Sampling

[Eric Scarfone]

hello list,
I've try to look into that question but I probably looked in the wrong places.

It is widely said that in order to respect the famous Nyquist criteria, one should sample signals at 3 times the resolution of the acquisition system.
Hence if ia calcuate correctly, using a 0.80NA objective to image fluorescein at 520nm (which gives a 390nm resolution), one should set the zoom of a laser scanning system so that xy pixels have a size of 130nm.
BUT, what if the structures I am interested in are in the 1 to 5 µm range? In that case I do not really need the ful resolving power of my objective and using say 300nm pixel size should provide enough sampling to correctly image 1000nm objects, or? What do the expert say??

There are  practical aspects to this question.
- using a high zoom factor exposes living preps to large amount of potentially harmfull photons  
- when one scans a limited amount of pixels to increase frame frequency, zooming to nyqvist drastically reduces the field that is imaged
- I am working on pretty old data that I have no possibility whatsoever to re-acquire with a higher sampling rate ... ;^)

Thanks for your thougts

Eric

 

 

 

Eric Scarfone, PhD, CNRS,
Center for Hearing and communication Research
Department of Clinical Neuroscience
Karolinska Institutet

Postal Address:
CFH, M1:02
Karolinska Hospital,
SE-171 76 Stockholm, Sweden

Work: [phone],
Cell: [phone]
Fax: [phone]

email: [hidden email]
http://www.ki.se/cfh/

[Craig Brideau]

For Nyquist as I heard it, you want to sample at a minimum of twice
the spatial frequency of the smallest detail of interest (to you).
Note this is the *minimum* for resolving the detail at all.  I suspect
the three times factor you suggested is to improve on this.  Usually
it is a matter of fiddling with the pixel size you get an 'acceptable'
image by your own standards.  Since your microscope is not an ideal
system rules like this "are more 'guidelines'", to quote Captain Jack
Sparrow. @:-)

Craig


On Wed, Oct 7, 2009 at 9:34 AM, Eric Scarfone <[hidden email]> wrote:

> hello list,
> I've try to look into that question but I probably looked in the wrong
> places.
>
> It is widely said that in order to respect the famous Nyquist criteria, one
> should sample signals at 3 times the resolution of the acquisition system.
> Hence if ia calcuate correctly, using a 0.80NA objective to image
> fluorescein at 520nm (which gives a 390nm resolution), one should set the
> zoom of a laser scanning system so that xy pixels have a size of 130nm.
> BUT, what if the structures I am interested in are in the 1 to 5 µm range?
> In that case I do not really need the ful resolving power of my objective
> and using say 300nm pixel size should provide enough sampling to correctly
> image 1000nm objects, or? What do the expert say??
>
> There are  practical aspects to this question.
> - using a high zoom factor exposes living preps to large amount
> of potentially harmfull photons
> - when one scans a limited amount of pixels to increase frame frequency,
> zooming to nyqvist drastically reduces the field that is imaged
> - I am working on pretty old data that I have no possibility whatsoever to
> re-acquire with a higher sampling rate ... ;^)
>
> Thanks for your thougts
>
> Eric
>
>
>
>
>
>
>
> Eric Scarfone, PhD, CNRS,
> Center for Hearing and communication Research
> Department of Clinical Neuroscience
> Karolinska Institutet
>
> Postal Address:
> CFH, M1:02
> Karolinska Hospital,
> SE-171 76 Stockholm, Sweden
>
> Work: [phone],
> Cell: [phone]
> Fax: [phone]
>
> email: [hidden email]
> http://www.ki.se/cfh/
>
>
>

[Greg Martin-8]

Hey Eric and Confocal Folks --

 

    There is no reason you have to sample at Nyquist.  Sample at the spatial frequency needed to answer the question at hand.

 

Be peace!  Greg.

