Nothing like using lensrentals.com and actually using all three cameras for a weekend. Really eye opening. Here are the quick notes and some test summaries. As an aside, I like to look at four review sets:
- Handling. Ming Thein and others who use the camers professionally
- Sensors. the specifications at DxOmark (which does great testing the sensor)
- Resolution. Photoreview.com.au (which does great testing the lenses at all ISOs, focal lengths and with RAW and JPEG as well as the full Imatest results I'm not quite sure yet how to interpret for sharpening and so forth
- Dynamic range. dpreview.com which looks at the various enhancement modes that help dynamic range alot.
- Color accuracy. Don't really have a good sense of this one, but nice to know the color biases of these cameras and get some hint for how to set the various "picture styles"
- Noise. That is what ISOs are really good. DPreview seems strong here with its various noise curves and also they test to see if the set ISO is really accurate. As an aside, it is interesting to see that for noise, all these cameras are about the same in RAW (except for the Fuji X100S) even cameras like the 5D3
So first on notes on how best to use the cameras that is what is their best settings for ISO, aperture, color etc:
- Used the Canon with my older EF-S 17-55 IS, we were all surprised how nice the camera really was even though it's specs are really pretty poor. With this lenses, DxOmark rates it as a Color depth 21.8, Dynamic Range 11.3 EV, Low-light ISO 843. As a comparison, the low-end Nikon D3200 is 24 bits color depth, 13 EVs dynamic range and low-light ISO of 1131! The main tweak is to turn up the highlight tone priority and the jpegs were pretty soft, so may want to turn up the sharpening as well. As an aside, it is hard as heck actually find this setting, it is in the Custom Functions *C.Fn-3" and you want to enable it to get a little more color depth. The good news is the colors out of the box settings aren't too pumped up. DxOMark gives it a 10P-Mpix rating and says the best resolution is 35mm focal length and f/2.8. As an aside, the really fair comparison would be with the Canon 40mm f/2.8 stm pancake. That is quite a lense, DxOMark gives it 10P-mpix at f/2.8 on the EOS-100D (vs 18MP native) and an incredible 18P-mpix on the 5D3 (vs 21MP native)
- The Ricoh GR needed some tweaking out of the box. Notably increasing the screen brightness to make it comparable, but technically, it scores right up there with the D3200 with color depth 23.6, 13.5 EV dynamic range and 972 low light ISO and unfortunately, Photoreview.com.au hasn't tested it yet but on the nearly identical Nikon Coolpix A it got a very flat resolution from with max at f/3.5 of 2550 lw/ph center, 2400 edge (9.5MP equivalent) so assuming the Ricoh will be similar at base ISO 100. DxOmark for instance sees the sharpness performance as nearly identical (an incredible 13MP rating which is remarkable given it is a 16MP sensor)
- The Fujifilm X100S didn't get these measured but the older X100 is right up there with the Ricoh with 22.9 bit color depth, 12.4 EV dynamic range and low light 1001 ISO. And is supposed to be incredible at F/4 and base ISO 200 (2700 lw/ph center, 2350 lw/ph edge or 10.4MP effective). The DR200 gives a 4.7 stop increase in dynamice range (on top of the very good 12.4Ev it has natively) and the DR400 is another stop. The tradeoff is that it actually underexposes by one stop in DR200 and then pushes brightness so you are limited to ISO 400 at DR200 and works mainly in situations that are brighter like good sunlight because you are really using shooting two stops less. DR200 at ISO 400 is actually shooting at base ISO 100. This camera has incredibly low noise up to ISO 3200!
For reference, for cameras I've already used,
- Sony RX-100. Wow, this thing is just so small. It really comes out looking very good compared with these larger APS-C and even full frame systems. The main limitation is the low light ISO with 22.6 bits color depth, 12.4 EV dynamic range and 390 low light ISO (the Mark II is about a third of a spot better on low light). It has a relatively low sharpness of 6P-mpix, but as we've talked about before, no one knows how DxO measures, but we do know that at resolution and dynamic range best aperture, it is incredibly good (F/3.2, ISO 125, 10.2mm, Raw) giving you very high resolution 2600 lw/ph center, 2400 lw/ph edge or 9.8MP effective). It doesn't push the colors as much as consumer cameras with yellow being the main weakness.
- Canon 5D Mark II. Ok, my trusty dusty (and now dead) body, amazing how well something five years old holds up, 23.7 bit color depth, 11.9 EV dynamic range (about the same as the Canon EOS-100D by the way and way less than these mirrorless ones), 1815 low light ISO (that is significant, about a stop lower).
- Canon 5D Mark III. The oe I will probably have to try next 24 bit color depth, 11.7 EV dynamic range, 2293 low light ISO. All these cameras need highlight tone priority it seems.
Notes from the web:
On the DxOMark test and what is meaningful:
Initially, I felt that the Ricoh GR's display wasn't as good as that on the Nikon, which rather surprised me -- this is, after all, an RGBW display, most likely a Sony WhiteMagic unit like that on my Sony RX100. It turned out, though, that Ricoh simply defaults to a very low brightness level. Manually turning up the brightness in the menu system yielded a display that was more attractive than that of the Coolpix A, with rich color and good detail. And if I set both displays to their maximum brightness, the Ricoh's display was noticeably the brighter off the duo. If you're willing to dial back the brightness, the lower default of the Ricoh coupled with the RGBW design likely contributes significantly to the extra 26% battery life you'll get after the Nikon runs its cell out, according to CIPA testing figures. Ricoh GR Review - Overview
The EOS 100D achieved a DxOMark sensor score of just 63, below that of all current entry-level rivals. Sensor metric scores for Color Depth and DR are all on the low side for an APS-C DSLR but Low-Light ISO score while low compared to the best is actually on a par with some models in DxOMark - Canon EOS 100D sensor performance
The higher the color sensitivity, the more color nuances that can be distinguished. As with dynamic range, color sensitivity is greatest when ISO speed is minimal, and falls rapidly with rising ISO settings. DxO Labs has focused on measuring only maximum color sensitivity.
A color sensitivity of 22bits is excellent, and differences below 1 bit are barely noticeable.
DxOMark - Use Case Scores
Dynamic range falls rapidly with higher ISO settings, as any analog or digital amplification performed will increase the noise in the darker areas, making it harder to distinguish between fine levels of contrast.
Maximum dynamic range is the greatest possible amplitude between light and dark details a given sensor can record, and is expressed in EVs (exposure values) or f-stops, with each increase of 1 EV (or one stop) corresponding to twice the amount of light.
Dynamic range corresponds to the ratio between the highest brightness a camera can capture (saturation) and the lowest brightness it can capture (typically when noise becomes more important than the signal, i.e., a signal-to-noise ratio below 0 dB).
A value of 12 EV is excellent, with differences below 0.5 EV usually not noticeable.
DxOMark - Use Case Scores
The SNR indicates how much noise is present in an image compared to the actual information (signal). The higher the SNR value, the better the image looks, because details aren't drowned by noise. SNR strength is given in dB, which is a logarithmic scale: an increase of 6 dB corresponds to doubling the SNR, which equates to half the noise for the same signal.
An SNR value of 30dB means excellent image quality. Thus low-light ISO is the highest ISO setting for a camera that allows it to achieve an SNR of 30dB while keeping a good dynamic range of 9 EVs and a color depth of 18bits.
A difference in low-light ISO of 25% represents 1/3 EV and is only slightly noticeable.
DxOMark - Use Case Scores