> Try running with -D3. You can see the raw sensor values then. The value of Sensor seems to be important. A measurement is OK if its value is not zero. I will try to modify spyd2.c further to increase the time as much as it is actually needed to have Sensor non-zero. I do not know if it can help with grey or white patches (probably your assumptions on the integration time are perfectly OK there). Graeme - is it possible to introduce two modes of operation for the Spyder? The first one would be very fast but inaccurate (something like your values, maybe even less than this). This mode would be used to get a rough settings used in initial calibration (RGB values, brightness etc.). The 2nd mode would be "as slow as it needs" and used for actual calibration and profiling. I am not asking you to implement it (it is to early to do so :] - I will need to have a solid proof that Spyder can actually do even better job than it does now), I am just asking about theoretical possibility - would it be tough to do?