Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

You can always expect a deviation at the high and low ends since there
will be, statistically, some pixels over 255 and under 0 which will
therefore be truncated.  But I must admit I'm surprised it reaches
down as far as 200 - playing about on an old Bio-Rad I found that
things seemed pretty linear as long as I kept the maximum below ~230.

On a modern 12-bit system you should anyhow have a pretty reasonable
linear range to play with.

                                                         Guy


Quoting Greg Martin <[log in to unmask]>:

>>     One more note on linearity:  We've found that on our Leica SP1 (8-bit
> system) that the response of the PMTs is only linear between 20 and 200 grey
> levels (these limits depend on how fussy one is about curve-fitting).  Same
> results whether we use a graded concentration of dye in solution, or adjust
> excitation with the AOTF (assuming AOTF attenuation is linear).  Does anyone
> know if this is typical?
>
> Peace!  Greg.


--
Associate Professor Guy Cox
Electron Microscope Unit, F09
University of Sydney NSW 2006
+61 2 9351 3176