Re: using a scanner as a densitometer?

From: Ryuji Suzuki ^lt;rs@silvergrain.org>
Date: 02/21/06-01:08:24 PM Z
Message-id: <20060221.140824.220136750.lifebook-4234377@silvergrain.org>

The best way is to use the RAW tiff image file (signals coming out of
the scanner sensors with very minimal of processing--not even gamma
correction) saved by VueScan. The reading aperture of VueScan's
densitometer function is rather too small and this will lead to an
unacceptable level of measurement noise in the read value. You would
want to apply a certain reading area and average the values within the
area on this raw tiff file.

Then read a calibrated step tablet to find the mapping from the
optical density to the scanner's value. Find a suitable polynomial to
interpolate the measured values. (Or, use a linear interpolation
between two adjacent points read off the calibrated step tablet, if
the curve is reasonably linear, especially in low density range.) Find
the inverse function of the interpolating function. Then apply this
inverse function to the value read off the sample of unknown density
to obtain the optical density reading.

I use a custom MATLAB program to do this, and I don't know of a good
software package that is designed to do the right things.

Also, it's probably not very difficult to make a blue- and green-
reading transmission densitometer using some of the preamplified
photodiodes incorporating lens and filter. These are sold by
TAOS. Combine these with a white LED. (I was told by TAOS that the
ones with analog output may be discontinued, but they still make ones
with digital output.) These sensors are also very useful in making
enlarger meters as well.

From: ryberg <cryberg@comcast.net>
Subject: using a scanner as a densitometer?
Date: Tue, 21 Feb 2006 09:27:53 -0800

> I've read a couple of casual mentions of using a scanner as a densitometer.
> My Epson 3200 reads the density as my choice of 0 to 100% or 0 to 255 units.
> The only way I can think of to get a log density reading is to scan a step
> tablet as the same time I scan a negative and make comparison of the
> readouts. Is there another way?
> Charles Portland OR
>
Received on Tue Feb 21 13:08:43 2006

This archive was generated by hypermail 2.1.8 : 03/13/06-10:42:58 AM Z CST