Digital BW, The Print

> <paul.roark@v...> wrote:
> > Notes: The results are expressed in Lab and were read with the X-
> Rite DTP 22
> > spectrophotometer.
> 
> Just had a quick question about your measurement technique...
> 
> When you are measuring the samples 3 times, are you reading the
> patches dark to light (or light to dark)in a complete pass, and then
> making 2 more passes, or are you measuring each patch 3 times and
> then moving to the next patch? Owning an old DTP22, the accuracy
> seems to be higher if you measure the entire strip, and then start
> over. Something about multiple readings of the same patch seems to
> make it less accurate. Why? I have no idea, my i1 does the same thing.
> 

I read each patch separately, not as a strip.  I thought I'd get the most
consistency by reading the 50% patch of the control strip, then the fade
test strip, then back to the control strip, etc.  From what I can tell, the
X-Rite is quite consistent within an individual session.  I suspect one of
the main sources of inaccuracy is heat -- including how long the instrument
has warmed up.  I'm not sure how long it needs to or if I'd be better off
just leaving it plugged in.  

Also, and a bit surprisingly, I found by manually calibrating it in quick
succession, before there would be time for the internal temperature to
change, it was not exactly the same after each calibration.  It's not that
it changes all that much, but for the small density changes I was trying to
measure, the minor inconsistencies are material.  Again, however, within a
single "session" -- same time and no re-calibration -- the meter is quite
consistent.  That is why I decided to go with the Control and Test strips as
opposed to the Test Strip's initial pre-test readings.  Although the test
strips were all read initially within a single session.  So, to the extent
whole session may have been affected by variance, it would be equally
affected.

At any rate, I try to eliminate as many variables as possible, since I'm
really only interested in comparing inksets and papers against each other.
I think either using the control or initial readings gives results that are
valid for that purpose even if the absolute readings are not entirely
consistent.

Paul
www.PaulRoark.com