Olympus E-30 high ISO comparisons



ISO 800, 1600 and 3200 from an Olympus E-30, E-3 and a Nikon D300 RAW and JPEG compared

There is no doubt in my mind that, despite an increase in pixels from 10 to 12MP, and so a reduction in pixel pitch, the new Olympus E-30 enjoys significantly improved high ISO image quality over any previous Four Thirds camera. Most telling is the absence of horizontal banding that can plague ISO 1600 and ISO 3200 images from previous Four Thirds DSLRs that used the Panasonic-made LiveMOS sensor, especially in images with large areas of darkness in the frame.

Four Thirds does have a reputation for poor high ISO image quality. The original Olympus E-1 was very noisy above ISO 400, although it was widely praised for the dynamic range and colour reproduction in its images at lower ISOs. That was with the Kodak FFT (full frame transfer) sensor, later used in higher resolution form with the E-300 and the E-500. I think the Kodak sensor was naturally quite noisy, but I'm also pretty sure that Olympus' image processing was not a match for Canon's Digic processor at the time.

The Panasonic tie-up, that resulted in the LiveMOS Four Thirds sensor, first seen in the E-330, did result in less noisy images, though high ISO banding became an issue. One aspect of the noise in LiveMOS generation E-System DSLRs has been really quite good control of chroma noise. This is much harder to correct in post processing than luminance grain. I have always thought that the banding problem was solvable, either through better management of the sensor data channels and related circuitry, or with improved image pipeline processing - better still, both. The E-30 has vindicated my hope and from a bandning point of view, high ISO images are now very uniform. This, alone, makes ISO 1600 and 3200 on the E-30 much more usable.

It seemed like an obvious thing to do - to compare the high ISO image quality of the new E-30 with the previous-generation E-3, which is still the E-System flagship model, of course. And just for fun, and as one was available, I decided to make a comparison with Nikon's much-praised D300 DSLR, which shares about the same pixel count as the E-30.

I'll be the first to admit that this test is not perfect. It's quite improvised, and the important issue of dynamic range is not catered for here, but I hope the test results will be of some use to those interested.

The cameras were mounted on a tripod and the shutters were fired after a self-timer delay to avoid operator-induced wobble. All image stabilisation modes were switched off and manual focus was employed, verified visually using magnified live view mode. Surprisingly, the D300, which has a superb 900K dot 3 inch screen, compared to the 230K pixel 2.7 inch screen of the E-30, displays a much lower resolution view for critical focus in live view, making crictical focus a lot harder to use, but that's another issue.

Low light test conditions

I chose a typical low light environment, illuminated with tungsten lighting. I set all three cameras to 3000K manual white balance. You will see from the JPEG samples that the Nikon produced a much warmer tone than the Olympus cameras, and less contrasty. It's possible that the D300's Picture Control settings, although set to 'Standard' may have been customised by the camera's owner - although he denies it!

It's worth pointing out that the D300 was fitted with an 18-200 Nikkor lens, which is not likely to be as sharp or as contrasty as a good quality shorter range standard zoom, like the 14-54 Zuiko Digital Mark II used on the Olympus bodies. All three cameras were set with noise filters set to 'standard'. This only affects the JPEGs, of course, but once again I have no idea if Nikon's 'standard' is the same as that of Olympus. Anyway, if you are seriously intersted in these tests, you will have to compare the RAW samples provided.

There was some variation in exposure, but the sample frames were matched to within aabout a third of an EV by evaluating the grey levels of the colour chart in the pictures in Photoshop, using a histogram view. In hindsight, I should have set the exposure to manual, but it was easier to leave it on auto, using centre-weighted metering.

JPEG crops gallery

There is a thumbnail gallery, with comments, of 100% crops to evaluate the JPEG samples on page 2 of this article, along with the links to the full size sample JPEGs and RAW files.

E-30 JPEGs look good at both ISO 800 and 1600

Fundamentally, the E-30 JPEGs look good at both ISO 800 and 1600. There is a noticeable increase in noise at ISO 3200 and the grain does look somewhat clumpy in dark areas at this setting, but light tones remain relatively clean. The increased resolution of the E-30 compared to the E-3 helps with the visibility of fine details and, if anything, I'm tempted to say that the E-30 has marginally less noise grain than the E-3. But I feel there is more smoothing being applied to the E-30 JPEGs so there is little, if any, increase in base resolution, despite the extra pixels. That may be disappointing, but the absence of banding makes a huge difference to the usability of ISO 1600 and 3200.

The Nikon D300 JPEGs are very smooth, even at ISO 3200, especially in the dark tones. Olympus still has some work to do there, but the Olympus results are more contrasty and detail stands out better. However, I do feel the Nikon is making better use of its sensor's resolving power - with JPEG output, at least.

Evaluating RAW

I haven't yet had time to evaluate the RAW files. Today's news that Olympus Studio (v.2.22) can now process E-30 raw files will make that task easier. Olympus Master 2.11 will also develop E-30 ORFs. I'll certainly be looking forward to your views based on your experiences of these samples.

Reader feedback:

Discuss this story: