[argyllcms] Re: color calibrating "ws2812b-led-panel" results in errors

  • From: Graeme Gill <graeme@xxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Mon, 28 Sep 2015 12:47:29 +1000

Martin Sperl wrote:

To generate some gamma curves for a ws2812b led panel I thought I could
use dispcal to get the corresponding values.

So I have placed my Spyder 3 on top of the LED Panel (in an almost dark room)
and ran:
./dispcal -v 9 -q l -y l -C ./transmit.sh ws2812b

White = XYZ 7151.499 16807.628 0.000
^^^^^ This is bad.

Looks like my Spyder driver hasn't recognized that it's too bright for
the instrument. I'd guess that all the other errors flow from that.

I can't actually find anything from DataColor that specifies what
maximum brightness it is rated for, but the sensor frequency
is well over its rated range for 1% accuracy (700Khz vs 10Khz).
Above some frequency I'd guess that the hardware stops counting
it properly, hence the whacky -ve Z value that gets clipped to 0.0.

I think I'll add a check to return an error if excessive brightness is
detected in any of the sensors.

So you either need a different instrument that can go as high
as 20000 cd/m^2, or you would need to interpose something like
a good (i.e. spectrally flat) neutral filter to reduce the light level.

Graeme Gill.

Other related posts: