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

    Any content, information, or advice found on social media platforms and the wider Internet, including forums such as AP, should NOT be acted upon unless checked against a reliable, authoritative source, and re-checked, particularly where personal health is at stake. Seek professional advice/confirmation before acting on such at all times.

Is the sensor is the most important factor?

Discussion in 'AP Magazine Feedback & Suggestions' started by idreamedof, Apr 27, 2011.

  1. idreamedof

    idreamedof Active Member

    Quote from AP on their website, highlighting an article comparing Nikon, Sony and Pentax cameras with the same 16.2 megapixel sensor unit:
    "Most photographers believe that the sensor is the most important factor in image quality, so what happened when Mat Gallagher compared three DSLRs that all use the same 16.2-million-pixel unit?"

    I haven't managed to pick up my copy yet, but look forward to reading this article as I'm interested in the Sony A580.

    I hope the point is made that the lens is very likely the most important factor, and here's why I've finally come to this conclusion.

    A few weeks ago, a friend wanted to buy the best camera available and favoured Nikon, so I agreed that the highly reviewed D7000 looked like a good bet. I advised getting it with the 16-85mm, but he got a good deal of a kit with the 18-105mm. We went out on a shoot on the first sunny day to give it a run through. I took along my Sony A200 (cheapest DSLR ever, with the Sony rebate) and, most of the day, I was using a manual focus Minolta 50mm 1:2 MC Rokkor (the kit lens from my SRT100x) which I have adapted to fit the AF mount. Unfortunately the images on 10mp from a lens I bought 34 years ago simply outclassed the results from the Nikon combination (both sets converted from RAW). He was gutted and I was slightly embarrassed. He's since shelled out on the 16-85 and got good results, but still not significantly better than those from the Sony/Rokkor combination. He bought the 50mm 1:1.8 Nikkor last week but I haven't heard how he got on over the weekend – my fingers are crossed for him.

    For me, the point is that it leaves me in a quandary about upgrading my A200 + 18-55mm, which I was hoping to do this year, to an outfit which should give images comparable to the D7000 – the Sony A580 + 16-80mm Zeiss. However, it won't be worth spending a load of money if it doesn't noticeably improve the picture quality I get from my existing kit.

  2. beejaybee

    beejaybee Marvin

    Well, it's important, but the lens is much more so.
  3. El_Sid

    El_Sid Well-Known Member

    I'd agree that the lens still remains possibly the more important factor when all other things are equal because if the lens is not delivering the goods it matters little how good or bad the sensor is. Nonetheless the sensor does have significant bearing on final image quality as does the processing - just as the choice of film and developer did in the days of wet chemistry. It will be interesting to see what the article reveals.

    As far as your friend's experience goes I'm not really surprised. By and large I would expect a prime lens from any reputable OEM to outperfom a typical kit lens in most situations - particlularly on a an APS format camera where the weaker portions of the image circle are not being used. I have a Nikon D50 which and I have the 18-70 f3.5-4.5 AF-S zoom (effectively the 16-85's predecessor) which gives very good results. Nonetheless it can still be outperformed by most of my manual focus Nikkors and other Nikon fit lenses - one example is my Nikkor-Q 135mm f3.5, I've no idea how old it is but it has been converted to the much later Ai fitting so is probably from the sixties, which is incredibly sharp by comparison to the zoom. It also performs the equally well on CMOS sensored EOS cameras too...
  4. daft_biker

    daft_biker Action Man!

    You may be surprised how different the images look from the different cameras. I was :)

    I don't think anyone is disputing how important the lens is but it's perhaps not the only factor worth considering;)
  5. P_Stoddart

    P_Stoddart Well-Known Member

    If you believe DxOMark then the sensor can show how good a lens is. Same lens different camera.

    But it also seems to confirm the old position that prime lens are better than zooms.

    IMHO with sensors getting better we are now see how good some lens were and are.

    Bottomline IMHO is that it is a marriage. Good lens with good sensor = good results. But put a poor lens on a good sensor it shows the weakness of the lens. Put a good lens on a lower res sensor you will hit the limit of the sensor.

  6. beejaybee

    beejaybee Marvin

    Same lens, same sensor, same subject, same lighting, different software - different results. In the old days you could take the same negative to two different master printers & they'd produce different images just using light from the enlarger & wet process chemistry. The number of things you can do to an image, digitally, means the range of results is pretty well limitless. (In terms of apparent sharpness, contrast, tonal range & saturation, etc.)

