Canon 5D Mkiii vs 5D Mkii High ISO Test

This is the first of a series of posts in which I will review the 5D Mkiii, which I very happily received earlier today. I have been looking forward to this new release by Canon for some time. Shooting mainly weddings I was really hoping for a body which improved on three key areas versus the 5D Mkii: low light / high ISO performance, focussing, and the bonus would be dual memory card slots. Well, the 5D Mkiii was announced a few weeks ago and the specifications seemed to fit like a glove. After the Nikon D800 announcement, I was praying that Canon did not follow suite with similar specification as the low light performance would be compromised and I for one have no need for that sort of pixel count. Regarding the dual card slots – I have never experienced card failure with my previous generation single-slot bodies, however having the peace of mind of an in-camera backup is very welcome and to be honest a bit of a surprise as although I was hoping for it I wasn’t really expecting it. I think I’ll set my camera up to write RAW files to the CF card and JPEGs to the SD. That way I can plug in a nice fat SD card and maybe only go through one or two cards on a typical wedding day.

Another great point about the Mkiii is that it used the same batteries as the Mkii (Lp-e6), of which I own four so thankfully there was no waiting for a battery to charge. I got stuck right in and the first impression was the superior look and feel of the Mkiii. Right now I also own the Mki and Mkii versions and the difference is distinct. The new button layout takes a little bit of getting used to and I won’t even go into the multitude of focussing options available. Very brief first impression of the focussing is that it is like a laser guided missile versus the old time “trebuchet” style focussing of the 5D Mkii. This is a huge improvement and for those people out there balking at the price of the 5D Mkiii, it is features like this that account for a lot of it, in my opinion.

Today I decided to to a comparison of high ISO performance versus the Mkii body. I want to stress that the crux of this test was to determine the performance of the RAW files alone and not in-camera JPEGs.

The test conditions and methodology were as follows:

  • Canon 85mm F1.8 lens was used on both bodies
  • Cameras mounted on a tripod
  • Mirror lockup and self-timer activated for each shutter release
  • Subject lit by tungsten light
  • Exposure approach: Manual. Aperture set to a constant F5.6 to ensure plenty of DOF. ISO from 400 to 25600 with corresponding increase in shutter speed for each ISO increment. Note: for the same aperture, shutter speed and ISO combination, the 5D Mkiii underexposed, versus the Mkii, by around 1/4 of a stop. This was compensated for by slowing down the shutter speed for each of the 5D Mkiii images, instead of adjusting in post processing.

After downloading the RAW files from each camera, they were opened up in Canon’s Digital Photo Professional. The only adjustments done to them are as follows:

  • 5D Mkii files: pulled down the noise reduction to zero (luminance and chrominance) and set sharpening to zero.
  • 5D Mkiii files: as above, set the noise reduction to zero. Here comes the first interesting observation – the Mkiii files are a bit softer than those of the Mkii so I set the sharpening level to “2″. If anything, this will degrade the observed noise but I decided to do this anyway. Also note that the softer images of the Mkiii sharpen up very nicely so I don;t see this as an issue at all.
  • All images were ten converted to JPEGs as maximum quality level.

Each image was then opened up in Photoshop CS5 and using an action I created, cropped a 100% magnification portion of each image. Each pair of images was then placed side by side, e.g. 5D Mkii ISO 400 versus 5D Mkiii ISO 400. The size of the crop is small enough so that there is absolutely no scaling of the images as the appear below on this blog post. These are purely 100% magnification views of each image taken for this test. Something I should have done was set a consistent white balance but I didn;t, so there is a slight colour shift between each set of images.

So, here we go, a series of images comparing the high (and not so high to begin with) ISO performance of the 5D Mkii to the 5D Mkiii under reasonable controlled conditions. All the Mkii images are on the left and the Mkiii images are on the right.

 

Now in the two images below, have a look at the detail breaking up at the lower edge of the red book in the centre of each 5D Mkii image. The Mkiii pulls away nicely here in retaining detail.

 

The final image below shows a comparison of a Mkii 3200 ISO image versus a Mkiii 6400 ISO image, in an attempt to gauge the level of improvement in handling of higher ISOs. To me these images look pretty similar with the noise rendering in different ways as evident in the above images as well….

 

I’ll have to view these with fresh eyes again, but my initial conclusion is that the “RAW file noise performance” of the 5D Mkiii is distinctly better than that of the Mkii. Not so sure about 2 stops been bandied about, certainly not in terms of the actual RAW files. To me the 5D Mkiii starts pulling away from as low as ISO 800 and certainly from ISO 1600. ISO 6400 on the Mkiii is maybe a stop better than ISO 3200 on the Mkii. Overall it looks promising and on the plenty of occasions I shot at ISO 3200 with the 5D Mkii at weddings, I can not only happily shoot at ISO 6400 with the Mkiii and expect pretty similar noise performance, but all lower ISO settings will produce cleaner images as well.

I am more than happy to take comments about the test setup or test methodology. Things which could maybe have been done better. This is just the way I tackled it today and I already see at least one improvement I could make to the test.

Now to get out there and do some real shooting!

If you enjoyed this post, please consider leaving a comment below. You can also like and share this post on Facebook, Tweet, Google +1, or email this post to a friend using the buttons directly below. It would be great if you could like our Facebook business page and follow us on Twitter:
BellissimaPhotoCanon 5D Mkiii vs 5D Mkii High ISO Test

Comments 17

  1. Yohan

    Thanks for doing and sharing this test. I think you’re a bit generous with the 1 stop difference. The comparison of 5DII @ 3200 vs. 5DIII @ 6400 is a win for the 5DII, which means the difference is less than a stop.

    The 2 stop difference was a claim straight from Canon but only in JPEG. In RAW, which is the real deal, it’s much less as you’ve shown.

    I assume more reviews are coming, including for the AF? That’s the big new feature for me.

  2. Mark A. Kathurima

    Many thanks for posting this. I am extremely pleased with my Mark II images at high ISO and am very happy to see the Mark III has taken it up yet another notch. Not sure I’d be in the market for one just at the moment but it is certainly a worthy addition to Canon’s already excellent stable.

  3. Post
    Author
    BellissimaPhoto

    Thanks Mark and Yohan. Yohan I tend to agree with you – I think it is a touch less than 1 stop improvement.

  4. Michael

    I can’t believe in the sharpness test … Maybe there is some other default NR filter which we can not turn off. Other thing is that some other test gave opposite results … ehhh
    Please write something more about AF.

    Thanks for everything
    M.

  5. Scott Martin

    Thanks for a great review and I’m looking foreword to your next installment. I’m still trying to decide whether the MK III performance is worth the price to upgrade.

  6. Federico

    I did the same test. And had the same result.
    I’m very disappointed about the mk III file that is softer than mk II.
    Even slightly better with noise, I prefer the mk II file. It shows just a clear look.
    I bought it because I needed second body and better autofocus but if you don’t need it, the mk II is a bargain.

  7. Post
    Author
    BellissimaPhoto

    Thanks for all your comments. The soft images are a function of Canon’s DPP. The default “high quality” setting softens the images considerably. Even low ISO Images. The alternative option is to use “high speed” which results in sharper images however this also disables the noise reduction sliders as well. Not good. I am pretty sure there will be a DPP update to address this issue.

    I have switched over to using my default ACR (only tried DPP briefly). This produces sharp images and of course has excellent noise reduction options (same engine as Lightroom).

  8. Bryan Attewell

    Thanks for taking and publishing.

    However, like some other respondents, I’m not seeing the improvements for the MK III. Maybe the slightly different exposure settings is affecting the results. It’s difficult to make comparisons when the images have such different tonality. Viewing on my laptop, whose screen is not the best, doesn’t help either!

    What interests me most is the shadow noise. This is a weak feature of the MKII and I would like to know if the new body brings any improvement. The highlight tone priority feature would be really useful if the shadows were not so noisy. Has Canon provided a fix?

  9. Nancy

    - Nice blog Paul keep up the good work! Yes the 5D MK iii is a stunning SLR and the video caiipilatbes are amazing. You just can’t beat the shallow DOF compared to what standard video cameras and their stock lenses can normally produce.

  10. Mike

    Gotta say I am disappointed about the softness of the MkIII, and the noise is worst than I would have hoped for.

    It would be good to see a comparison done with matched white balance, and also a variety of light levels. I photograph bands on stage, and have yet to find a DSLR that comes close to Fuji 800 Pro.

  11. Andy Holdsworth

    I know this is an old post, but regarding the difference in sharpness lies in the AA filter. It is slightly thicker on the mk3 than the mk 2. But I’m sure by now you already know this .
    Cheers,
    Andy

    1. Post
      Author
      BellissimaPhoto

      Thanks for your comments everyone. Andy I agree about the AA filter, but the biggest factor is Canon’s DPP which at the time resulted in very non-ideal images in terms of sharpness. Processing the same RAW files in Lightroom gave a better result.

Leave a Reply

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