Here you can see a generated GretagMacbeth ColorChecker chart, place your mouse over any of the labels below it to see the color reproduction in that mode. Select a camera/setting combination from the 'Compared to' drop-down to comparative boxes inside each patch.
Pentax's decision to use 'Bright' as the default custom image style on the K20D means that, by definition, it does not represent color neutrally. Additional sharpening and contrast can be added to the 'Natural' custom image mode if you'd want neutral colors but sharp images with good contrast.
|Pentax K20D||Compare to:|
|Portrait||Bright Adobe RGB|
Artificial light White Balance
As with previous Pentax DSLRs, the automatic White Balance setting does a pretty poor job, neither succeeding in representing the scene as it appeared, nor counteracting the light sources' color casts. The presets are excellent, however, helped by the provision of three Fluorescent lamp settings to let you closely match the behavior of the light source. Helpfully, the presets can be applied to RAW files after shooting if you don't want to spend valuable shooting time switching between them (which takes a couple of button presses, given the lack of a dedicated WB button).
|Incandescent - Auto WB
Red: 15.5%, Blue: -24.1%, Poor
|Incandescent - Incandescent preset WB
Red: -0.4%, Blue: 0.5%, Excellent
|Fluorescent - Auto WB
Red: 3.1%, Blue: -13.6%, Average
|Fluorescent - Fluorescent preset WB
Red: -0.5%, Blue: 0.7%, Excellent
Long Exposure noise reduction / Night shots
A typical thirty second long exposure (night shot) at ISO 100 did produce hot pixels, though only in certain areas of the frame, suggesting the camera had tried to cancel those it had identified. Somewhat unhelpfully, the K20D does not allow you to turn off long exposure noise reduction (normally the subtraction of any positive data that appears in a second exposure taken with the shutter closed).
Instead the options are: ON (applies long exposure noise reduction on all exposures longer than 0.3 seconds), or Auto (applies long exposure noise reduction depending on shutter speed, sensitivity and internal temperature). Below is the result of long exposure noise reduction turned on. A couple of hot pixels are still visible in certain locations. Processing from the accompanying RAW files removes these, so is worth doing for long exposures.
|JPEG - 30 second exposure, F10||100% crop|
|RAW conv. - 30 second exposure, F10||100% crop|
A typical 'modern digital SLR' performance, no color cast or white balance issues, good exposure in a portrait shot with well metered flash power and slightly under-exposed shooting our color chart (because of the white background). Overall, good.
|Built-in flash||Built-in flash|
Overall Image Quality / Specifics
The K10D was criticized for its rather soft JPEG output, a problem that appears to have been corrected in its successor. However, by changing the default image mode to 'Bright' (in most regions, anyway), not only has Pentax increased the default level of sharpening, it has also changed the camera's color response. Thankfully the K20D shows, on the 'custom image' screen how it is biasing its color response as well as the other parameter settings, so you can tailor the settings to suit your tastes.
At high ISO settings (from ISO 3200 upwards), the images produced by one of the three K20D's we tested were prone to banding, with a grid-like pattern appearing on the images. The banding appears to be mainly in the green channel and is most prominent in a vertical band down the middle third of the sensor. This phenomenon wasn't present in the other two test bodies but the affected one was a production model, however, discussion of the issue on the Pentax SLR Talk forum suggests this isn't the only camera to experience the problem.
The biggest surprise in shooting with the K20D was the appearance of a large number of aberrant pixels that were particularly visible when shooting our test charts. Upon discovering the problem, we co-operated with Pentax to give the company an opportunity to solve the problem. Part of the problem (identified by members of the Pentax SLR Talk forum), related to the use of the camera's self-timer, an issue that has been rectified in version v1.01 firmware. However, even after loading the firmware and running the camera's pixel mapping feature, we are still seeing stuck pixels in images.
In response, Pentax has said: "Over the past week, our technicians have been investigating the “stuck pixel issue” that arose from your tests. As highlighted by yourselves, these appear when there is a strong black and white contrast. In these circumstances, the camera is trying to compensate for the stuck pixel by using the information from its neighboring pixels. In situations when adjacent pixels are a completely different contrast, especially black and white, the black pixel may appear white and vice-versa. This is why these stuck pixels are not appearing in the same place."
These pixels can appear in real-world images, in areas of high-contrast and are not correctly mapped-out by either Adobe Camera Raw or Pentax's own software. They certainly aren't noticeable in the majority of images (even when you know where to look) and are unlikely to be a major problem in most situations. However, they do occur and are not routinely removed even if processing RAW files. They may cause problems for copy-stand work or product photographers in which there will regularly be examples of high contrast and mis-colored pixels are unacceptable.
Below are some crops from our test chart and the most badly-affected real-world image that we found, to give an impression of the most destructive the problem is likely to be. Because they appear to be related to contrast, they cannot be identified by taking images of uniform surfaces, making them virtually impossible to quantify - however it is clear that the problem affects the K20D more than any other DSLR we've ever tested.
|Resolution Chart (JPEG)||100% crop|
|Real-world example (JPEG)||100% crop|
|Real-world example (ACR RAW conversion)||100% crop|
- 19 Photographic tests (Noise)
- 20 Photographic tests (DR)
- 21 Photographic tests (DR)
- 22 Photographic tests
- 23 Compared to
- 24 Compared to (JPEG)
- 25 Compared to (JPEG)
- 26 Compared to (JPEG)
- 27 Compared to (JPEG)
- 28 Compared to (RAW)
- 29 Compared to (RAW)
- 30 Compared to (RAW)
- 31 Compared to (RAW)
- 32 Compared to (Higher ISO)
- 33 Compared to (Resolution)
- 34 Conclusion
- 35 Samples