ISO impact on sensor performance

Each of us has our own criteria we use when selecting which camera to buy. Sensor performance, handling, ergonomics, auto-focus, lens selection and other factors can all play a part. Many photographers are very interested in the sensor performance of their cameras, and this short article discusses ISO impact on sensor performance.

NOTE: Click on images to enlarge.

While it is impractical to provide sensor performance details on a wide selection of cameras, this article references DxO test data for a small number of cameras to reinforce a critical fact. Shooting at higher ISO’s has a significant, negative impact on the dynamic range and colour depth performance of a digital camera.

To start, let’s have a look at dynamic range.

dynamic range comparision

The above chart highlights DxO dynamic range test scores for a few different cameras including the Nikon 1 J5 and V2, Panasonic GH4, Olympus E-M1, Nikon D7200, Canon 760D, Sony Alpha 77II, and the Nikon D810. I chose these cameras for a few reasons. The Nikon 1 J5 and V2 are both cameras that I own and use. They also represent examples of different sensor technologies. Each of the other cameras in the chart, whether they have M4/3 sensors (GH4, EM-1), APS-C cropped sensors (D7200, 760D), or full frame (Alpha 77 II, D810), has the highest dynamic range score in its category and/or in its brand based on sensor size.

You’ll notice that as ISO is increased every camera suffers from a loss of dynamic range. DxO suggests that a difference of 0.5EV in dynamic range will be noticeable for most people.

As you compare the sensor performance of various cameras noted in the chart some potentially surprising facts will begin to emerge. For example, the Nikon D810 and D7200 are the dynamic range leaders at base ISO. When shot at even ISO-400 their respective dynamic range performances drop quickly. At ISO-400 they become basically equal to M4/3 cameras like the GH4 and E-M1 when those cameras are shot at ISO-200.

Another interesting comparison is between the J5 and 760D. When shot up to ISO-400 there is no discernible difference in dynamic range, even though there is a significant one in terms of sensor size between the two cameras.

It may be a bit shocking to many folks when examining the data to discover that shooting a Nikon 1 J5 at ISO-200 produces dynamic range performance equal to a D810 when shot at ISO-800.

Now, let’s have a look at the percentage of dynamic range loss when various cameras are shot at common ISO’s such as ISO-400, ISO-800 and ISO-1600.

percent loss of dynamic range

As you can see in the table above, even moving to ISO-400 causes a D7200, Alpha 77 II and D810 to lose over 10% of their dynamic range. Push that further to ISO-1600 and these same cameras lose 25% or more of their dynamic range.

What is clear is that when we are ‘sloppy shooters’ and use higher ISO’s than necessary it will have a significant, negative impact on the dynamic range performance of the sensor in our cameras and resulting images.

Let’s examine colour depth.

colour depth comparison

All cameras lose colour depth as they are shot at higher ISO’s. DxO suggests that a difference of 1.0-bits in colour depth is needed to be noticeable for most people. It is interesting to note that the colour depth of a D810 when shot at ISO-400 is similar to that of an E-M1 when shot at ISO-200. A Nikon 1 J5 shot at ISO-200 is similar to that of a D810 when shot at ISO-800.

percent loss of colour depth

As we can see in the chart above, all cameras lose a significant percentage of their colour depth as ISO’s are increased from base ISO. This reinforces the notion that regardless of the cameras each of us own, shooting at the lowest possible ISO (given subject matter and allowable lighting) will produce the best quality images.

Practical considerations.
There are some practical considerations that emerge from examining these data for dynamic range and colour depth. First is the importance of shooting cameras at their base ISO whenever possible to maximize sensor performance. The second is the importance of using a monopod or tripod whenever necessary to help lower the ISO at which an image is captured. And, the third is how critical it is to have good hand-holding technique as this allows a photographer to use slower shutter speeds and potentially lower ISO’s to attain their desired exposure.

Owners of cameras with larger sensors like APS-C and full frame can sometimes have a rather condescending attitude towards owners of smaller format cameras like M4/3, and crow about how much better the sensors in their cameras perform. The caveat of course is whether these same larger sensor camera owners are actually using their gear in a way to maximize sensor performance by always using the lowest ISO’s possible.

Using smaller, lighter gear can result in it being much easier to capture images hand-held at slow shutter speeds when compared to using heavier gear. Based on the weight variances between camera set-ups differences in shutter speeds of 1 or 2 stops are certainly possible. This could mean using lower ISO’s and better dynamic range and colour depth performance.

As we’ve seen in the data above, that difference in ISO could be a great equalizer in terms of actual dynamic range and colour depth sensor performance achieved when shooting with smaller sensor cameras compared to their larger sensor brethren.

My intent is to keep this photography blog advertising free. If you enjoyed this article and/or my website and would like to make a modest $10 donation through PayPal to support my work it would be most appreciated. You can use the Donate button below. Larger donations can be made to tom@tomstirr.com through PayPal.

You can also support my efforts when you purchase anything from B&H by using the Thomas Stirr affiliate link. Even the smallest purchases will help support this web site.

As a reminder to our Canadian readers, you can get a special 5% discount when ordering Tamron or Rokinon lenses and other products directly from the Amplis Store by using promotion code AMPLIS52018TS.

Article and all images are Copyright Thomas Stirr. All rights reserved. No use, duplication or adaptation of any kind is allowed without written consent. While we do allow some pre-authorized links to our site from folks like Nikon Canada and Mirrorlessons.com, if you see this article reproduced anywhere else it is an unauthorized and illegal use.

8 thoughts on “ISO impact on sensor performance”

  1. Hi Thomas,

    I’m a 1″ sensor fan too – tho, I’m in the Sony camp – RX10 ii.

    I’m occasionally enticed to consider a full-frame in the A7 series, but when I pick one up, with lens attached, I realise what a great package I have in the more compact, and much lighter RX10ii body (constant f2.8, 24-200mm FFE – which is plenty of range for my needs).

    Main reason for writing is to say what a clever approach you took to comparing the DxO Dynamic Range data – as % from BaseISO … Thanks for that tip !

    Also note you’re a fellow DxO OpticsPro/PhotoLab + NikCollection user … So, I’m sure you’ll be as impressed as I am in DxO’s implementation of U-Point technology in new PhotoLab. I don’t think it’s quite there yet (c/w Nik ColorPro) – but, knowing DxO, they’ll soon get that sorted out.

    Regards, John M – Melbourne, Australia

    PS. Will follow you – Much of your 1 Nikon commentary translates well to my experience (tho, I don’t understand how you’re working without an EVF).

    1. Hi John,

      Thanks for adding to the discussion with your comment! As far as DxO PhotoLab goes, I haven’t had time to do much with it yet. I make it a habit to update DxO OpticsPro when new versions come out just to make sure that my main RAW processor is as current as possible. Other than using the Spot Weighted DxO Smart Lighting tool I very seldom do any spot adjustments with my images so I doubt that the U-Point technology will make that much difference to how I process my files.

      I do use a Nikon V3 for birds-in-flight and having an EVF is very important for that type of photography. Other than that I use a J5 for all of my other photography. I very quickly got used to using the rear screen to compose images and it really is a non-issue now.

      Tom

    2. Hi Tom, A bit more on this topic …

      With cameras such as your 1 Nikons and my Sony RX10ii (both being ISO-Invariant – they even map very similarly on DxOmark!) there’s definite benefit in keeping ISO low as possible – even to the point of “shooting dark” and brightening in post-processing (instead of raising ISO – as Canon users need to do).

      Something else I’ve recently learned wrt impact of ISO on Dynamic Range is that, as always with photography, there’s a compromise involved: A bloke called Bill Claff has a site where he publishes noise and dynamic-range numbers by camera model.

      One of his charts shows what he calls “Photographic Dynamic Range Shadow Improvement vs ISO setting”: http://photonstophotos.net/Charts/PDR_Shadow.htm … wherein, again, your 1 Nikon J5 maps almost identically to my RX10ii.

      As I understand it, this chart is an indication of degree of improvement of detail retention in shadows as ISO is increased … BUT, one needs to also keep in mind overall Dynamic Range vs ISO: http://photonstophotos.net/Charts/PDR.htm.

      Looking at these charts *in combination* tells us (for our cameras): We’re losing about 1EV of Dynamic Range with each doubling of ISO (100 to 200, 200 to 400, etc) … BUT, on the other hand, there’s some small improvement in shadow detail as ISO is increased!

      I’m taking this to mean I can happily allow my ISO to increase to 1600 (AutoISO Range on RX10ii facilitates this) … Your J5 shows benefit even further, up to 3200. Beyond 1600 for me, I’m happy to brighten in DxO-OP/PL.

