Jonathan,

Don't know if there is a "simple review." You don't mention the
context either as in optical sectioning, deconvolution, etc.

The statistical noise for an 8 bit image is 6.25% and for 12 bit it
is 1.56%. Clearly the latter will have better signal to noise. But, a
single 8-bit image even with bright objects in it always looks noisy.
That is why confocal users invariably use something to reduce the
noise such as Kalman averaging. In this case, you are essentially
displaying the accumulation of maybe a 1000 counts (3% error) fit
into an 8-bit (256 levels) dynamic range.

So it really depends on what you mean. You need 16 scans with 8-bit
ranges to get the same range as with a single 12 bit CCD image. But
if you have to deconvolve the latter, or you have bleaching in the
former, quantitation can go either way. There is no simple response.

Personally, since I like using red fluors (better S/N relative to
autofluorescence) and the quantum detection sensitivity is much
better for CCDs in the red than just about any PMT, I go for the CCD.
But I have also used averaged 8-bit confocal images, which really
make it not 8-bit.

Email me if you want more specifics.

Mario


>Hello,
>
>Does anyone out there know of a simple review of 12 bit CCD
>quantitation vs. 8 bit quantitation.
>
>I need to address a critique that feels 8 bit confocal images are better
>for quantitation than 12 bit cooled (-35 C) CCD images.
>
>Thanks in advance.
>
>Jonathan
>
>
>--
>Jonathan T. Boyd
>Fox Chase Cancer Center
>7701 Burholme Ave.
>Philadelphia, PA 19111
>[log in to unmask]

--

Mario M. Moronne, Ph.D.
NanoMed Technologies
ph (510) 528-2400
FAX (510) 528-8076
Berkeley, CA
94706