Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
You guys are so helpful, so thank you. I guess I should provide
additional information. I'm using the zeiss LSM to image fixed cells,
I'm using a oil immersion lens and my RI match is 97% matched. I've done
a PSF, its FWHM was close to theoritical, though I can not remember off
hand it exact value (the bead was 200nm). I will be looking at
co-localization and DO NOT want to reduce my z-resolution. Right now my
stacks are at 200nm thick (Nyquist required 250 nm) I'm scanning slowly
and line averaging, using 1 Airy disk pinhole (which I believe is the
diameter of the pinhole). I'm of the believe that you try and collect
the best confocal image before decon.
-----Original Message-----
From: Confocal Microscopy List [mailto:[log in to unmask]] On
Behalf Of James Pawley
Sent: Sunday, May 23, 2004 7:14 AM
To: [log in to unmask]
Subject: Re: 1 airy vs. 2
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
>Hi all,
To me, Guy's comments seem to cover only part of the problem.
As a "one Airy" pinhole will accept most of the light from the excited
volume that originates from near the plane of focus, clearly any signal
increase must represent signal that originates from either above or
below the focus plane. Therefore, one could consider it noise.
However, particularly if one deconvolves the raw data (as you really
should if you are to meet the Nyquist Criterion for reconstruction and
eliminate the out of bandwidth signal generated by Poisson Noise) then
it is not so clear that this signal is noise noise. After all, very
good 3D reconstructions can be made by deconvolving widefield data where
most fo the signal woudl be noise by this definition. Surely, even
partially confocal data (and a 2 Airy pinhole or even a 10 Airy pinhole)
is better than this (see paper by Agard group on "...Partial confocal
imaging"" in widefield.)
I think that the basis of our disagreement may have to do with what we
think we are trying to achieve: is it best "optical resolution" or best
"practical resolution" on a specimen capable of emitting only so many
photons. If you see the problem of spatial resolution as one almost
exclusively as one of optics/diffraction, then I agree with Guy. If you
believe that Poisson noise as a bigger factor in ones ability to
distinguish two closely-separated features, then the extra signal you
get from the larger pinhole (a process that makes z-resolution worse
faster than it degrades x-y resolution) may come out ahead. After all,
in confocal imaging,we very seldom collect even 100 photons/bright
pixel. The noise on this signal is 10%, This makes seeing the 25%
contrast implied by the Rayleigh Criterion very
problematic: especially after this 25% has been cut to about 10% by
averaging over the 2 or so pixels need by Nyquist sampling. It is my
belief that on most live-cell preparations, the balance favors larger
pinholes followed by deconvolution at the appropriate big-pinhole PSF.
Nyquist and Poisson really make Abbe complex. Decon reduces, but does
not eliminate this complexity.
Apart from this, one should use a larger pinhole whenever time or
photo-damage considerations prevents you from sampling in the
z-direction using less planes than Nyquist suggests (or when imaging
only one plane in a weakly-fluorescent specimen: Ca++ imaging?). This
just iallows yo to take some account of signal that would have been
missed entirely.
By the way, is "One-Airy" defined at the diameter or the radius of the
Airy figure?
Do all the Manufacturers agree?
Jim P.
>Search the CONFOCAL archive at
>http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
>
>I would propose that the sole reason for that suggestion is to make it
>look as if the software is doing something! It is totally crazy to
>open the pinhole above 1 Airy since you will axiomatically NOT gain
>signal only noise by doing so.
>
>There is one (significant) caveat here, though. If you are imaging
>under conditions of index mismatch or at depths which don't match the
>correction of your lens then spherical aberration will significantly
>increase the size of your spot. In that case you will want to open
>your pinhole to more that 1 Airy since the signal is no longer
>contained in an Airy disk. BUT - it is much better to avoid this
>condition if at all possible! Ie, use a water immersion / coverslip
>corrected lens with living samples in water and an oil immersion lens
>on permanent mounts.
>
> Guy Cox
>
>
>Quoting Perveen Biln <[log in to unmask]>:
>
>> Search the CONFOCAL archive at
>> http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
>>
>> Hello,
>>
>> I'm not sure if this topic has been discussed before, as I'm new to
>> the list. I was hoping for some feedback on the choice of pinhole
>> size. Right now I image at 1 airy disk, this provides nice clean
>> images (peaks are high compared to the valleys) I would like to
>> deconvolve the images,
> > and have been told to increase the pinhole to 2 airy disk, to
> provide
>> more information for the deconvolution. However, I feel that doing
>> so will degrade my image quality, I don't intuitively see an benefit
>> in deconvolving a "poor image". Any thoughts
>>
>> Perveen Biln
>> MSc. Candidate
>> Cardiac Membrane Research Laboratory
>> Department of Kinesiology
>> Simon Fraser University
>>
>
>
>--
>Associate Professor Guy Cox
>Electron Microscope Unit, F09
>University of Sydney NSW 2006
>+61 2 9351 3176
--
James Pawley (at home)
21 North Prospect Ave.
Madison, WI, 53726
|
