1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.
By dirk on Sep 12, 2017 at 1:02 AM
  1. dirk

    dirk CI-Founder

    Top Poster Of Month

    Joined:
    May 15, 2002
    Messages:
    1,554
    Likes Received:
    9
    Camera brand:
    many/ see my profile
    City/ Country:
    Germany
    Very often I can read on the internet, that MFT or other systems with a smaller sensor are not as good as APS-C or other bigger sensor sizes just because of the inherent smaller sensor size. Of course there are differences. Of course bigger sensors have advantages regarding tonality, dynamic range, noise etc.

    But the question is, does this matter nowadays at all? And if yes, when?

    I think we all can agree that the sensor technology is very mature nowadays. The level of "good enough" for 99% of the users is already reached. For ca. 70% of the users already with 1 inch sensors and the bigger you want to go in sensor size depends on what you are shooting and how big you want to print.

    If you want to photograph a black cat in a dark cole mine, you need obviously a fullframe sensor. If you want to have noise free images at very high ISO settings, too.

    But how many of us want to do that really?

    Did you check already in Lightroom, how many of your photos are done with ISO numbers above ISO 1600? And if you look at those photos with ISO number of 3200 and higher, what was the reason why you had this high ISO number? Was it really necessary?

    I looked at my images and I came to the conclusion, that with very few exceptions, I do not shoot higher than ISO 1600. Surprisingly In most cases not higher than ISO 800. The more resoltion my camera had, the higher the ISO number was becauss I needed a higher shutter speed to make sure everything is tack sharp with 36MP. High MP sensor is not always an advantage, I can tell you ;)

    The other few exceptions have been necessary because of lack of image stabilization or just because I needed a very high shutter speed of 1/1000s and faster for indoor sport events, which I no longer do.

    Having realized that, this gave me the opportunity to rethink my current gear and I made some drastic shifts towards smaller sensor formats.

    There is a site called "Imaging Resource" which does a lot of gear reviews. Among others, they make test prints on inkjet printers to see how big you can print with each camera they tested and how good the print result will look like with different IS settings. I find the results very interesting.

    As an example I show you an exverpt of a print test made with the Pansonic Lumix GX807GX85 (microfourthird system) and one made with teh Nikon D7200 (APS-C sensor).

    The results of the APS-C sensor camera is of course better. But under which circumstances and at what print size until which ISO setting? Look at this:

    Panasonic_GX80_vs_Nikon_D7200_IR_print_analyse_markiert.jpg

    So here we are not talking anymore theoretically or from one fanboy to the other. Here we see real life results. Not at 100% photoshop, but in a print.

    I have only an Epson A3+ printer. And my house is not big enough for all the prints I would love to hang on my walls. I do not know how it is about you. But A3+ (13x19 inches) will be my biggest print size forever. Most prints of mine will be significantly smaller. Not enough empty space on my walls :(

    In this screenshot from IR you can see what it means if you put the bottle neck on the prints size and a borderline for the maximum ISO setting you will use in 95% of your images. For me its A3+ and ISO 1600.

    Will the system you are interested in be able to achieve the print quality you need? (not what you theoretically want).

    I would assume you want to have at least "very good" image quality, right? If that is correct, you can achieve this already with a "small" sensor size of the MFT system, if you do not want to print larger than A3+ and stick at or below ISO 1600.

    Of course you can print larger. You can upscale the image in Photoshop etc. and still have very good image quality at larger print sizes. But the print-tests of IR show you the results without doing any photo retouching.

    Do not get fooled by differences which you can measure or which you can see on teh screen at 100% magnification. Look for visible differences in your desired print size.
     

Comments

Discussion in 'General buying advise - The first steps' started by dirk, Sep 12, 2017.

    1. dirk

      dirk CI-Founder

      Top Poster Of Month

      Joined:
      May 15, 2002
      Messages:
      1,554
      Likes Received:
      9
      Camera brand:
      many/ see my profile
      City/ Country:
      Germany
      There is another good source for information on the internet: The website of Thom Hogan. It started as a Nikon only site, but became later an informational site about all the different brands, with and without mirror.

      Thom Hogan started in 2007 to talk about "good enough" for prints and made an overview table specifically for different print sizes and MP on the sensor.

      thom_hogan_printsize_2007.jpg

      Source: http://www.bythom.com/printsizes.htm?=


      Quote:

      "...But let's get back out to the main point I'm trying to make: if you set up a shoot correctly (exposure, camera settings, shot discipline, etc.), use the base or next ISO value of the camera, manage the post processing correctly, do only modest up-sizing (if any), and pick the right options from your printer driver, then you should be able to get that good or excellent print out of virtually any of the currently available DSLRs (attention: as of 2007, so 10 years ago) on the market at up to the maximum size the desktop inkjets can produce. Many of us manage to do better than that. I've produced and seen 36" prints from a 12mp camera that look excellent, though it takes a great deal more control over every variable from shoot-to-print to do that with any consistency..."



      And in another article of 2014 he was talking about a statement from Olympus, which was printed a few years back of 2014, in which Olympus said that 12MP is enough for most users. Thom Hogan then invents two categories of shooters, the "optimal shooter" and the "non-optimal shooter". He is then analysing which shooter category needs which MP to print their desired print size.

      Quote:

      "... Let’s start with the “optimal” shooter. This person has the lenses they need, composes carefully in the field, understands their camera’s dynamic range and how to maximize it, and is a good post processor of raw data. What can they currently do with 16mp (m4/3 and APS)?

      • Print to 15” with no resizing (e.g. presenting the printer with 300 dpi)
      • Print to ~24” with good resizing techniques
      • Easily shoot to ISO 1600 without worrying much about noise, probably ISO 3200 with care
      • Shoot at 4 fps, maybe faster depending upon the camera,
      • Shoot with anywhere from adequate to very good focus capabilities, also depending upon the camera
      Do these users need more than 16mp? Likely not. Looking at the above list, most everyone in the optimal shooter category would probably say that would want better focus (bullet 5) or perhaps better low light capability (bullet 3) before more pixels (bullets 1 and 2).

      What would they get from more pixels? Larger print sizes, basically. Possibly more edge acuity at the same print size. Though these would have to balanced against issues that might arise from using smaller photosites, e.g. noise. Sensor technologies do get better with regularity, so the noise issues would probably be mitigated somewhat with time. ..."



      and

      "... Okay, what about the “non-optimal” shooter?

      First, I guess I have to explain what I mean here. The biggest category of non-optimal shooter is what I call the “heavy cropper.” They take shots with wider-than-needed lenses and then attempt to discover the picture via cropping in the digital darkroom....
      "

      "....
      Are these non-optimals printing large? Probably not. Most are sharing via the Internet as far as I can tell. But even in the example I just gave they could get an 8” print without resizing and probably an 11” print with resizing.

      To a large degree, the Olympus executive was right: 12mp is a good stopping point for many. If you’re using all those pixels, you can easily print an 8x10” print, probably do quite fine to 11x14”. ..."



      Source: http://www.sansmirror.com/newsviews/how-many-megapixels.html?=

      Thom Hogan sums it up with his statement: "...I’d tend to say that 16mp was enough. But that won’t stop the floodgates of more megapixels...."


      Now lets look at another often quoted website on the internet. DXOmarks. DXO is known for its RAW converter. But they also do have a website, on which they test sensor performance of almost all camera bodies and lenses in combination with those lenses. Unfortunately they only test Bayer sensors. So no Fuji X-Trans sensor test nor Sigma Foveon tests. But in the case of Fuji, you can look at the Nikon APS-C sensor results. Do expect the same results in DR etc,. since Fuji and Nikon in that generation both use a Sony sensor at the underlying and only put some modifications on it, which should not result in meaningful differences regarding noise, DR etc.

      This site is kind of interesting, not because of their final results, but because no matter which model you look at, you can compare them with each other on the same site and make a judgement about how much they differ to each other. So not the absolute results of DXOmark tests are important for me, but the relative difference between the diffeernt models tested with the same standard test method.

      Here is a screenhsot with 3 different cameras side by side. You can only compare 3 models at the same time side by side on that site, so I will add another two more below to have all numbers together to understand how this can be used for the best insight.

      First 3 different sensor formats. As a good example of the current status quo of the sensor technology in 2016 I use here a Nikon 1 J5 (20 MP 1 inch sensor, no AA filter, released in 2015), Panasonic Lumix GX80/85 (16 MP MFT sensor, no AA filter, released in 2016) and a Nikon D7200 (24MP APS-C sensor, no AA filter, released in 2016):


      dxomarks_comparison_sensorsize_J5_GX80_d7200.jpg



      and here the well know Sony RX100 Version 5 (20MP 1 inch sensor, no AA filter, released in 2016), Nikon D750 (24MP fullframe with AA filter, released in 2015) and Nikon D800 (36MP fullframe with AA filter, released in 2012):

      dxomarks_comparison_sensorsize_rx100V_d750_d800.jpg


      So we have now something to compare to objectively. But do not look at the "overall score", focus more on the marked numbers in red (color depth) and orange (DR) circles.

      There you get the information you want to know. You can dig even deeper at dxomarks, but this is beyond the purpose here. What is important is not just the absolute difference between the different sub-categories, but also whether you are able to see these differences in real life shootings and in prints. What does a DR-number of 12.6 EV mean vs. 14.6 EV? Is this good or bad? Will you be able to see it and if yes, when exactly?

      DXO claims that for the measured category DR (orange), every camera with a number above 12 EV is "excellent". Differences below 0.5 EV usually not noticeable!

      For color depth (red) they say every camera above 22 bits is "excellent". Differences below 1 Bit are barely noticable!

      No matter which sensor size, we are always in the "excellent" terretory. Even the 1 inch sensor of the Nikon J5 which came out in 2015! Bear this in mind if you discuss any kind of differences.

      I do not say that you should not aim for more and better. I just say we are talking here about differences on a very very high level. You can therefore relax. You need to make sure that you have the right lenses and shooting discipline to exploit the full potential of the cameras of 2016.

      Full explanation of how DXO is testing:

      https://www.dxomark.com/dxomark-camera-sensor-testing-protocol-and-scores/?=

      Quotes:

      "...The higher the color sensitivity, the more color nuances can be distinguished. As with dynamic range, color sensitivity is greatest when ISO speed is minimal, and tends to decrease rapidly with rising ISO settings. In DxOMark testing we measure only the maximum color sensitivity. A color sensitivity of 22bits is excellent, and differences below 1 bit are barely noticeable..."

      and

      "...Dynamic range corresponds to the ratio between the highest brightness a camera can capture (saturation) and the lowest brightness it can capture (which is typically when noise becomes more important than the signal — that is, a signal-to-noise ratio below 0 dB). A value of 12 EV is excellent, with differences below 0.5 EV usually not noticeable. Dynamic range is an open scale...."
       
    2. dirk

      dirk CI-Founder

      Top Poster Of Month

      Joined:
      May 15, 2002
      Messages:
      1,554
      Likes Received:
      9
      Camera brand:
      many/ see my profile
      City/ Country:
      Germany
      To show how the values of dynamic range, tonal range and color sensitivity change with different ISO values, look at these charts. I picked the ones for prints, not for screen.

      dxomarks_comparison_sensorsize_J5_GX80_d7200_printDR.jpg

      dxomarks_comparison_sensorsize_J5_GX80_d7200_printTR.jpg


      dxomarks_comparison_sensorsize_J5_GX80_d7200_printCS.jpg
       
    3. dtepper

      dtepper New Member

      Joined:
      May 17, 2011
      Messages:
      3
      Likes Received:
      0
      I am not a luddite. I embrace technology. With that being said, I do not have the most current model digital camera models. I have a Hasselblad H4D40 system, and a Nikon D800 system as well.
      I use the Hasselblad for studio work (portraits mainly) and for light travel (car/walking). The Nikon system is for international travel. Both systems are excellent. I am a platinum/palladium printer as well. I do make large inkjet prints as well. I have made 60 inch prints with the Hasselblad, that are gorgeous. I mainly print in the 20 to 24 inch and both do a fine job. However, I do see a difference between my d800 and the H4D40. The image files from medium format systems have true 16 bit color, and have the other benefits as well: Depth of field, etc. My medium format sensor is CCD based, and I don't use it past ISO 800, so my experience might not be as relevant as others who more up to date kits.

      We live in an exciting age, with so many excellent choices.

      I hope this helps.
      Best,
      Dave
       
    4. dirk

      dirk CI-Founder

      Top Poster Of Month

      Joined:
      May 15, 2002
      Messages:
      1,554
      Likes Received:
      9
      Camera brand:
      many/ see my profile
      City/ Country:
      Germany
      Hi Dave,

      of course there is a visible difference between MF and a D800 and between a D800 and an APS-C Kamera etc.. Same as in the old analogue time with film. The same film delivered better prints from a 6x6 MF camera than from a 24x36 SLR. The point I wanted to make is that since around 2012, there is no "garbage" anymore on the market. It is like everybody has a Velvia or Provia slide film now in his hands. Image quality is really good nowadays.

      But of course the gap between each sensor size stays always the same because the sensor improvemenst are for all sensor sizes available within the same sensor generation. So everybody improves from one generation to the other.
       
    5. dtepper

      dtepper New Member

      Joined:
      May 17, 2011
      Messages:
      3
      Likes Received:
      0

      And I agree with you. There may be differences between sensor sizes, but does it really matter? Image quality from all these newer sensors and apps, is amazing. We live in exciting times. Moore's law states that processors (and by default, computers and cameras) double in power every 18 months (I know I am oversimplifying his law). And I know that I am restating your original question.....does size really matter? I'm not sure anymore.....
       
    6. dirk

      dirk CI-Founder

      Top Poster Of Month

      Joined:
      May 15, 2002
      Messages:
      1,554
      Likes Received:
      9
      Camera brand:
      many/ see my profile
      City/ Country:
      Germany
      That depends on the user, his experience in photography and his shooting style. An expereinced user will see immediately the difference of i.e. of a Medium Format camera with its wider dynamic range, smoother background blurring, creamier bokeh, more "clarity" etc. of those "fatter" pixels and more pixels in general. Same goes if you look at images between i.e. a Nikon D7100 and a Nikon D800. Same generation, but a totally different "atmosphere" on the D800 picture.

      Still the D7100 image is excellent. It has just another "juice" than one taken with a D800 or Hasselblad X1D. A trained eye sees this and enjoys it. As always, the last 20% in image quality gain is the most expensive. whether it is worth the premium price, everybody has to decide for himself.

      Surprisingly I find the difference/ the gap between 1 inch sensors (Nikon J5 in my case), MFT sensors (Oly/Pana) and APS-C sensor (Nikon, Fuji) not as big as the difference/gap APS-C vs fullframe vs. Medium Format.

      I sold my D800, although I loved the results. But with travelling around and 2 kids, it is not a pleasure to schlepp around. I have at the moment only Fuji-X, MFT and Nikon 1 left (and a few P&S). My plan is to simplify everything until the end of this year.

      My goal is to have only one small system for travel (either MFT or Fuji X or a future Nikon APS-C mirrorless) and on top of that small system maybe one medium format camera with only 1 or maximum 2 lenses in the long run when it is not that expensive anymore.

      Of course a high-end P&S with a fixed lens will always be with me :)

      Thanks to this, I will have a system for the "good enough" level. Easy to carry around. Small & light lenses. For the daily stuff with the family and memories. :z04_photos:

      For the "special moments", when I have time for me and without deadlines I then can have the "pure photography" like HCB but with a medium format camera or a high-end fixed lens camera. And then I will enjoy this step-up in image quality with a larger sensor... :z04_sabber:
       
    7. dtepper

      dtepper New Member

      Joined:
      May 17, 2011
      Messages:
      3
      Likes Received:
      0
      I completely agree and understand.

      I love my H4D40, and my D800. But a nice small camera that wouldn't be too heavy, full frame and a just a few lenses would suit me just fine. I tried EVF on Sony A7, and by the time I added fast lens and battery grip the weight was almost (not quite) as much as D800.
      My wife has a PS for family moments. I have been looking at high end PS, but haven't taken the plunge yet. If I just wait long enough, and technology advances just a little more.....:D
       
    8. dirk

      dirk CI-Founder

      Top Poster Of Month

      Joined:
      May 15, 2002
      Messages:
      1,554
      Likes Received:
      9
      Camera brand:
      many/ see my profile
      City/ Country:
      Germany
      this waiting will never end like with computers... :z04_Flucht:

      if nothing special will be announced within the next 1-2 months, you should look closer at the Ricoh GRV MK2, RX100 and Canon G7X MK2
       
    9. bytesmiths

      bytesmiths New Member

      Joined:
      Jan 14, 2008
      Messages:
      3
      Likes Received:
      0
      Another thing to consider is the modern miracle of In-Body Image Stabilization (IBIS).

