How an S100 compares to a 600D at high iso? You're te judge! (Warning: jpeg only!)
I recently acquired a 600D, to replace (or not?) my 50D which developed an Error50 problem, and I am quite satisfied, but even though high ISO seems better than on my 50D, it is not as good as I might have thought at first. When I bought my S100 a couple of months ago, I was really surprised at the high iso jpeg output, regarding the tiny sensor size.
So I decided to make a quick high ISO comparison from both camera's JPEG output. I know, I know, people who shoot dslr shoot raw to get the best out of... I know. But many people also prefer shooting jpeg most of the time because it is much more convenient. I am one of those, even though I occasionally shoot raw, too.
Please note that the S100 is using a Digic 5 processor vs Digic 4 for the 600D.
Both cameras are on a tripod, with a 2sec self-timer. For the 600D, I used a 15-85 focused manually in live view. White balance was set to daylight, but for some reason, the colour balance between both cameras is quite different. The S100 is closer to reality. I'll have to check the WB settings for the 600D.
NR on the 600D is set to normal, and low on the S100.
The images from the 600D were resized to 4200*2800, roughly 12 megapixels, using bicubic sharper in photoshop.
The S100 images always on the left, 600D on the right.
I'll let you draw your own conclusions, but here is what I think:
- At low iso, there is not much difference, apart from the fact that the S100 cannot ouput a completely grain-free image, but there is no chroma noise at all. The Digic 5 is doing a great job.
- I thought a 2 stop comparison would be a good starting point. And indeed, IMO, the S100 images at 400ISO can compete with the 600D at 1600. But the interesting thing is that this trend changes as the ISO goes up. When comparing 800ISO VS 3200 on the 600D, you can clearly see more noise in the Eos image. But there is slightly more detail as well. So that's down to noise reduction.
- At higher iso settings, something interesting happens. The S100 at 1600 looks much better than the 600D at 6400 iso. There is slightly less detail in the pictures, but much much less noise as well. I prefer the S100's output by a mile.
-Then what about 3200 ISO vs 1600? Well, both camera are actually quite close. The S100 has a bit less chroma noise, but also loses some detail. But I think they are quite comparable.
That comes a bit as a surprise to me, regarding the difference in sensor size! And considering the S100 has an F/2.0 lens at wide angle, I can use a lower ISO than the 600D with its kit lens (f/3.5). That means that in low light, where I can shoot at f/2.0 with the S100 at, say, 800 ISO, I will get a very similar image quality to the 600D that has to be shot at 3200ISO to compensate for the slower lens. Quite a feat for such a small camera!
Of course, I can use fast lenses on the 600D, and I certainly do, but this small test is just about sensor low light performance.
I'm now dreaming of a new aps sensor which would have the same quantum efficiency as the S100 little CMOS. Imagine a APS-C 20-30 megapixels camera that would be able to shoot 12800 noiseless pictures!
Also, there are some processing differences, but the new 650D with its Digic 5 is VERY close in IQ to the 600D, so that's not the decisive factor.
PS: STUPID AUTOMATIC SPACING AND STUPID NEW FORUM ARE DRIVING ME MAAAAAD! It took me hours to find how to post pictures so that you can see the names and references. I had to upload every picture one by one in my gallery! Now if that's progress, let me get back to the good old things!
Also, make sure you view the original pictures in my gallery, because it seems that the new stupid forum can't handle full size pictures.