As the Photolab standard setting is hopelessly oversharpened, I also sharpened the Nikon jpg for a better comparison. The grey walls still remain gray in the Nikon version, while the Photolab version has all kind of weird colors:
I also compared with DeepPrime (left) vs the sharpened Nikon jpg. With DeepPrime, the problem does not exist, it looks very similar to the Nikon jpg (maybe even a touch better). However, I’d not like to use DeepPrime on every image (especially since no noise reduction has been used at all on the Nikon file).
This is a common artifact, not exclusive to PhotoLab. It’s good that you tried moire removal to see if that helps. Have you tried adjusting other noise reduction parameters? (It’s probably low-frequency chroma noise you’re seeing.) Or chromatic aberration removal?
I wouldn’t call any sharpening in PhotoLab hopeless. It’s very easy to reduce the amount of sharpening applied.
JPEG files from your camera are already demosaiced and processed for artifacts. RAW files aren’t. Having to remove moire is normal but not necessary for every image, so you’ll have to decide what’s best. If 25% does no harm, why not make it your default setting?
I recall asking DxO Support (some years ago) if there might be any downside to having Moire activated by default - and they advised against that … But, perhaps that was because they assumed my default would be applied by the “Magic Wand” (?)
FWIW, I don’t think DxO ever promised that its standard preset was a panacea or was even entirely tuned to a particular camera-lens combo. To some extent, yes - but not entirely. As with moire correction, too much or not enough noise reduction is also bad. The standard preset of 40 (which isn’t ISO-dependent) is way too much for most of my images, while my own preset of 15 generally does no harm but is occasionally not enough. Each of us has to find what’s right for our own photography and tastes.
Not true. Some corrections are applied even if all adjustments are deactivated. For example, white balance and color rendering.