      Regards, John M – Melbourne, Australia

      1. Hi John,
        You are much more technically inclined than I am when it comes to photography, and I commend you for it! I am much more experiential and experimental in my approach. Since I’ve been using DxO OpticsPro/PhotoLab Prime I never hesitate to shoot my Nikon 1 gear up to a camera setting of ISO-3200. As indicated in DxO published testing, there is difference between manufacturer stated ISO and measured ISO. I’m not sure if the ISO ratings mentioned in your comment are manufacturer stated, or measured ISO. I originally stumbled onto these ISO differences when shooting with various Nikon 1 camera models, finding about a 2/5 of a stop difference between the V2 and the V3 or J5 (i.e. the V2’s manufacturer stated ISO and measured ISO were much closer than either the V3 or J5). I then went to the DxO site to check it out further and discovered that their test data also showed a difference of roughly that amount.
        Tom

        1. Mmm – I might be technically inclined – but that doesn’t make me technically competent 😉 … I’m still learning !

          Yes, difference between manufacturer claimed ISO and measured ISO is common – and claim is consistently less than actual. Apparently (I have read), there’s some leeway in the ISO standard that camera manufacturers all take advantage of (none wanting to give any advantage to the other !)

          I see why that’s important when swapping same lens between different bodies – but, in consideration of a specific camera in isolation it’s not really an issue – there’s no harm in accepting each ISO setting for what it’s claimed to be.

          Bill Claff’s charts take this into account: He charts the various measurements at claimed ISO values (100, 200, 400, etc) but includes the actual ISO, as measured, in a “pop-up” when cursor hovers over the graph point.

          He notes usage of DxOmark for a lot of his base data,

          Well worth taking a look … (No affiliation with me).

          Regards, John M

          1. Hi John,

            I agree that differences between manufacturer stated and measured ISO is a non-issue in most cases, especially when shooting with one camera in isolation. I shoot all of my client videos using manual settings including ISO. Under this type of situation differences in measured ISO between camera bodies can be noticeable and cause more work in post if exposure correction is needed. That’s one of the reasons why I use three Nikon 1 V2s to shoot client videos and do my best not to mix in other camera bodies with noticeably different measured ISO performance if I can help it. One of the good things with the V2 is that it was pretty much identical to the measured ISO performance of a D800, which at one time was my primary video camera. This allowed me to capture video clips of the same scene from different angles with the two cameras, then mix the clips into a video assembly without the need to do exposure correction, saving me time in post.

            Some folks love to brag about the low light performance of their cameras, and in some cases they have no idea at which ISO they are actually shooting. A good example is the Olympus OM-D E-M1 Mark II. At a manufacturer stated ISO of 1600, the camera is actually shooting at ISO-688, more than a full stop lower. A Nikon 1 V2 shot at manufacturer stated ISO-1600 is actually shooting at ISO-1236 and at manufacturer stated ISO-800 is actually shooting at ISO-619.

            Many folks look at low light performance when considering a camera purchase. So, to more accurately compare the noise between two cameras I believe one needs to understand at what ISO the cameras are actually shooting, then match up the closest measured ISO settings to get a more accurate assessment. In this case using files from an OM-D E-M1 Mark II shot at ISO-1600 would be best compared against a V2 shot at a manufacturer stated ISO-800. Don’t get me wrong…the Olympus would still win out, but the difference would likely not be as great as would be expected if identical manufacture stated ISOs were used for the noise comparison.

            Tom

  2. “Using smaller, lighter gear can result in it being much easier to capture images hand-held at slow shutter speeds when compared to using heavier gear. Based on the weight variances between camera set-ups differences in shutter speeds of 1 or 2 stops are certainly possible. This could mean using lower ISO’s and better dynamic range and colour depth performance.”

    Hate to say, but said with respect, no. Heavier cameras have more momentum and do not tend to be as easy to be “shakey” as smaller cameras. Also shooting high ISO has real world practical applications that force even a conscious photographer to shoot with it.

    I think more accurate to say would be – to be clear- that there are many out there that may not know better and shoot high ISO unnecessarily, but high ISO does have its place.

    Cameras with good built in IBIS like the Olympus OMD EM5MKII, the now new Panasonic GX85, or even the Full Frame Sony A7RII in some situations can sure make of the stop difference, though that doesn’t apply to moving subjects.

    1. Hi Ricardo,

      I agree with you that being able to shoot at slower shutter speeds does not apply to moving subjects. I can’t comment about bodies like the EM5 MKII or GX85 as I have never used them. I did shoot with a GH4 for a very limited time and found that I could not shoot that camera at the same slow shutter speeds compared to my Nikon 1 V2’s.

      I have heard other folks, DSLR owners primarily, talk about the momentum of heavier cameras and that they are easier to control. My experience may be atypical of course, but I have not found this to be the case from a practical standpoint at all. I find that it is far easier to control my Nikon 1 V2’s and shoot at slower shutter speeds than it was when I used to shoot with a D800 and FX lenses. This is especially true when using the 1 Nikon CX 70-300mm or the 1 Nikon 30-110mm with extension tubes when doing macro-type work hand-held. That’s one of the reasons that I had no problem selling my D800 and all of my FX glass about a year ago.

      Tom

Leave a Reply

Your email address will not be published. Required fields are marked *