High ISO tests with the Sony RX-100 II

The new version of Sony's RX100 camera, the II or M2 depending upon which Sony literature you are reading, boasts an increased ISO range up to 12,800 from 6400 in the prior model. That is a lot to expect from a camera with a 1" sensor. So naturally I wanted to see for myself just how high I could take it and still satisfy my own needs and standards.

The RX100 II set to an ISO of 12,800 took a portrait of itself in the reverse mirror of a medicine cabinet by using its 10 second delay. The grain effect is noise that was greatly exaggerated by cropping and then using the extreme contrast of a curves layer in Photoshop.

Shooting at a high ISO levels usually coincides with lighting circumstances having very high contrast and deep shadows with little or no detail, and this tends to exaggerate the effects of noise even more than an evenly lit scene would. The photo below was taken in just such a poorly lit scene, a dimly lit bar with mixed and contrasty light. The RX100 was set to ISO 3200. You can download a full res version of the photo by clicking here.  The dark area of the beers were in the deepest shadows and were brightened up in post-processing. Doing this is essentially the same as having shot that area at a higher ISO to begin with, and consequently causes much more exaggerated noise to occur there.

For the purposes of this demonstration I did not use any noise reduction in post-processing. Had I done so, I could have reduced the noise to acceptable levels in photos shot at ISO 3200 and below. Above that level, artfacting occurs that is difficult to correct without loss of detail.

The area in the dark part of the beers show excessive noise when opened up in post-processing. The High Res version of this file can be downloaded by clicking here.

The test photos below were taken at ISO's ranging from 400 to 12,800. I did not feel the need to go below ISO 400 because no noise was apparent below that level. In fact, using the camera in an ISO range of 100-400 produces the highest quality with no visible noise needing correction. Once you hit ISO 800 and beyond, however, noise begins to creep in and will require some reduction when using the image at larger sizes.

Also note that in the situation below, although I used ISO's up to 12,800, this scene did not demand it because it had much more ambient light available than the bar scene above. Nonetheless, downloading the hi res version of the various sample images will provide a good example of what you can expect noise to look like at the various ISO setting of this camera.

The ISO test was done from ISO 400-12,800 of the scene above. Download the hi res samples by clicking on the links below.

Download hi res test images here:

RX-100 II - ISO 400
RX-100 II - ISO 800
RX-100 II - ISO 1600
RX-100 II - ISO 3200
RX-100 II - ISO 6400
RX-100 II - ISO 12,800

Conclusion:

If we compare the ISO levels of the RX100 II to those of a high end full frame camera, it isn't going to look so good. After all, cramming 20.7mp onto a 1" sensor has to have some repercussions in noise levels. But, if we take it for what it is, a small, portable camera, convenient enough to always have with us to grab shots we would otherwise miss, it does an admirable job. And a bit of extra post-processing to deal with noise is a small price to pay for getting what would otherwise be a missed opportunity.

Using the RX100 II up to an ISO level of 1600 should not present any real noise problems that are not easily fixable in post-processing. At ISO 3200 and above some artifacting creeps in that makes the images more problematic, but not impossible, to deal with. Ultimately, the shooting circumstances will determine just how bad the final noise will be. It could become very difficult to correct in dimly lit, contrasty scenes with ISO levels of 3200 and beyond. As a result of my own tests, I will try to keep the maximum ISO level of the RX100 II topping out at 1600 whenever possible. This will be helped by using the f/1.8 aperture the camera has at its widest zoom range.