In the red corner: the Sony RX100m2, weighing in at 265 grams
In the blue corner: the Canon 5Dmk3 with 24-70mm f/2.8L lens (version 1), weighing in at 1975 grams
As I recounted earlier, I used a Sony RX100m2 large sensor compact exclusively on a recent trip to Istanbul. I was very happy with the image quality. So happy in fact, that on my return a nagging doubt set in: was the diminutive Sony actually better, at least in bright light conditions, than my “real” camera, a Canon 5Dmk3? Such heresy is not inconceivable, especially as my standard lens is a version 1 of the 24-70 f/2.8L. It’s a nice lens, but not one of Canon’s best performers (the version 2 is said to be substantially better), and my copy has seen some significant abuse over the years.
Testing cameras and lenses is a tricky business, and although I am a scientist what follows is not at all scientific. Rather, I shot a few frames from my office with parameters mimicking what I would consider to be my “normal” usage of these two very different tools. For the 5D3, I would hardly ever shoot wide open in broad daylight as the depth of field would be too shallow. So I set the camera to f/5.6 at 28mm at the base ISO 100. Depth of field is less of a concern for the Sony, with its small sensor, so I stopped down about one stop from wide open for an exposure of f/3.2 at 28mm at its base ISO of 160. Departing a little from the “real world” theme, both cameras were mounted to a tripod and three frames were shot using the self-timer to minimize vibration. Frame-to-frame variation was negligible. The RAW images were then minimally processed in Lightroom to set equivalent white balances, and 100% crops of the center and corner regions exported with “low” output sharpening only.
And the winner is… Well, as far as center sharpness goes, there’s really not much in it to my eye. The Canon image has slightly more contrast, but if one adjusted the Sony file to match that I suspect it would be hard to tell them apart in a real world situation. In the corner the Canon image is clearly sharper, though there’s also more than a bit of chromatic aberration that from now on I’ll try and remind myself to correct (quite possibly the Sony is correcting this silently in post, but if Sony and Lightroom are having a tete to tete and improving the image without telling me I have no complaints).
What about at higher ISO? To my mind, ISO 3200 and beyond is for situations that are pretty seriously dark, and unlikely to be encountered casually. If I go somewhere with the thought of taking pictures by moonlight, for example, I’d certainly pack my 5D3 (and a tripod!). ISO 1600, on the other hand, is a setting you might often need in an interior with not too much light – in fact I used it quite often with the Sony in churches and mosques in Istanbul. Here are 100% crops of frames shot later in the day at ISO 1600 (these are at about 1/30s, so long exposure noise considerations are not a factor here).
Obviously the Canon’s full frame sensor dominates over the Sony here. There’s markedly less noise, and better resolution.
Evidently I’m comparing apples with oranges here, so there’s no real conclusion to be drawn. It’s no surprise that the laws of physics have not been repealed, and that a state of the art full frame sensor returns a better high ISO image than a compact, even one with a relatively large sensor. If you have both cameras, as I do, it’s also somewhat reassuring that even under bright conditions Goliath has the edge over David. That said, the center sharpness of the Sony lens – which much of the time is the only thing that really matters – seems for practical purposes to be as good as my Canon L glass at typical shooting apertures. Given that the battle is not really David vs Goliath, but rather a situation where you have to carry either David or Goliath into battle, there’s clearly a use case for both.