Well, thanks to Lensrentals.com, I did a big test of all four cameras. It wasn’t super precise since I didn’t have the RX-100 with me, but this isn’t a technical review, but more of actual usage by a not very good photographer. I shot in raw and in jpg and used DxO to process for all but the Fuji (I had to use Lightroom as DxO doesn’t support the X100s). I used a bit more image sharpening in DxO and mainly the defaults in Lightroom (although its autotone really does wierd things). I used my judgement on exposure compensation (mainly what the camera wanted, but sometimes used their smart correction, usually resulting in a -0.5 EV correction) and they have a lighting thing called DxO lighting which I used.
The reviews have the Ricoh GR as a camera people love and the Fujifilm X100S as the one that has amazing image quality. All deliver a true 11MP or so (although the Sony is less at 10MP). They should deliver about the same low light performance with the APS-C sensor being one stop better, but the Sony have a one stop better lense (F/1.8 vs. F/2.8).
All of these cameras did a great job in lots of light and the photos were actually (surprisingly) quite identical, it was when it was a tougher shooting scenario. As an aside, it will take me a while to really get used to no more viewfinder. For whatever reason, it is easier to focus when you are looking in and concentrating.
Here are the results in Jay Leno order:
4. Fuji X100S. This is the coolest camera technically and I really wanted to be blown away by the image quality. I had to use Lightroom 5 to develop the images. And I would say that in camera JPEG as Ming Thein has pointed out are good, but you still get quite a bit more dynamic range with RAW. The main issue with the camera is that it is not image stabilized. I took quite a few photos in low light and this definitely makes it hard to take an photo at say 1/10. The camera is a little wierd to use in terms of not having a mode dial (it uses separate dials for everything else). The big issues with it are that it is big, larger than the dSLR and not pocketable and the image quality isn’t that much nicer (at least with my inexpert use of Lightroom).
As an aside, I also set the camera according to what the web said were the best defaults, so used Adobe RGB (and yes I have a color corrected monitor) and also DR200 which is supposed to pull more detail out. I didn’t do an A/B comparison, but that could affect some results.
I’m giving you the actual 5-6MB images, so this page is going to be slow to load, but here is a photo in good light (and no, it doesn’t win any awards, but was the only photo I took that didn’t have people etc in it):
Here is the low light photo taken as raw and then processed. It is pretty good I think and was sharp because I really concentrated. it is hard to hold a non-EVF camera well:
Here is the out of the camera JPEG, you can see it is quite a bit darker. I did some cropping on the other one so you will see it isn’t quite the same (sorry couldn’t help but be a little artistic):
3. Ricoh GR. This is the camera I really wanted to love. There is actually some sort of bug in the DxO processing of the GR (in version 8.3 I was running in trial mode). It for some reason read all the RAWs as much darker. Probably -1EV less. I fixed it by turning DxO smart lighting to Strong and this seemed to make it closer. The JPEG out of this camera are very good by the way. I though much closer than the X100s which is strange as that is not what the reviews were showing, but YMMV. Control-wise, this is a camera that you can definitely use with one hand and was quite easy to figure out. And having the 28mm F/2.8 fixed lense felt Ok most of the time, but you definitely would like a zoom sometimes as well. It is barely pocketable but is still maybe a third bigger than the RX100. The main issue is that again, it isn’t stabilized, so you have to be careful how slow you go.
Here is an image taken in good light that show what detail this camera can achieve. Amazing!
The RAW processed definitely shows more detail and it is about the same brightness as the out of the camera for low light.
This is the out of the camera, it is brighter and definitely not as sharp. I did turn up the Raw sharpening as other reviews have said the DxO images are soft. Went from -0.5 (whatever that means) to +0.5 (seemed symmetric:
2. Canon SL-1 (EOS-100D) with 17-55mm EF-S F/2.8 IS
This was probably the big surprise. This camera body is super small, but in this case, we tried a good lense on it. The DxOmark shows this to be about an 11MP camera with this lense. We also turned on highlight tone priority (buried in the custom menus) to try to get more dynamic range out of it. I’ve been using Canon’s for a long time and it is certainly nice to be back to the familiar dials and so forth. The viewfinder is definitely not a 5D2 size and it has the same auto focus system as my old Canon 350D it feels like, but it worked pretty well. What was a surprise is that image quality was very good. It certainly helps to have image stabilization and zoom as well. The only drawback is that the camera default is pretty red. I had to turn this down for portraits quite a lot. And having an effect 1.5x crop factor, the 17-55 becomes a very usable 22mm to 70mm lense.
I don’t have any photos I can share with you (eg of neutral subjects) during the low light portion unfortunately. But these are in daylight and are representative of the incredible picture quality of all of these cameras in good light. The main advantage for this camera is the ease of composing with a viewfinder but it is big. Something like this in micro-4/3 would be great or maybe given how the 5D3 and Magic Lantern is developing, I should try that next as the high mix camera.
1. Sony
This was probably the most surprising result. After all a small 1″ sensor shouldn’t be able to be the top choice compared with these monster APS-C sensors, but the Sony is a pretty amazing package. Theoretically it is about 10% less resolution (10MP effective vs. 11MP), but particularly in daylight you don’t really see this kind of difference. And as Scott has pointed out camera shake is a huge determinant of resolution. Which might be why the two stabilized cameras are top choices, I guess. The camera is just about impossible to figure out with the Sony interface, so you have to redo all the keys and buttons to make it work. It certainly has a long learning curve. The big advantages though are the size. It is really a pocket camera and in low light, it is very close in performance to the other cameras. Dpreview shows the Fuji shold be much better, but in my real world shooting, it was hard to see it. It is definitely expensive at 750 for the RX-100 Mark II and $650 for the RX-100, but it does take pictures (particulary in bright light) that rival these other ones. In low light, I was surprised to see how close it came, although it is obviously nothing compared with a big dSLR like the Nikon D800 or the Canon 5D3. Finally, it is the one camera here that does videos well. We didn’t talk much about this, but it does shoot 1080p60 HD and its contrast detection isn’t bad for casual use.]
DxO is a good match for this camera and this is definitely one where you want to use the RAW processing. The JPGs out of camera are OK, but you definitely get more quality by post processing. Between the RX100-II and the RX-100, it seems to me that swallowing the price increase makes some sense. The camera is about 1/3 EV faster which it needs, but more importantly it has a 49mm filter built in (no funky glue) and a Sony grip. Although I really liked the Franiac. Combined with a Giotto lense protector (it is glass) and most importantly a clearview viewer and it is decent for casual shooting.
Here is a good light photo, hard to see the 10% resolution loss unless you go side-by-side I think
Finally, in low light, the image stabilization really does help, here is a New orleans image (I didn’t have the camera for the comparison shoot), so it isn’t super fair, but this gives a good sense of the 2,000 or so photos I shot with it: