 |
 |
 |
 |
The data are compressed by the log amplifier, and digitised to a linear
value. Both the number of decades that the amplifier works over and the
range of digitised values vary from system to system. In all cases you can
calculate the antilog, and hence the original signal level, if you know how
many decades are used (in your case four), and how many digitisation levels
("channels") there are (in your case 1024).
You then need to get these numbers to map out so that the maximum channel
number gives you the right exponent for antilogging, that is you need to
rescale the channels so that you have a full scale reading equal to the
number of log decades. In your case you divide by 256, in David Coder's
example
(http://nucleus.immunol.washington.edu/Research_facilities/Apps/logscale.htm
l)
where maximum channel number is 256 with four decades, the scale factor is
64. Finding the exponent to the base ten of your result gives you what you
want.
If you use a spreadsheet such as Excel, type in an equation something like:
=10^(ch/256)
where "^" is the caret you get by typing shift-6, and ch is the channel
number you're trying to convert.
Actually, if you want channel 0 to map out as 1, and channel 1023 to map
out as 10,000 like they do on B-D graphs, then you should use 255.75
(1023/4) as the denominator in the exponent, but this gives odd looking
intermediates and no-one seems to do it
Ray
At 9:14 pm 4/9/96, john@rm101.demon.co.uk wrote:
>Hi all,
>
>Can anybody help in converting channel numbers from listmode data
>(i.e. 0 - 1023) to their correct 4 decade log value.
>
>Many thanks
>
>
>John
|
|
|
|