Saturday, 5 December 2009

Ricoh GRD III Review – Part 3

After looking at the features it's time to see how the image quality compares between the GRDs.
The first test is between the three GRDs at low ISO and color.

Image Comparison GRDs – Part 1 (Color: ISO 64/80, 100 and 200)

This is not a scientific test and there are no lab tests either, it is a test to see how the cameras behave under normal shooting conditions. All cameras have been set to A-mode, EV -0.3, ISO 64 (80 for the GRD II), 100 or 200 respectively, Normal image parameters with NR off (GRD II and III) and Auto WB. The JPGs are unprocessed and the RAW files were batch processed with RAW Therapee 2.4 using the default profile.
The pictures have all been taken without a tripod.

Let's start with the first picture.

At ISO 64/80 and JPG you would not expect a huge difference between the cameras and while the details captured are very similar it is interesting to see that the GRD II underexposed the scene. This could be explained due to the different metering but what is strange is that it had the same shutter speed and aperture as the GRD III and a higher ISO with 80 vs 64, yet the picture is almost one stop underexposed and looks very noisy with smudgy noise reduction artefacts. As you will see further down, the GRD II always tends to underexpose even if there is no reason for it. I will need to ask Ricoh about this strange behaviour.

Lets see how the RAW files of this scene turned out and if there is a lot of difference between the JPGs.

The surprising thing is that the GRD III RAW is not so different from the JPG and both show similar levels of detail but the RAW file retains slightly more highlights. The GRD I RAW file is very good and although it is 4 years old and has 2 megapixel less, it is very crisp and shows the same amount of details. The GRD II RAW file is not as dark as the JPG but is full of noise without showing more details, if you use Noise Ninja to remove the color noise you will end up with less details.

The first picture shows that the difference between the JPG and RAW files of the GRD III is not very high at ISO 64. The GRD I is still an excellent camera and performs well in both RAW and JPG. The GRD II on the other hand is producing an underexposed picture which can't really hold up in either JPG or RAW.
Since this behaviour was evident in quite a lot of pictures I decided not to retake the pictures in the hope that the GRD II will get it right at some point.

Let's move on to the next picture, again taken at ISO 64/80.

The GRD II again produces a slightly underexposed picture but this time it's not quite as extreme. Aside from that all three GRDs show a similar level of detail. The only thing that stands out here is the way the GRD III renders the shirt of the woman in the front. She was wearing a black shirt and both the GRD I and GRD II show this but the GRD III shows this slightly purple/reddish, almost like the Leica M8 with the weak IR filter.
I have not seen this before so it could either be a problem with the new Multi-Pattern AWB or it can indeed be a weaker IR filter on the GRD III. Again, I will try to speak with Ricoh about this and see if they have an explanation.

Next is the RAW file and the question if the GRD III will still render the shirt slightly purple.

And the answer is yes, the shirt is purple even in the RAW file so this does point to a weaker IR filter. The GRD II again shows the most noise due to underexposing the picture. All three cameras however benefit here from using RAW, if you look at the trees in the background all cameras show more details when using RAW.

This picture did reveal a strange issue with the GRD III which seems to have a weaker IR filter and renders certain black textiles slightly purple. I have not noticed this before so it could have been a random problem with this particular scene or my GRD III sample.
Aside from this, all cameras performed well, with the GRD II again underexposing the picture and showing more noise in RAW.

The next picture is still at ISO 64/80.

Nothing too unusual here but this picture shows the heavy handed noise reduction of the GRD II quite well in the bleeding reds of the lights and the mushy details in the walls. The GRD III has the least amount of visible noise and shows the most details but rendered the scene slightly more orange.

The RAW file is next.

All three GRDs are very similar here when using RAW although the GRD III shows less noise, the new and improved sensor definitely pays off. The slightly orange cast is also only visible in the JPG file so it is only a different processing which introduced the slight orange cast.

Looking at this picture there really is not much to choose between the GRDs but the new sensor in the GRD III manages to pull out ahead by showing slightly more details and at the same time having less noise.

Now it's time to move on to ISO 100 and see how the cameras perform.

Again the GRD II underexposed but manages to still capture enough details. The GRD III shows the least amount of noise while still showing enough details.

Next is the RAW file.

Not much difference from the JPG here although the GRD II shows quite a lot of color noise.

This picture is more difficult to judge because the GRD II underexposed it yet again so has more noise. Overall though, if you shoot at ISO 100, you will not see much difference between the GRDs (but keep an eye on the GRD II exposure).

The last picture in this part is taken at ISO 200.

Although the GRD I seems to suffer from slight camera shake, it shows slightly more details in the map but also more noise. The GRD II does not show much noise but also not much in the way of details. The GRD III manages just fine but could still do with a less aggressive noise reduction or proper noise reduction 'OFF' setting.

Let's see how the RAW files compares.

The RAW file shows how big the difference is between the JPGs from the GRD II and III. Aside from that all three GRDs do a very good job when using RAW although here again the GRD III has the least amount of noise.

This brings us to the end of Part 3. A few interesting observations are that the GRD II tends to underexpose quite often and the GRD III shows that having a new sensor has paid off, by having less noise and more details than the older GRDs. I will need to look closer at the strange issue with the purple shirt and try to find out what could have caused it.

Overall, if you only shoot up to ISO 200 and color you can't go wrong with either GRD but getting the GRD I or GRD III is better because you won't have any surprises with underexposed pictures.

All pictures used here and a few more can be downloaded here.

Continue to Part 4...

No comments:

Post a Comment