[3dem] streaks in FFTs

[Henning Stahlberg]

Hi,

Similar streaks can be seen in STEM images that are recorded at pixel  
dwell times that are shorter than the reaction times of the  
photomultiplier. Our interpretation in that case was that the STEM  
detector didn't react fast enough to the varying signal, so that non- 
periodic variations in the intensity of sections of horizontal pixels  
lines in the images were produced that overlaid the image signal. In  
other words, stronger individual pixels are smeared out horizontally.  
This, however, does not affect all pixels, so that it is not producing  
a general modulation transfer function that overlays the image data.
My guess would be that in your CCD at least one of your ADCs is  
occasionally smearing the data out, or that your data still during  
transfer of the analogue signal  have some noise affecting it, before  
ADC conversion. That noise would be non-periodic and at a lower  
frequency than pixel read-out time (presumably lower than 1MHz).
It would be interesting to see the original image, and if any point  
spread function (PSF) problem affects equally all four quadrants, in  
case you have a multiport CCD.

In any case, I would contact Gatan first.

Henning.



Henning Stahlberg,
Molecular & Cellular Biology, Briggs Hall 5,
University of California at Davis, 1 Shields Ave., Davis, CA 95616, USA
Tel: +1-530-752 8282 (office), +1-530-754 8285 (lab), Fax: +1-530-752  
3085
mailto:HStahlberg at ucdavis.edu, Skype:henningstahlberg
http://stahlberglab.ucdavis.edu
http://2dx.org
_____________________________________________________________



On Jul 9, 2008, at 12:32 PM, Norm Olson wrote:

>
> -- 
> 3DEM:  We are seeing vertical streaks in FFTs from images recorded  
> with our Gatan 4K CCD camera.  We were wondering if any of you have  
> seen this and how we can solve the problem.
>
> Take a look at this image here for an example:
> http://cryoem.ucsd.edu/streak.jpg
>
> We are using our program, ROBEM, to box virus particles from our  
> recorded images.  From there we calculate an averaged FFT from the  
> individual image FFTs to determine the defocus of our image.  This  
> is just standard stuff...
>
> The streaks are very apparent in the averaged FFTs but we also see  
> them in FFTs calculated in ImageJ from our entire 4K X 4K images.  
> The intensity of the streaks is greater in extremely low dose images  
> but they are also visible in normal, and high dose images.  The FFT  
> included here is from an image taken at 300KV (problem appears to be  
> KV independent), ~59,000X, and with a dose of ~27 electrons/A  
> squared.  The sample was a carbon film.
>
> Any ideas?  Thanks
>
> Norm Olson
>
> ______________________________________________________________
> Norm Olson
> Cryoelectron Microscopy Facilities Manager
> 1510 Bonner Hall
> Department of Chemistry & Biochemistry, MC-0378
> University of California San Diego
> La Jolla, CA 92093-0378
> nholson at ucsd.edu
> http://cryoem.ucsd.edu
> Cell:  (858)220-2183
> (858)822-6718 - Office; (858)534-5846 - Fax
> ______________________________________________________________
> _______________________________________________
> 3dem mailing list
> 3dem at ncmir.ucsd.edu
> https://mail.ncmir.ucsd.edu/mailman/listinfo/3dem