I went up to the Kerry Park viewpoint after the Ben Folds concert to get some shots of the city. I thought it would be good chance to test out the S90’s low-light performance as well, with an easy comparison to the D90, which I had with me as well. With a tripod set up, I started with the D90 and a Micro-Nikkor 60mm AF-S and moved on to the Canon S90. I shot from ISO 200 through ISO 1600 in RAW with the D90. With the S90, I shot from ISO 80 through ISO 1600, in both RAW and JPEG. I figure the S90’s the type of camera that could be used by both serious amateurs, who don’t mind post processing, and the more casual user, who simply wants great photos straight out of camera.
For now, I’ll set up some quick samples – the full set of ISO 80 – 1600 will have to wait to when I have a bit more time to process everything. I only used the RAWs from the two cameras for this example, with the D90 images processed in Capture NX2 and the S90 images processed in Canon DPP. The default noise reduction settings were used (none in NX2 and 2 Luminance and 4 Chrominance for DPP). No sharpening was applied to either set of photos.
So, without further ado, the canonical Seattle skyline shot.
Canon S90 – ISO 80. Click on the thumbnail for a 1600px wide version, or click here for the original size. Use sparingly, and be kind on my bandwidth!
Canon S90 – ISO 200. Click on the thumbnail for a 1600px wide version, or click here for the original size. Use sparingly, and be kind on my bandwidth!
Nikon D90 – ISO 200. Click on the thumbnail for a 1600px wide version, or click here for the original size. Use sparingly, and be kind on my bandwidth!
And for a closeup (100% crop) –
Canon S90 – ISO 200
Nikon D90 – ISO 200
Clearly, the Nikon D90 is pulling out much better image quality here. That’s to be expected. You can also see the difference in dynamic range between a compact camera sensor and the larger APS-C of the Nikon D90. At what ISO does the D90 have comparable quality to the S90 at ISO 200? I found it to be somewhere above ISO 800. The following is a 100% crop of the Nikon D90 at ISO 800.
Nikon D90 – ISO 800
Compared to the Canon S90, there’s more detail here (probably due to the fact that some noise reduction was applied to the Canon S90’s RAW images in DPP), but noise levels are about the same. The D90’s ISO 1600 shot is significantly noisier than the S90’s ISO 200.
What can we take from this brief look at the differences in image quality from a large hulking DSLR kit compared to the Canon S90? There’s around a 2.5 stop (or slightly more) advantage for the Nikon D90 in this situation (I’m adding half a stop for the fact that some NR was applied to the S90 RAWs), but that’s also factoring in the 60mm AF-S lens used, which is a decently sharp macro. What I think is pretty interesting is this comment from a reader of my S90 review:
Your comparison with the D40 SLR is an interesting one because Iâ€™ve been using one for a [D40] couple years now, and I just ordered a S90 today (replacing my ancient SD200). Of course I donâ€™t expect the S90 to take better photos than the D40, but I think it will do better in low-light high-ISO situations than the D40 w/kit lens based on sample shots, and also:
– Lens has a 1.8 stop advantage over kit at wide end
– Being 3 years newer probably gains it ~1 stop (for example, compare D90 to D40) w/ better tech and software
– 1/6 area sensor size probably means ~2.5 stop worse
– Andrew S
The D90 is probably close to a stop better than the D40 in low-light, which means the D40 is probably around 2 stops better in low light than the S90. That’s fairly decent performance out of a compact camera, that weighs less than the 200g+ kit lens that comes with the D40. With the fast f/2.0 lens at the wide-end, as Andrew S remarks, the S90 could get pretty close to a D40 kit’s performance. I’d argue that although there’s still a good half stop or more difference there, it shows that in certain circumstances, you can get pretty close. Of course, this is with some pretty wonky math and some assumptions of performance on my part (I’ve long since parted with my D40, so I can’t do any scientific testing), but I think it’s within a pretty small margin of error. I’d be very interested in seeing what the image quality difference is between the S90 and an older DSLR.
Again, as I mentioned in the review, I imagine the majority of users of the S90 will end up posting their photos on the internet, with a much smaller subset doing any printing. At any reasonable web size, say 1600px wide, as I’ve done for the linked images at the beginning of the post, noise really doesn’t play a part at ISO 200. From my brief look at ISO 400, it should be absolutely no problem either at 1600px. In fact, with DPP’s default 3 Luminance and 6 Chrominance NR, there is not much visible noise in this same scene at ISO 400 at 1600px wide. Very impressive.
And remember, everything I showed here is a baseline. I pretty much used every default; there’s more image quality to be squeezed out of the S90. I did my tests in this manner to keep both the S90 and D90 on a level playing field, from a post-processing point of view.