Olympus EVOLT E-520

This section of our report focuses on the intensive image quality testing we put the E-520 through, to see how it stacks up against the other cameras we have reviewed.

The E-520 scored very well on color performance and resolution, but gave unimpressive results on the noise, white balance and dynamic range tests. These poor scores coupled with a number of average ones gave the E-520 an overall disappointing result.

Color (10.29)
The E-520 performed its best in the important area of color reproduction, producing a result that was highly accurate, especially in the skin tones. The E-520 was tested for color fidelity in our labs by shooting the industry standard Gretag Macbeth color charts under strictly controlled lighting. These images are analyzed using Imatest, a program designed to tell how accurately the camera performs in a number of areas, including color reproduction, resolution and image quality. Imatest produces a series of graphs and charts to display its findings, including the image below, which shows how the color value recorded by the camera differs from the values on the original chart. The outer rectangle shows the captured color, the luminance-corrected color is shown in the inner square and the ideal color is the small rectangle on the right.

The ideal color in this test result represents the original color,
the outer rectangle the camera's rendition.

The E-520 performed well in this category, especially considering the camera's relatively low price point, as it matched all the colors very closely.

This second chart maps out the differences between ideal and captured colors, by showing the ideal in the square and the captured color in the circle. The length of the line connecting the two corresponds to the color difference. This means that a short line indicates the ideal and captured color levels are very close to one another.

The squares indicate the original color chart values,
the circles what the E-520 captured.

Resolution (9.85)
In this test we look at the level of detail a camera can capture. A camera that takes images with higher resolution gives you the option to crop and zoom in on your images, allowing a closer look at minutiae. While the number of megapixels is a significant factor in determining resolution, this figure is not an accurate predictor of actual image quality. We test resolution by taking hundreds of photos of a standard resolution chart under a number of camera settings and distances from the chart, and analyzing the images using Imatest, which determines the number of alternating horizontal and vertical lines the camera can reproduce, measured in line widths per picture height, or lw/ph.

 The industry-standard resolution chart that we use for testing

The E-520 performed relatively well on this test, showing some improvement over Olympus’ previous model, the E-510. It placed comfortably between two other cameras, the Alpha 200 and D60, but significantly below the Rebel XSi. The maximum resolution found was 1871 lw/ph horizontal, and 1764 vertical with slight undersharpening for both, but this result worsened substantially at longer zoom ranges, bottoming out at 854 lw/ph.

Noise
With digital cameras, noise appears as speckling or static, and is especially noticeable in large areas of even color. Noise is more of a problem when shooting at high ISO’s, and can mar images quite substantially in low light conditions.

We test the noise level of a camera by shooting the Gretag Macbeth color chart at every available ISO, and use the Imatest application to measure image noise and trends across levels of light sensitivity. Most cameras have some noise reduction processing built in, so we shoot at all noise reduction levels in order to see how well the camera can cope in difficult lighting conditions.

Manual ISO (4.89)
The E-520 did poorly in our manual ISO test. The noise at ISO 100 was tolerable, if a little high, but it rapidly rose to an unacceptable level when not using noise reduction. A reading of 4% noise at ISO 1600 puts the E-520 performance far worse than competitors’ cameras. While noise reduction technology dramatically improved results, it barely brought the E-520 result to the level of the Sony Alpha 200 performance without noise reduction.

Auto Noise (1.85)
The Auto Noise test gauges the camera’s ability to automatically choose the correct ISO for the lighting conditions, and produce an image with the lowest possible noise level. The E-520 fared abysmally in this test, performing far worse than any of the other cameras, including the older E-510. The reason for this low score is that the E-520 consistently chose to shoot at ISO 400 while in auto ISO mode, even in conditions that were bright enough for ISO 100. This led to overly noisy images, and a very low score.

White Balance (3.74)
The human brain has the ability to adjust to any sort of light source, letting us perceive a white object as white whether illuminated by a fluorescent bulb or bright sunlight, even though these different sources actually give off light with distinctly different colors. Cameras get around this problem with white balance adjustment, either using one of a number of preset white values, or else metering the light and compensating for white balance based on this reading. In this section we test the accuracy of this function.

Shooting the Gretag Macbeth chart under fluorescent, tungsten (incandescent lighting found in most homes), daylight and flash lighting conditions tests the white balance abilities of the camera. Each of these tests is run through Imatest to measure how far the photos deviate from known color values, and how well the camera's white balance system copes. For each light source we shoot both automatic white balance and presets for that source to test how accurately both function.

Auto White Balance (3.17)
The chart below is similar to the color charts shown previously, with the outside square showing the captured color, the center square the color corrected for luminance, and the ideal color shown in the inset rectangle. These images are all exaggerated to show the difference more clearly, and you won't actually see this level of error in reality.

The auto white balance values for the E-520 are universally poor, with the camera inaccurately accounting for the light source. Shooting in the shade provided the worst results for this test. The best came from using a tungsten light, but were still unimpressive.
   

Preset (4.3)
The preset white balances fared slightly better than the automatic, but still not very well. The E-520 does have three fluorescent settings, and when the correct one was chosen, the only decent result for this test was achieved. However, no information is given on the camera or in the manual about which preset matches up with what type of fluorescent bulb. The preset shade setting was far better than the automatic, but was still a long way from ideal. Tungsten lighting provided the lowest score for this test, but this is a stumbling block for white balance on many cameras.

Exaggerated White Balance Errors (Presets)

Flash Preset WB - Flash Illumination

Florescent Preset WB - Fluorescent Illumination

Cloudy Daylight preset WB - Daylight illumination

Tungsten preset WB - Tungsten illumination

 
Still Life
In order to provide a useful comparison of shooting at different ISO levels, we’ve taken two standard shots at the full range of ISO’s available. You can click any of these files to see the full version, but be aware that they are large (some over 4MB) so they may take some time to load. 

Low Light (8.26)
Shooting in low light is a difficult task, and one that many cameras struggle with. We simulate a number of common low light conditions in order evaluate color fidelity as light levels drop, and changes in noise levels as exposure time increases.

Our score for low light is a combination of two tests. We shoot at a range of brightnesses, from 60 lux (which approximates normal indoor lighting) to 5 lux (the light produced by a single candle in a dark room). We test color accuracy in these varied light conditions, as well as noise levels. The E-520 scored below average in this test, due to the camera's ongoing problems with producing large amounts of image noise while shooting at high ISOs.

The second tier of the low light testing is taking long exposure shots, from 1 to 30 seconds, and testing color fidelity and noise levels. The E-520 scored much better in this test, showing high color fidelity, and bringing up the average score for this section, placing the camera on par with the Sony Alpha 200 and between the Rebel XSi and Nikon D60.

Once again, the E-520 showed its weakness in low light conditions, rapidly losing dynamic range with increasing ISO. While some deterioration is expected, the result here is extreme, and it falls far short of the levels set by other cameras.