[argyllcms] Re: dispcal: Warning - RegOpenKeyEx failed with 2

  • From: Graeme Gill <graeme@xxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Fri, 21 Dec 2007 11:26:48 +1100

lee scratchy wrote:

C:\argyllV0.70Beta7_win32_exe\bin>dispcal.exe -v -c1 -yc test
Setting up the instrument
dispcal: Warning - RegOpenKeyEx failed with 2

any idea what I could do to fix that ?

Hmm. I suspect this is my fault, for assuming that all systems
have at least one serial port.

Is it correct that you have no serial ports on your machine ?

I'll fix the next release to cope more gracefully, in the meantime,
the easiest workaround would be to provide a serial port if your
machine has the hardware for it.

The only other suggestion I could make to get it running with the
current software is more risky, and that would be to add a dummy
serial port to your registry.

You would probably have to add a new key under "HKEY_LOCAL_MACHINE"
"HARDWARE" called "SERIALCOMM", and then add under that a string
with the name "\Device\Serial0".
You could probably leave the value unset (Normally it would have a value of 
If you aren't familiar and confident about making such changes
directly to the registry, then I wouldn't go in that direction though,
and I'm not sure if adding such an entry might upset other software.

Otherwise I guess you're stuck until I build a new release.

and once that will work, can I use full screen test patterns so I can also
> create a .cal file for my HC3100 DLP videoprojector ?

The display utilities are really set up to do one test patch at a time
driven by the software. dispcal can't really work any other way.
You might be able to create a crude profile with a static test chart
(more like the printer test charts), and measure each patch by hand
using the latest chartread utility (sorry, I haven't built
an MSWindows version of that yet), but if you can get the
computer driving the display, it's much more convenient, and
much more flexible.

For projectors, it's more about arranging the geometry so that the
instrument has it's field of view filled by the test patch, without
self shadowing, and making sure any ambient light in the room doesn't
interfere. (I tried this with the Huey and a friends projector a week or
so back, and the measurements at least looked sane, and the Huey has
the advantage that it's small).

One thing to investigate first with a projector, is whether it
uses the VideoLUTs of the computer or not using dispwin. If it
doesn't, there's no point in creating calibration curves with
dispcal, although the initial controls setup procedure can
be very useful, in that they can guide white point and contrast

Graeme Gill.

Other related posts: