In part 2 of the Sigma DP2 shootout, we'll look at how the DP2 lens performs at a variety of apertures and compare the overall detail captured by the DP2 with that of the Panasonic GH1 and Olympus M. Zuiko 17mm f/2.8 lens.
To begin, I'll explain my choice of RAW processors. The DP2 files were processed using Sigma Photo Pro (latest version) because there are no other RAW processors which demosaic DP2 files at this time. The GH1 files, on the other hand, can be processed using a number of different applications. I chose to use Adobe Lightroom (latest version) because it automatically addresses the considerable barrel distortion of the M. Zuiko lens. In doing so, one loses some resolution, and the blurriest extreme corners of the image get cropped out, but overall I think it gives the fairest comparison to the Sigma DP2.
How much do processing decisions affect apparent detail? Here's an example of an f/2.8 center 100% crop processed with Lightroom (right) compared to the same file processed for maximum apparent detail in Raw Developer (left):
Most of the difference there comes down to sharpening, though I do find that Raw Developer squeezes out a touch more detail. This is true for every camera I have tried to date.
Other testing details:
- Subject distance adjusted for roughly the same framing with each camera (greater distance for the longer lens)
- GH1 used in 3:2 aspect ratio
- DP2 files upsized using Genuine Fractals 5 to match the native GH1 picture dimensions
- Tripod, self-timer, base ISO, aperture-priority exposure, good light (high shutter speed)
Even looking at just the resized left half of the scene, there is a clear difference between the DP2 and GH1-17mm files at f/2.8:

Let's look next at the 100% crop comparisons, keeping in mind that the DP2 files have been upsized to match the GH1 file sizes.
f/2.8
Center:

More detail in the GH1 file as would be expected with the higher megapixel sensor.
Near edge:

DP2 outclassing the M. Zuiko here and also showing greater lens contrast. There is some color fringing in the GH1 crop.
Extreme corner:

Impressive wide open performance from the DP2 in the extreme corner of the frame while the M. Zuiko lags far behind. Again there is some color fringing in the GH1 crop.
f/4
Center:

GH1-17mm looking very good here with the lens resolving sufficiently to take advantage of the higher megapixel sensor.
Near edge:

Zuiko has caught up here in peripheral performance.
Extreme corner:

The Zuiko has improved but has a ways to go still.
f/5.6
Center:

As expected, the higher GH1 sensor is still capturing more detail at this aperture.
Near edge:

Still very close in this zone.
Extreme corner:

DP2 still has an edge in the extreme corner.
f/8
Center:

GH1-17mm still has an edge, but diffraction is bring the level of detail down closer to that of the DP2.
Near edge:

DP2 looks the better of the two here.
Extreme corner:

DP2 is clearly better.
Conclusions:
- The Sigma DP2 lens has balanced performance across the frame wide open at f/2.8, and by f/4 even the extreme corner peformance is quite good. Anything less would have been a disappointment given the precedent set by the outstanding DP1 lens.
- The M. Zuiko 17mm lens leverages the GH1 sensor potential in the center of the frame from f/2.8, puts in a decent peripheral performance stopped down to f/4 or f/5.6, but never quite nails the extreme corner.
- Softening due to diffraction is evident at f/8 with the Micro Four Thirds kit and not with the DP2.
- There seems to be higher lens contrast with the DP2 although the contribution of processing by Sigma Photo Pro cannot be ascertained. Application of wide-radius unsharp mask to a GH1 file gives it a similar look to that of the DP2 files.
Tariq · 811 weeks ago
amin 67p · 811 weeks ago
Raw Developer is fantastic. It doesn't have the intelligent auto functions of mainstream RAW apps, but the results are top notch. I'm going to try your DP workflow with RD.
Linh 36p · 811 weeks ago
I'm really banking on Panasonic doing good by their 20/1.7 since I preordered that GF1 kit. So far the samples seem to show it being better than the 17/2.8, so it's a good sign. Too bad it wasn't in time for this little shootout.
Robert · 811 weeks ago
I really hate cameras with clunky interfaces, and the E-P1 might the worst camera I've used, in this resprect. But no problems with the DP2.
Andrew · 811 weeks ago
Jason · 811 weeks ago
Could you at some point post some further insight into how you use RAW Developer? I'm personally not familiar with it, but just the results of your sample for the GH1 were quite convincing. If you processed the image with similar sharpening in Camera Raw/Lightroom, would they be closer?
amin 67p · 811 weeks ago
Ranger 9 · 811 weeks ago
In particular, I suspect that Genuine Fractals may be synthesizing some "detail" into the DP2 shots that is not actually present in the image formed by the lens (or in the original subject.)
I'd be interested to see what the shots would look like if this step were skipped.
amin 67p · 811 weeks ago
See here:
http://www.seriouscompacts.com/2008/04/dp1-shooto...
http://www.seriouscompacts.com/2008/04/dp1-shooto...
AfR · 811 weeks ago
Gman · 811 weeks ago
jason · 811 weeks ago
dp2 fan · 811 weeks ago
wolfie · 810 weeks ago
gianco · 808 weeks ago
Leo · 789 weeks ago