      One reason to desire low-noise, high-ISO shooting is to keep the shutter speed up in order to shoot hand-held.

      If IBIS lets you shoot two stops slower, it negates the advantage of low noise at high ISO, since a lower ISO can be used.

      The physics of IBIS work against full-frame sensors. It takes four times the energy to rapidly move a sensor that is twice as big, which means larger batteries and lower battery life. This is why industry-leading IBIS is only found on smaller sensors.

      A few other innovations make the "FF advantage" less so. Olympus's† "Live View" allows a low-noise exposure to be built up from multiple exposures while you watch. The image data stacks geometrically, while the noise averages out, so the result is a long-exposure, low-noise image. Another nice innovation is Olympus's† "HD" mode, whereby the image sensor is shifted 1/2 pixel for each of eight shots, resulting in an 80 MP image from a 20 MP sensor. (It does require a tripod, though.) And the electronic viewfinder (now appearing on some FF cameras) has its own set of wonders — as well as consequences.

      I'm not saying there's no use for FF anymore, but for those in the middle of the bell curve, the need for FF is vastly diminished these days. I'm happy as a clam with my OMD E1.2, like I haven't been since I started with tiny, sturdy OM film cameras back in the 70s.

      †Other small-sensor camera makers may have similar things. I am most familiar with only Olympus.
       
    10. dirk

      dirk CI-Founder

      Top Poster Of Month

      Joined:
      May 15, 2002
      Messages:
      1,554
      Likes Received:
      9
      Camera brand:
      many/ see my profile
      City/ Country:
      Germany
      Although I agree to most of the points, IBIS only helps, if you have non-moving subjects.

      But there are also other advantages of bigger sensors. The bigger the sensor is the better the image is looking with the same amount of pixel. No matter at which ISO. I would love to have a Hasselblad X1Dmini with only 16MP or 24MP but the same sensor size than the X1D with its 50MP.

      DOF will be in the future no differentiation anymore between sensor sizes IMHO. If you look at what smartphones can already achieve with background blurring only with software (Google Pixel 2 for example), you realize that it is only a question of time, when this will be ready and even better implemented for real cameras.
       
    11. ilkka

      ilkka New Member

      Joined:
      Aug 8, 2004
      Messages:
      12
      Likes Received:
      0
      Size matters ore not, when i started whit Nikon full frame cameras i have not looked back. So maybe the size matters, never given any thought about it anyway. I usually change camera body in two ore three years because they get lot of beating from my work as freelance. I usually buy the latest model because they are sturdy and fast. My work is most press and feature!
       
    12. David Sw

      David Sw New Member

      Joined:
      Oct 14, 2010
      Messages:
      16
      Likes Received:
      0
      City/ Country:
      Coquitlam, Canada
      ...coal mine... Cole would agree.
       
    13. foveonfan

      CI-Club+

      Joined:
      May 25, 2007
      Messages:
      322
      Likes Received:
      0
      Hi!

      I have just bought a new Canon 6D. Could not resist the bargain price resulting from it being superseded by the 6DII. Before this, Idid all my work with Sigma's Foveon tech. My current Sigma SD1 is great for my landscapes. In other words, anything that does not move. But since gaining an interest in bird photography, the SD1 began to show its weakness, slow and inaccurate AF. Manual focus is not one of my strong skills anymore as my eyes get older and older. This meant many lost opportunities with the quick as lightening little creatures I so sorely wished to hang on my wall.

      Enter the 6D. Careful use of the central or spot AF mode has seen my lost opportunities vanish.

      But that is not what this thread is about. We're really discussing whether the size of sensors matters. My experience tells me yes and no. Many factors to consider. How big do you wish to print, etc. My SD1 might have a APS-C sized sensor, but boy can that sensor render detail, with only 15 mp spatial locations and I've many AO+ sized prints to prove it. I'm confident that there will be many landscapes opportunities in my future that will find me preferring the SD1 out of my pack instead of the 6D.