 

Greg Martin

 

Keck Microscopy Facility
University of Washington School of Medicine
www.depts.washington.edu/keck

 

And

 

Lynn and Mike Garvey Cell Imaging Laboratory
University of Washington Institute for Stem Cell and Regenerative Medicine
www.depts.washington.edu/garveyic

 

206-685-8784 (office)
425-344-2632 (cell)

----- Original Message -----

From: [hidden email]

To: [hidden email]

Sent: Wednesday, October 07, 2009 8:34 AM

Subject: Sampling

hello list,
I've try to look into that question but I probably looked in the wrong places.

It is widely said that in order to respect the famous Nyquist criteria, one should sample signals at 3 times the resolution of the acquisition system.
Hence if ia calcuate correctly, using a 0.80NA objective to image fluorescein at 520nm (which gives a 390nm resolution), one should set the zoom of a laser scanning system so that xy pixels have a size of 130nm.
BUT, what if the structures I am interested in are in the 1 to 5 µm range? In that case I do not really need the ful resolving power of my objective and using say 300nm pixel size should provide enough sampling to correctly image 1000nm objects, or? What do the expert say??

There are  practical aspects to this question.
- using a high zoom factor exposes living preps to large amount of potentially harmfull photons  
- when one scans a limited amount of pixels to increase frame frequency, zooming to nyqvist drastically reduces the field that is imaged
- I am working on pretty old data that I have no possibility whatsoever to re-acquire with a higher sampling rate ... ;^)

Thanks for your thougts

Eric

 

 

 

Eric Scarfone, PhD, CNRS,
Center for Hearing and communication Research
Department of Clinical Neuroscience
Karolinska Institutet

Postal Address:
CFH, M1:02
Karolinska Hospital,
SE-171 76 Stockholm, Sweden

Work: [phone],
Cell: [phone]
Fax: [phone]

email: [hidden email]
http://www.ki.se/cfh/

[Guy Cox]

So why use that objective?  Maybe you need the higher resolution in Z that it provides, but if not why not use a x20 NA 0.5 objective?

 

However, if Z resolution is an issue the using a high NA objective and undersampling in XY can give you a nice dataset with equal resolution in all directions.  There is absolutely nothing wring with that.

 

                                                        Guy

 

Optical Imaging Techniques in Cell Biology
by Guy Cox    CRC Press / Taylor & Francis
    http://www.guycox.com/optical.htm
______________________________________________
Associate Professor Guy Cox, MA, DPhil(Oxon)
Electron Microscope Unit, Madsen Building F09,
University of Sydney, NSW 2006
______________________________________________
Phone [phone]     Fax [phone]
Mobile [phone]
______________________________________________
     http://www.guycox.net

 

---

From: Confocal Microscopy List [mailto:[hidden email]] On Behalf Of Eric Scarfone
Sent: Thursday, 8 October 2009 1:35 AM
To: [hidden email]
Subject: Sampling

hello list,
I've try to look into that question but I probably looked in the wrong places.

It is widely said that in order to respect the famous Nyquist criteria, one should sample signals at 3 times the resolution of the acquisition system.
Hence if ia calcuate correctly, using a 0.80NA objective to image fluorescein at 520nm (which gives a 390nm resolution), one should set the zoom of a laser scanning system so that xy pixels have a size of 130nm.
BUT, what if the structures I am interested in are in the 1 to 5 µm range? In that case I do not really need the ful resolving power of my objective and using say 300nm pixel size should provide enough sampling to correctly image 1000nm objects, or? What do the expert say??

There are  practical aspects to this question.
- using a high zoom factor exposes living preps to large amount of potentially harmfull photons  
- when one scans a limited amount of pixels to increase frame frequency, zooming to nyqvist drastically reduces the field that is imaged
- I am working on pretty old data that I have no possibility whatsoever to re-acquire with a higher sampling rate ... ;^)

Thanks for your thougts

Eric

 

 

 

Eric Scarfone, PhD, CNRS,
Center for Hearing and communication Research
Department of Clinical Neuroscience
Karolinska Institutet

Postal Address:
CFH, M1:02
Karolinska Hospital,
SE-171 76 Stockholm, Sweden

Work: [phone],
Cell: [phone]
Fax: [phone]

email: [hidden email]
http://www.ki.se/cfh/