Image Sensor
Without a doubt, the most important feature of a digital camera today is its image sensor. You could put the most advanced autofocus and metering systems with a boatload of great features into a camera, but at the end of the day, they are all more or less secondary – the sensor performance is still looked at first. Things like resolution, dynamic range, diffraction, color depth, and ISO performance are all tightly related to the sensor and its physical size.
Canon claims the 5D Mark III to have a much better sensor with a two-stop improvement over the Canon 5D Mark II – its “native” ISO range has been expanded from ISO 6,400 to 25,600. Sounds like a huge difference, which should make images from the Canon 5D Mark III look much better in comparison, especially when high ISO images are down-sampled to a smaller resolution (see my “how to resize images in Lightroom” and “how to resize images in Photoshop” for an in-depth explanation of the down-sampling process).
The sensor went through some physical changes. Its physical dimensions stayed the same, but the resolution increased by over a megapixel – from 21.1 MP to 22.3 MP. So technically, this means more resolution while delivering better low-light performance. Does the new sensor deliver? How does it truly compare to the previous generation Canon 5D Mark II? See the “Camera Comparisons” page for more details.
Dynamic Range
Measuring the dynamic range of a camera is a complex process that requires a good, consistent methodology, especially when doing cross-brand comparisons. Since I have been shooting with both the Canon 5D Mark III and the Nikon D800 side by side, I can say that the D800 clearly has the lead in dynamic range. This difference was obvious when I shot the same scene with both cameras, at very similar camera settings. The Canon 5D Mark III consistently overexposed highlights, while the D800 rarely did (the exposure was similar on both). The dynamic range difference was even more obvious when post-processing images in Lightroom – I clearly had more options for recovering data on D800 images than I did with the 5D Mark III.
Because I do not have a way to actually measure dynamic range in numbers, I rely on DxOMark for sensor rankings – I find their results to be on par with what I can get when recovering images in Lightroom and Photoshop. Unfortunately, the Canon 5D Mark III does not seem to perform well according to DxOMark. They rated the Canon 5D Mark III 51st among all cameras, which puts it below many Nikon, Sony and Pentax cameras and sadly, below even many APS-C size sensors. Even the older Canon 5D Mark II yields a better dynamic range in comparison. Here is a screenshot from the Dynamic Range tab from DxOMark’s Camera Sensor Ratings page:
Some people take DxOMark measurements with a grain of salt and I agree – you cannot always trust one source for these types of measurements. In fact, the same applies to my reviews as well – my experience with a particular piece of equipment might differ substantially from another reviewer. However, after looking at many different images from both cameras, I do agree with DxOMark’s assessment. For example, take a look at the below two image crops from the 5D Mark III and D800:
The original images (RAW) were exposed the same on both cameras, both at base ISO of 100. After I imported them into Lightroom, I moved the “Shadows” slider all the way to 100 and then picked the darkest part of the image for the above crop. As you can see, the Canon 5D Mark III crop looks much noisier in comparison and retains fewer colors and details compared to the Nikon D800 crop. When pulling details from shadows, the Nikon D800 has a lot more information to work with.
Table of Contents