Richard Hughes wrote:
When argyll does a calibration of the sensor for my ColorMunki, where is the calibration stored? On the device itself or somewhere on the system?
It stores it in a file named ".mk_XXXXXX.cal" in the current directory, where XXXXXX is the instrument serial numbers.
I ask because I'm currently using -N for spotread if the user has already calibrated once since starting gnome-color-manager. Is this a safe thing to do?
It is if you are running consecutive invocations of tools with the same current directory.
If it's stored on the device, I could fallback to requiring calibration once a week or so, rather than every single time you start the program, which is what I have now.
I wouldn't really recommend that. For these sorts of instruments the calibrations should be more frequent than that, since things like black level are very temperature dependent, and LED output levels are also temperature dependent (The ColorMunki actually measures its LED temperature by measuring it's forward voltage drop, and compensates for the relative warm up effects when it is turned on for each strip reading. There is no absolute temperature compensation though, it depends on the white reference for that.) In fact temperature effects is one of the known problems with measuring deep display blacks, since often displays are warmer than ambient temperature (and high brightness can also directly warm the instrument sensor), so the black level tends to rise from the point the instrument is attached to the display. So calibrating on just every use is a bit of a compromise - those after high repeatability will actually want to calibrate more often than that. (The Spectroscan automatically recalibrates every 50 readings for instance. I suspect the Isis does something similar.) Laboratory grade instruments do things like having a mechanical calibration shutter, so that they can calibrate before every reading. Graeme Gill.