Presently, the HCFR based sensor and application uses a calibration profile (sensor conversion matrix coefficients) for each display it is to be used with. If this is your first or only colormeter, the HCFR application (and website) includes many pre-made probe calibration profiles for various displays. The very low unit-to-unit variation of the TAOS photo diode array affords a method of creating a conversion matrix based on the spectrum characteristics of a display rather than an individual colormeter.
Alternately and perhaps ideally, the probe / app can to be trained to a known D65 reference (after you’ve used a different probe to establish a D65 white point) and a custom calibration file created.
To create a new calibration file, a short probe calibration routine is performed at first use where a D65 white field and the component colors (RGB) will be presented to the probe. The offset coefficients for the sensor conversion matrix will be calculated and a calibration file will be created specific to that display. This only needs to be done once per display.
The absolute ‘accuracy’ of the probe will depend on the calibration file and the accuracy of the reference instrumentation used to verify that. As I used a SpyderII to set my reference for my HCFR probe calibration, it won’t be anymore ‘accurate’ than the SpyderII it was trained against. Although it can’t be anymore accurate than the reference, it can perhaps track better or have a better range, better resolution or other benefits relative to the reference meter.
The nearest specs I could surmise would come from the TAOS TCS230 filtered photo diode array datasheet HERE and app note HERE.
It has been suggested by the HCFR developers that the need for a HCFR probe calibration file may not be necessary and are exploring methods of eliminating it. A document further discussing HCFR calibration files can be found HERE.