I can't remember anyone worying about this in the days when film ruled the world.

    Me too. That said, I can't remember worrying about going for a drive without my cell phone or letting my kids come home late. Standards has changed.

    Happy shooting,
    Yakim.
     
  14. Let's not the forget the nifty fifty (Canon 50mm f1.8). It's the cheapest lens out there and will easily perform on the 5D MkII.
     
  15. Not addressed in the above responses (I'm pretty sure) is the OP's last question:

    "I'm guessing, if you use a less than perfect lens, on such a high resolution sensor, you would get poorer results, than if the same lens was used on a sensor with lesser resolution, or am I completely mistaken?"

    I'm afraid you're mistaken, but it's a common misperception. This is more a question of what is the weak link in the chain. If you put an oversized link in a chain (analogous to a higher resolution sensor), that doesn't make the chain weaker. But neither does it make it stronger. If you have a sensor that can resolve all of the details in the image of one of your lenses, then a higher resolution sensor will not improve your image quality. But neither will it give you a lesser image quality. Does that make sense? What a higher resolution sensor will do for you is to resolve more detail in better lenses. Think of it like buying a better sound system for someone who is hard of hearing. Will the music sound worse? No. But the listener might not be able to discern any improvement.

    Turning this issue around, if you have a nice lens whose details are not fully resolved by the sensor you have, and if you move up to a higher resolution sensor, then you'll get more detail in your images.

    Finally, looking at the issue in still a different way, if you have a lens with certain image defects, you may resolve those defects a bit better with a higher resolution sensor. For instance, if your lens has a bit of chromatic aberration to it, then you'll see that problem a bit better with a higher resolution sensor. However, you'll only see it better if you blow the image up larger or "pixel peep." (If you keep the image size the same, you might not notice much difference.) It is for this reason some photographers like to say a higher resolution sensor is "harder on" lenses. Some people incorrectly read this as meaning that higher resolution camera bodies somehow wear out lesser lenses faster. However, put more accurately, a higher resolution sensor tends to reveal the deficiencies of a given lens -- again, if you look closely enough.
     
  16. zml

    zml

    There is more to lens performance on digital than pure resolving power, so a lens that tests great in terms of lpm may perform poorly on a high res sensor (esp. when pixel-peeping at 100%) Say, the spectral characteristics and internal reflections of lenses play a major role here (my guess is that that's the reason for the new 24/1.4L II and new Canon lens coatings in general) Having said that, the only lens I’ve used and consider unusable on a 21 mpixel rig (1Ds3, sorry, no 5D2 yet to test) is 50/1.8 which simply lacks contrast and has too many aberrations up to f/4 (it is OK at f/8 or so, but other lenses beat it by a wide margin even at f/8.) All other lenses I've used perform great. Now, when pixel peep, one can see some differences in contrast, edge/corner sharpness, the amount of CA, and color rendition, but most, if not all, of these differences are not visible in printed output (or in any other human-sized output.) But yes, I suspect that a $200 kit zoom may totally fall apart in terms of IQ on a high-res camera. The bulk of my shooting is done with 24 T/S, 24-70 and 24-105 with great results, which is not the conclusion one would arrive to by reading some internet newsgroups :) <br><br>As for film, throw a slab of film on a color corrected light table, take a 4x APO viewing loupe and amuse yourself with CA, lack of contrast/detail, corner sharpness, tiny dynamic range, etc. on many a slide you consider perfect. And yes, some general purpose films (Tech Pan, Agfa Ortho) did outresolve most general purpose lenses (in terms of pure lpm in optimal conditions.)
     
  17. Even the oldest EOS lenses worked just fine with Kodachrome film, and I'm guessing that there are relatively few of them that will be beyond their usefulness with any existing digital camera. Anyway, listen to Sarah, she's got it spot on.
     
  18. Recently I found a link of Ken Rockwell'e tests about one of the cheapest 28-80mm Nikkor kit lens in one PN threst ( http://www.kenrockwell.com/nikon/2880.htm ), That was very interesting for me because I have one second hand that one. Earlier I used to thought that that is not giving me its best and and may be little low at resolution because when I shake it, it makes some loose like noise and it is old and may be used heavily and roughly by wedding professional. Its result were good when I bought it but I lost that quality because I started using it like my other lenses. But when I am writing this, I am scanning my recent sunlight shot-out negatives in , and the results are again nice.

    So I found that every lens has its con and pros, and results depend on their proper use, and we have to decide when, how and what to use use at where and when by testing the lens ourselves.

    I am still trying to understand lenses.
     

Share This Page