Dynamic range of a CCD
Locked
|
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Hi
everybody, I
don't understand why the dynamic range of a CCD is defined as the full-well
capacity divided by the camera noise, and not minus the camera
noise. Thanks
a lot for your help, Laurent. ___________________________________________ Laurent Gelman, PhD Friedrich
Miescher Institut Facility
for Advanced Imaging and Microscopy WRO
1066.2.16 Maulbeerstrasse
66 CH-4058
Basel Tel.:
061 696 43 38 / Cell phone: 079 618 73 69 |
|
|
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
oversimplifying: you cannot distinguish between two
intensity levels if their diffrence is not greater than the noise...
Leoncio A. Vergara
MD
Assistant
Professor Director of the Optical Imaging Lab. (OIL), Dept. of Neuroscience and Cell Biology University of Texas Medical Branch (UTMB) 301 University Blvd Galveston, Texas 77555-0641 OIL phone: 409-772-3970 Lab-PMD phone: 409-7470019 fax: 409-7470015 From: Confocal Microscopy List [mailto:[hidden email]] On Behalf Of Gelman, Laurent Sent: Wednesday, July 23, 2008 9:46 AM To: [hidden email] Subject: Dynamic range of a CCD Hi
everybody, I don't understand
why the dynamic range of a CCD is defined as the full-well capacity
divided by the camera noise, and not minus the camera
noise. Thanks a lot for
your help, Laurent. ___________________________________________ Laurent
Gelman, PhD Friedrich Miescher
Institut Facility for
Advanced Imaging and Microscopy WRO
1066.2.16 Maulbeerstrasse
66 CH-4058
Basel Tel.: 061 696
43 38 / Cell phone: 079 618 73
69 |
|
|
In reply to this post by lgelman
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
The noise is variation in the signal,
which prohibits your ability to discriminate between signals varying less than
this. Dividing the FWC by less than this does not add to the ability to detect
differences, even though technically the bit depth of the image could be
higher. Equally, dividing by greater than the noise results in “binned”
values in the bit range which could, technically, be accurately detected as
different values if the bit depth were greater. Typically, if a camera has a
noise of 7 electrons and a FWC of 21,000, then the dynamic range (your ability
to discriminate to signals of different intensity) is 3000. In this case, the
manufacturer would set the bit-depth at 12 (4096). The offset is subtracted from the full
well capacity, this is set by the manufacturer (typically 50-200 counts in the
image). There are many more detailed descriptions
of this in various books (Video Microscopy by Inoue and Spring for instance). Hope this helps Paul S. Maddox, PhD From: Hi everybody, I don't understand why the dynamic range of a CCD is
defined as the full-well capacity divided by the camera noise, and not minus
the camera noise. Thanks a lot for your help, Laurent. ___________________________________________ Laurent Gelman, PhD Friedrich Miescher Institut Facility for Advanced Imaging and Microscopy WRO 1066.2.16 Maulbeerstrasse 66 CH-4058 Basel Tel.: 061 696 43 38 / Cell phone: 079 618 73 69 |
|
|
In reply to this post by lgelman
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal Hi everybody, I don't understand why the dynamic range of a CCD is defined as the full-well capacity divided by the camera noise, and not minus the camera noise. Thanks a lot for your help, Laurent. ___________________________________________ Laurent Gelman, PhD Friedrich Miescher Institut Facility for Advanced Imaging and Microscopy WRO 1066.2.16 Maulbeerstrasse 66 CH-4058 Basel Tel.: 061 696 43 38 / Cell phone: 079 618 73 69 www.fmi.ch Dear Laurent.
Actually it is a misleading way to define dynamic range.
The idea is that the read noise gives you the error in the intensity
of any measurement. So this definition of dynamic range implies that
it is a measure of the number of separable intensity levels that might
be visible in the data.
The problem is that, once the signal level exceeds the square of
the read noise (in electron/pixel) the ability of discriminate
different signal levels becomes limited more by Poisson Noise
than by the "dynamic range".
So, in sum, it is just a convention that works pretty well to
separate better CCDs from worse.
For microscopists however, the read noise is a more useful
measure because we seldom have specimens bright enough to produce a
"full-well" signal.
Jim P.
-- Jim Pawley (Summer address) c/o Postmaster, Egmont, BC, Canada,
V0N-1N0 604-883-2095, [hidden email]
|
|
|
In reply to this post by lgelman
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal The dynamic range is defined as the ratio between the largest detectable signal and the smallest detectable signal. For a small signal to be detectable, it needs to be at least as big as the read noise (that's just the common definition, we're dealing with probabilities/uncertainties and not with binary true/false here). Thus the dynamic range is equal to the ratio between FWC (which obviously determines the largest detectable signal) and the read noise. BTW, the dynamic range should not be confused with the maximum SNR or the number of gray levels, because for these, the photon noise (which is intrinsic to the signal) must be taken into account also. Beat At 16:46 23-07-2008, you wrote: >Search the CONFOCAL archive at >http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal >Hi everybody, > >I don't understand why the dynamic range of a CCD is defined as the >full-well capacity divided by the camera noise, and not minus the camera noise. > >Thanks a lot for your help, > >Laurent. > >___________________________________________ >Laurent Gelman, PhD >Friedrich Miescher Institut >Facility for Advanced Imaging and Microscopy >WRO 1066.2.16 >Maulbeerstrasse 66 >CH-4058 Basel >Tel.: 061 696 43 38 / Cell phone: 079 618 73 69 ><http://www.fmi.ch/>www.fmi.ch > > |
| Free forum by Nabble | Edit this page |