    But you still need a good lens ... without a good image for the sensor to capture, you won't get a good image output; or, "garbage in, garbage out" in the time-honoured phrase from the computer business.
  7. idreamedof

    idreamedof Active Member

    I've suggested to my friend that he tries an old Nikkor lens, just so he's got something to compare, so it's interesting to hear your experiences. Being new to photography he's finding it hard to accept that such old lenses can be better than new ones. His thinking is: there must have been advances in optical science since the 70s. I take the cynical view that most of the advances have been in marketing :)
    We work in graphic design, so regularly get supplied very high quality digital images from professional photographers, using full frame equipment, so we can see that modern high end equipment is capable of delivering the goods – but the cost is too high to be practical for most amateurs. My answer is to keep the prints small and boost the contrast!

    (Thanks also to everyone for their contributions – all very helpful – keep them coming!)
  8. beejaybee

    beejaybee Marvin

    True, very true (and depressing).
  9. Benchista

    Benchista Which Tyler

    Optical laws haven't changed, so it's perhaps hardly surprising that if you compare a modern kit lens to an old, easy-to-design-and-build prime that the latter perform better. However, if you do a like-for-like comparison, you'll see where there have been real improvements. Zooms at all levels are better than 70s ones, which for the most part were dreadful - only the 70-210s were respectable back then (OK, that's a generalisation...), whereas today even the cheapest kit lenses produce pretty reasonable results at around f8-f11 (although that wasn't necessarily true even 5 years ago - more densely-populated sensors appear to have killed off the real lemons fairly recently). At the top end, Canon and Nikon's current f2.8 70-200s are absolutely stellar. Similarly, new prime lenses have raised the bar several times over the last 20-30 years

    And in all honesty, that's all that most people ever want to do - print relatively small. To do that, kit zooms do a very respectable job and provide decent enough results for most people at a price in real terms unimaginable in the 70s. These days, I use pretty high-level lenses most of the time, both zooms and primes, on fairly high MP bodies (EOS 5D II and 7D) and they're certainly capable of producing respectable print sizes - I find I'm printing more and more at large sizes - I think I've done more 30"x"20" prints this year than through 20 years of using film, and most of 'em have been shot with zooms - albeit Canon L zooms (but not the most costly ones). In my 35mm days, grain tended to be the limiter on large print sizes, so I used to use medium format for anything around that size - I certainly don't feel I'm missing out on resolution at these sorts of print sizes now. I do sometimes use old primes, mostly Carl Zeiss, and they're very good - but then they're very good, if you see what I mean. I also sometimes use quite grotty lenses - a Canon 28-200 on my 5D II, or the kit 18-55 IS on the 7D. Both of these perform surprisingly well, although I suspect the 28-200 quality would fall apart on the 7D. But both have produced usable A3s, and for most people would be quite good enough - the average user producing prints up to A4 would be perfectly well served with cheaper lenses.
  10. MickLL

    MickLL Well-Known Member

    In one sense I have a problem with this article - even though it's a fascinating read.

    When I was doing my 'O' level physics I was taught that if you wanted to see the impact of one variable you kept all of the other variables constant and altered just the one in which you were interested.

    This test set out to see if the different 'treatment' of the same sensor by three manufacturers made any difference to the end output.

    All well and good except that one of the cameras - the Sony - has a pellicle mirror (hope I've got that right) that is said, throughout the article, to impact the results from the sensor.

    It seems to me that the article compares two apples and an orange and I don't quite understand the logic when Sony has another model, using the same sensor, that is an ordinary DSLR and therefore properly comparable with the other two.

    Having said that I found the article to be one of the most interesting for a very long time. More similar please.


    PS Before anyone points it out - yes I am a Sony user (well Minolta actually at the moment) but I hope that the above comments can be divorced from that.
  11. Benchista

    Benchista Which Tyler

    Well I didn't really see it as a comparison test in that sense, and from a detached point of view (not owning any of the brands in question), was much more interested with it the way it was structured, with the non-pellicle pellicle mirrored Sony (they call it soemthing else, don't they?) - perhaps it would've been more interesting still with both Sony flavours checked in, though.
  12. MickLL

    MickLL Well-Known Member

    In the old days ;) the average user took the cheapest print film, went to the cheapest D&P shop, often had the film in the camera all year, was happy with heads chopped off and quality that we on this forum would hate.

    I am led to believe that the average user today isn't so different and, in particular, prints very few photos.

    We tend to forget that we here are NOT average users.

    My point is that lens development hasn't been just about ultimate image quality. It's also been about size, weight and cost and in that respect a good job has been done.

    From a personal perspective though I wouldn't be seen dead using some modern lenses - even ones that the non-average folk on this forum seem to think acceptable. The visible distortion and the impact on verticals and so on would drive me to drink.;););)

  13. MickLL

    MickLL Well-Known Member

    I'm in difficulty here because it's all too easy to be accused of being a Sony fan-boy - especially as the Sony didn't seem (to me) to stack up too well against the other two. However I am where I am and I'm trying to 'speak' as a scientist rather than a camera owner.

    Having got that out of the way let me agree with you that including both flavours of Sony would have added hugely to the value of the test. Firstly it would have allowed a true comparison of the different implementations of the same sensor and also allowed a comparison of the impact of the 'mirrorless DSLR' (I'm so much of a Sony fan that I don't even know what they call it ;);) )

  14. El_Sid

    El_Sid Well-Known Member

    Certainly there have been advances in the science - far more efficient lens coatings and very sophisticated glasses with very controlled dispersion characteristics for instance - but the biggest advance is the advent of computer aided design and manufacturing. Modern lens designs are calculated mainly by computers which can calculate and analyse optical layouts far faster than even the finest human lens designers ever could. The net result is that lenses with very complex designs, such as zooms, have gone through more design iterations in the last twenty years than human designers probably could have managed in a century. Couple this to the inhuman levels of accuracy and consistency of computer controlled manufacturing equipment and the net result is that modern zooms are considerably more capable than they were when your Rokkor was made.

    Having said all this zoom lenses are still far more compromised than any prime lens since they must, or at least should, deliver acceptable quality across their whole focal length range while primes can be optimised specifically for the one focal length they offer. Many optical designs for primes were pretty much perfected long ago and other then adjustments for new types of glass, multi coating and manufacturing techniques have remained more or less the same. This is why old primes can still appear to equal or out-perform modern lenses. Where a difference might be seen, for example, is in extreme conditions such as contre jour lighting, where modern lenses designed for digital use are less prone to flare and ghosting caused by back reflection from the sensor, or on some very high pixel count sensors where the sensor can resolve finer detail than an older lens can perhaps deliver.

    Unfortunately there is some truth in this though I'd suggest there's more far hype in the marketing of digiboxes than the glass...
  15. Benchista

    Benchista Which Tyler

    No fanboy accusation from me, I understand entirely where you're coming from - and part of me is scientist enough to agree, but the rest of me feels that the comparison with the pellicle is just more interesting, entertaining and perhaps informative than a straight scientific comparison, that's all. Yes, if you wanted to choose a brand based on the results, a straight comparison would be better, but I feel the article as it is has broadened my knowledge more than that straight comparison.
  16. MickLL

    MickLL Well-Known Member

    That's perfectly true - I learned a lot and, as I said in an earlier post, would welcome more articles in a similar vein.

  17. beejaybee

    beejaybee Marvin

    In which case 1 MP is more than enough - so why do mass market compact cameras compromise themselves by having far denser sensor arrays with a consequent performance penalty in poor light?
  18. AlexMonro

    AlexMonro Old Grand Part Deux

    Because mass market compacts are not about photography, they're about marketing - and big numbers sell. :(
  19. P_Stoddart

    P_Stoddart Well-Known Member

    I don't think 1MP is enough though.

    A 6x4" dye-sub at 300ppi needs 2.1MP to avoid any stretching by the printer. Then a 10x8 at 300ppi needs 7.2MP. So really makers don't need to make low end compact above 8MP to keep the mass market happy.

    You can get away with stretching a 8MP image on A3 dye-sub at 300ppi.

    I think most low end compact users don't do alot of post shot work like cropping.

    But I agree some reasonable camera designs get hurt by pushing more pixels on the same size sensor at the moment.
  20. idreamedof

    idreamedof Active Member

    I know it may seem to some to be a cynical view, but I entirely agree. There is a huge new market for the quickly evolving digital camera and it's the manufacturers that capture that market that will survive.

Share This Page