Advertisements

Four Thirds sensor resolutions tested

Our tests show the same megapixels don't necessarily deliver the same resolution

There has been some recent discussion about sensor resolution on the Four Thirds User forum, with one poster fearing that the E-420 resolved significantly less detail than the top of the range E-3. At the same time there has been some debate over the discover by several reviewers that the Panasonic Lumix DMC-L10 can resolve significantly more detail than any other Four Thirds DSLR, including the E-3, prompting the choice of the L10 as the default body for testing Four Thirds lenses.

I though it would be interesting to compare sensor resolutions of as many Four Thirds DSLR bodies as possible. I've almost managed to cover the entire gamut, missing only the Olympus E-300 (which I'd expect to be similar to the E-500 as they use the same 8MP Kodak sensor), Panasonic DMC-L1 and the Leica Digilux 3.

Testing was done use a standard ISO 12233 resolution test target, but placed at approximately double the distance normally required. In fact the correcting factor turned out to be 2.08x. The reason for doubling the distance was that 10MP cameras just about out-perform this target at normal distance. We hope to be getting a double resolution version of the target in due course.

The same 50mm f/2.0 Macro Olympus Digital Zuiko prime lens was chosen for maximum optical quality. With 2KW of tungsten floods illuminating the test target, the exposure was determined to be 1/30th second @ f/5.6, ISO 100, and set manually on all cameras. White balance was set manually to 3000K. Where possible, the cameras were triggered remotely or by self-timer. Again, where possible, mirror lock-up (anti-shock mode) was used, with a 5 second delay.

The results below are all taken from RAW files and exported from Olympus Studio 2.0, apart from the L10 file, which was exported from Adobe Photoshop Lightroom. Most of the test images were sharpened and contrast-enhanced, and noise filter settings switched off, though there were a few exceptions (see notes below).

As expected, the L10 delivered the highest technical result by some margin. A bit of a surprise is that the E-3 resolved slightly less than any of the other 10MP models. These results suggest that Panasonic has opted for very light anti-aliasing (deliberate blurring via a glass filter placed directly in front of the sensor) of their sensor in the L10, while Olympus has opted for stronger anti-aliasing on the E-3's sensor. The theory is that greater anti-aliasing will reduce the prospect of undesirable moiré patterns forming when photographic fine repeating pattern in subjects like textiles. It's also been suggest that increasing the anti-aliasing strength helps smooth tonal gradations. The cost is resolution.

Anyway, here are the results, and I hope you will find them useful!:

10MP Olympus E-520, corrected resolution: (lines per frame height): 2180 lpfh (lines per frame height).

 

10MP Olympus E-420, corrected resolution: 2180 lpfh.

 

10MP Olympus E-3, corrected resolution: 2080 lpfh.

 

10MP Panasonic Lumix DMC-L10, corrected resolution: 2600 lpfh.

It was quite hard to interpret the L10's result as the lines looked rather artificial at the higher end of the resolution scale. The lines are also clearly aliased (jagged) at the lower end of the resolution scale, much more so than the other cameras tested.

 

10MP Olympus E-510, corrected resolution: 2180 lpfh.

With the E-510 I found I needed to reduce sharpness compared to the others to avoid halos.

 

10MP Olympus E-410, corrected resolution: 2180lpfh.

 

10MP Olympus E-400, corrected resolution: 2180 lpfh.

The E-400 produced a very soft result, but responded well to sharpening and the result shown here is after additional sharpening compared to the other Olympus models, except for the E-510.

 

7.5MP Olympus E-330, corrected resolution: 1870 lpfh.

 

8.0MP Olympus E-500, corrected resolution: 1980 lpfh.

 

5.0MP Olympus E-1, corrected resolution: 1560 lpfh.

 

Reader feedback:

Discuss this story: