Hi Alessandro,

As far as I know Confocal microscopy, that is marketed for the 
broad “Biological-Biomedical” market, and quantitative imaging is a joke, isn’t 
it? Every optical physicist knows that!

There is a little doubt that 100% of Biologists can discriminate millions of 
colours but fail at the greys (“going down from resolving power at 5 bit before 
the graduation (B.S.) to 4 bit after studying hard at the University (Ph.D.)” – 
sorry for another joke).

It would be great if someone who is responsible (or close to a successful
retirement) would be able to accurately evaluate and compare all the optical 
hardware including noisy PMT’s or 16 micron/pixel EM-CCD cameras, the latter as 
a standard even equipped with binning capabilities!!! – and Biologists should 
also learn how to “linearly” correct for the huge undersampling, let say, with 
the 60x N.A. 1.4 objective and 2x2 binning for the top of the line 16 um/pixel 
EM-CCD/ICCD camera. It sounds like a good company – Quantitative Imaging, Inc., 
that would only survive providing that a highly integrated business and 
academic environment would be enforced – please ask your Boss, Alessandro, if 
this statement is both scientifically and politically correct.
 
Unknowingly “swimming in artefacts” is a great pleasure, or criminal 
carelessness?
Broad education is another joke, sorry for the excessive and plain “joking”. 

Have a nice weekend,

Vitaly
NCI-Frederick,
301-846-6575


> Date: Fri, 24 Jul 2009 08:38:59 -0500
> From: [log in to unmask]
> Subject: Leica SP5 and detector linearity
> To: [log in to unmask]
> 
> Dear all,
> I hoped that Leica was pointing this out to their customers, but time 
> is passing and therefore decided to post this information.
> 
> For everybody that is using the SP5 for quantitative measurements 
> (eg., seFRET, apbFRET, FRAP, etc...) this may be relevant. With 
> typical biological samples we found the relation between detected 
> signal and light arriving to the detectors (PMTs) non linear (*). This 
> non-linearity largely depends on
the 
> SNR of the sample, but on three SP5 I checked, it may effect most 
> experiments.
> 
> This non-linearity depends on the read-out electronics that is 
> sampling the PMT signal at 40MHz. These samples are then averaged by 
> an FPGA. However, because the signal collected in 25ns is necessarily 
> noisy, the ADC has a comparatevely high probability to be saturated 
> (sotchastically just because
of 
> Poissonian noise). I do not enter further details to avoid to bore you 
> but
the 
> result is a non linearity.
> 
> Solution: we got from Leica a new software Data Transfer option
called "Direct 
> Overflow", below "enhanced". This happened a while ago and therefore I 
> expect most of you may have it and could try and check. With that 
> options, the saturated pixels are the ones in which at least one of 
> the 25ns samples got saturated, i.e. the "non-linear pixels".
> 
> However, because the probability of having at least one of those 25ns 
> samples saturated is typically high, you are going to loose dynamic 
> range. I was
trying 
> to get Leica adding an option to this read out mode that could allow 
> the FPGA to tollerate a certain number of saturated (not just 1, but 
> also not waiting that are all saturated) 25ns samples before indicating 
saturation on screen.
> Somehow this was not implemented and therefore, I recomand to use 
> "direct mode" for your measurements, but check carefully with the "direct 
overflow"
> mode and all other imaging options the same, if you suffer from this problem.
> 
> Alternatvely, keep your gain much lower than the "saturation level" 
> when not using the "direct overflow" mode.
> 
> If you are unaware of this, you may be underestimating FRET 
> efficiencies and having different recovery times in FRAP when 
> comparing bright to dim areas, just because of how the microscope operates.
> 
> Regards,
> 
> Alessandro Esposito
> University of Cambridge
> www.quantitative-microscopy.org


-------------------------------------------------