Per Baekgaard wrote:
As I've been on holidays, I haven't actively followed this discussion... and then of course, I come back home and find out that I am, too, being hit by this error, and should have been tracing it ;-(
Anyway, in my case, I'm using the NVidia server, and have it set up to _duplicate_ (clone) the same image on both the DVI and analog output. The reason being that I can then use my old, analog, switchbox to change between the PC's connect, but then also improve the picture quality (when and if I want) by pressing a key on the LCD display to select the digital output.
I have been looking forward to calibrate my system using a DTP-94, but of course now I cannot do so without changing the setup somehow, I think?
You simply can't calibrate multiple monitors independently, if the underlying video drivers don't allow programmatic access to the individual VideoLUTs that implement calibration.
It appears that while the standard X11 multi-screen and Xinerama multi-screen extension allow access to each screens VideoLUTs independently, NVidia TwinView, and ATI MergeFB do not. There's therefore nothing that can be done about this, other than not use TwinView or MergeFB if you want to calibrate multiple monitors independently.
You can try and apply the one calibration to all screens under these conditions (assuming that TwinView and/or MergeFB apply the main displays VideoLUT settings to all screens), and of course none of this stops you profiling the screens independently.
One thing that I have promised to address, is that currently the utilities bomb at the "scan for screens" stage when TwinView or MergeFB are running with multiple monitors, because the VideoLUT access extension is used to get a name to identify each screen. I will shortly issue a patch to at least make things a bit more forgiving than this, and only bomb on TwinViews or MergeFB if the non main screens VideoLUTs are actually accessed. This should at least address some peoples problems sufficiently I hope, to be useful.
Graeme Gill.