      But after all is said and done, your personal taste and choice may vary from that and I respect that.

      Sincerest regards, Jim Roelofs
       
    14. dirk

      dirk CI-Founder

      Top Poster Of Month

      Joined:
      May 15, 2002
      Messages:
      1,554
      Likes Received:
      9
      Camera brand:
      many/ see my profile
      City/ Country:
      Germany
      I do think the Foveon sensor is very special in this regard and not comparable with other sensor independencies between sensor sizes. The Sigma Merrill sensor is competing with the Nikon D800 sensor at basic ISO settings although it is only APS-C size and even then my D800 could often not match the clarity and 3-D effect of the Merrill images. So this would be unfair to compare ;)
       
    15. foveonfan

      CI-Club+

      Joined:
      May 25, 2007
      Messages:
      322
      Likes Received:
      0
      It's not until you've witnessed some large prints (AO+) that you begin to appreciate what this supposedly lightweight APS-C sized sensor in the SD1-M is capable of. It's not possible to post this image (even at 100dpi) here. The best I can offer is the address at my portfolio site. 162621766.IOwVwZG7.jpg Select "original" viewing size and you'll begin to see what I mean.

      Sincerest regards, Jim Roelofs

      I sincerely hope this is not inappropriate, Dirk, as I have not intended to offend you in any way.
       
    16. dirk

      dirk CI-Founder

      Top Poster Of Month

      Joined:
      May 15, 2002
      Messages:
      1,554
      Likes Received:
      9
      Camera brand:
      many/ see my profile
      City/ Country:
      Germany
      Hi,

      do not worry, there is nothing inappropriate in your posting. We do have nothing against posting a link to a supersized image. But most users do want to see additionally the image in a nomal viewing size within that posting to know beforehand what kind of image they can expect. you can not discuss images, if you do not see them in the thread themselves.

      I know that many Sigma users started with the introduction of the Merrill Foveon sensor generation to argue, that the benefits of Merrill sesnor can only be seen at full resolution. I strongly disagree. The same aruement can be made with Nikon D800 files or Hasselblad medium format files. Of course it is always better, to see it as big as possible. THis is why you should add a link to the original file.

      But at 1000 pixel on the longest side (not dpi), an image is big enough to see the advantages in a thread. I can see it even at 600 pixel, if the postprocessing is done correctly.

      Just try it out. Upload a Foveon image with 1000pixel on the longest side in the "having fun with Sigma" thread and put under it a link to the original file.

      The image you linked in the posting above, is only 100x160 pixel by the way. So way too small...
       
    17. afshalders

      afshalders New Member

      Joined:
      Jun 1, 2011
      Messages:
      26
      Likes Received:
      3
      Camera brand:
      Many
      City/ Country:
      Rio de Janeiro/Brazil
      Again and again, it depends if we're talking about prints or screen. The human eye sight limit is something like 300 dpi viewed from a distance of 15-20 cm. Any resolution beyond this and you're wasting information, and this is a well proven fact. For larger prints, we need lower dpi. Nobody looks at a 30x60 print from 15cm unless for pixel peeping.

      There are many calculators that helps to choose the proper resolution x size x distance.

      To be honest, if I ever need an A0 size high quality print, I'll do something far more logical, at leat to me: Use a medium format camera and do a high quality scan. No digital camera beats a 6x6 or 6x9 low grain film scanned at 3600+ dpi

      Talking about Sigma, I'm using their cameras since the advent of the SD9. Still have a SD10, SD15,DP2, DP2-M and SD-1M and Foveon images can be easily upscaled to 150% with very little loss in image quality. Just choose the right algorithm like Lanczos with an 8x8 matrix or Mitchell-Netravali. The lack of color alias is one of the reasons that makes this possible. Even at 200% enlargements the results are better than with Bayer or X-Trans sensors.

      Higher sensor resolutions like 20-30 megapixels sounds good on paper BUT sometimes you get a bad deal, for example in the form of noise or optics limitation due astigmatism and diffraction. Some lenses I have are excellent on my SD14 and really sucks with the SD1.

      And, of course, the post processing makes a huge difference, specially if using raw. I'll not dig deep into this subject. =)

      But the best way is to check by ourselves the final result using our eyes. If you like it, don't care about what other people are saying.
       

Share This Page