You need a densitometer if you want to calibrate your B&W film / developer pair, if you are using Ansel Adam's Zone System, if you are using BTZS method, or if you want to monitor your color processing (E-6, C-41, and RA-4). The good news is that if you own any scanner, you already have one.
All you need to know is what optical density is.The formula is:
where is I0 is the light intensity before it enters the film, and I is the light intensity after it leaves the film. You can use any scanner to measure those values and then you can use any spreadsheet program to calculate the density. However, there is the catch: the scanner must work in linear B&W mode (gamma=1.00) without any modifications to the data which comes from sharpening, auto exposure, ICE, GANE and so on. If you are using SilverFast, you can use 16-bit HDR mode. I'm not familiar with other software, so use your manual. (By the way, this is the same mode you should use with negfix8 script for black-and-white images.)
If you try to use any other mode, your result will be meaningless.
So, the measuring procedure is as follows:
How accurate are the results? To check it out, I've used the IT8 transparent target. Each target comes with text file containing patch's L*a*b* actual values. You can convert L* value to the density, using this calculator, or you can use the following spreadsheet's formula (substitute the $$ with the L* value):
I've measured each gray patch using ImageJ Histogram feature (select the patch then press Ctr+H):
In this example the GS11 patch is measured: D=LOG10(65535/8413)=0.89
Here are the results:
As you can see, my scanner (Epson V750) is pretty accurate in the range of 0.2 - 2.0 D, which is enough for the intended purpose.You should note, however, that this is not a standard densitometer, so you cannot compare those results with measures from certified densitometers (Status A, E, I, M or T). You can use it to monitor your own processing.