How accurate is SpyderPrint

I have been using a Datacolor SpyderPrint for the last two months. I’m following Keith Schreiber’s instructions for using the SpyderPrint to read calibration targets for digital negatives. In Keith’s method, you use the Spyder in manual mode to measure Lab* values on a target, and you use the L* values to make your curves. I decided to test the accuracy of the L* values with the following experiment.

I made a series of 20 measurements of the same sample (what I actually did was to place the spyder back in its cradle and click 20 times). Here are the results:

minimum L* value = 88.05

maximum L* value = 88.67

average L* value = 88.37

standard deviation = 0.21

Given that the Spyder measures L* values to two decimal places of precision, the lack of accuracy is shocking. Has anyone else noticed this? Or is it possible that my Spyder is defective?

The short-term repeatability rating of the i1Pro2 (a $1500 dollar instrument) is .1 so a spectro that is 5 times cheaper is fine at .21.

Remember, you are measuring from L90ish to L24ish. So your standard deviation is only 1/314th + you are smoothing the measurement from there (partly due to this deviation but mostly for other reasons).

aka, don’t worry about it. Your brush-mark deviation is way more of a thing than the measurement device.