Thanks for the reply, Baddass - a fellow Bristol-dweller, I see!
I'm not actually a professional graphics bod, just an amateur digital photographer who's maybe a bit fussy! Today I tried looking at some of the same test images & pages on a friend's Samsung 913N, connected to his Mac mini. Different computer, room and lighting - trickier lighting for the Samsung - so not a direct comparison, but they all looked better on the Samsung than on the Dell. The Samsung reproduced a wider range of grays on my grayscale test page (which has 256 shades), only losing the lightest 2-3 shades as white whereas the Dell loses the lightest 8-12, depending on gamma. Grayscale images maintained subtle shading on highlights that were bleached out on the Dell. No minty tinges! Pretty good subtlety of skin tones on highlights where the Dell was burnt out and oversaturated.
In general images on the Samsung looked more like those on my Iiyama CRT. I was surprised by how good the tonal range and colours were, considering that it has, I believe, a 6-bit panel. I'd like to evaluate it more carefully in better lighting conditions and side-by-side with the Dell, but first impressions were favourable.
Thanks for the link to the calibration tool, however this seems to be for Windows and I'm running Linux. As far as I can see the only calibration I have available at the driver level is gamma setting for each colour channel. I've played with this and calibrated the monitor with a gamma test image that I found (I'm learning about this stuff as I go along, not an expert). As I said, changing gamma brings some improvement but not enough, and at the expense of dark colours. What sort of parameters do proper calibration tools play around with?
I found a review of an Iiyama TFT where image quality in general, and tonal range in particular, was said to be much better with DVI than with D-sub:
http://www.iiyama.co.uk/default.asp?SID=&NAV=1384&PCAT=0
- Follow link "CRT v LCD". Maybe I'll try the Dell on the DVI connection on my friends Mac. Perhaps I just got a sub-standard unit.
Thanks again for the response.