Archived posting to the Leica Users Group, 2001/10/17

[Author Prev] [Author Next] [Thread Prev] [Thread Next] [Author Index] [Topic Index] [Home] [Search]

Subject: Re: [Leica] Leica Users digest V21 #29
From: Alexey Merz <alexey@webcom.com>
Date: Wed, 17 Oct 2001 08:43:21 -0400

On Wednesday, October 17, 2001, at 03:01  AM, Leica Users digest wrote:

> Date: Tue, 16 Oct 2001 18:28:54 -0700
> From: Jim Brick <jim@brick.org>
> Subject: [Leica] RE:  Re: 24x36 CCD
> Message-ID: <5.1.0.14.2.20011016182022.032f0488@pop.alink.net>
> References: <5.1.0.14.2.20011016092630.032ea9a8@pop.alink.net>
>
> I cannot say for sure. But my educated guess is:
>
> The scanning backs no as single line CCD's can be made very dense, eg;
> Leica S1, better light, etc...
>
> Instant (non scanning) such as Phase 1, (can't remember the others) I
> suspect yes. Since you are using the center of the lens (small 
> CCD) you are
> using the best part and the lens frequency would cause fringing, 
> aliasing,
> etc., unless  cut to a frequency lower than the lens. Even true 
> with very
> large CCD's as they cannot be as dense as smaller CCD's.

Hm. I am skeptical, Jim. I certainly know of no case where a 
serious optical microscopist has used any kind of "frequency 
cutoff" filter when projecting an image from a microscope objective 
onto a scientific-grade CCD. We do generally correct for the 
overall resolution of a given lens - we try to match the 
objective's Rayleigh limit to the apparent CCD pixel size - but 
this is only to avoid the loss of high-frequency data, and we 
commonly bin 2X2 or 3x3 on the CCD with only a linear loss of 
spatial resolution. Certainly I'm not aware of artifacts other than 
linear resolution loss that arise through the use of less dense 
arrays.

As for the divergence issue, I could see that being much more 
serious. Many of the really hot CCD's on the market now, especially 
the Sonys, use a tiny plastic microlens (not a fiber-optic) over 
each well to increase the effective collection area (and thus the 
apparent quantum efficiency) of the wells. Rays hitting pixels at 
the corners would come in at acute angles, and therefore be more 
likely to be reflected and scattered instead of refracted into the 
well. This would result in both loss of information and increased 
noise.

The presence of a stepdown lens in the body could solve this 
problem AND protect the CCD surface from contamination. I know that 
people will screab about additional glass, but they are not 
screaming about the additional glass required when they use 
projectors and enlargers - a more serious problem!

Finally, another concern is the full well capacity of the sensor 
wells. The smaller they are, the smaller their dynamic range. 
There's a tradeoff between spatial resolution and intrascene 
dynamic range, and I'm not certain how much that tradeoff can be 
improved upon without radically new CCD designs. But we should not 
discount how fast the new CMOS chips might improve. Compared to 
CCDs which are apparently a pretty mature technology, CMOS is just 
in its infancy.

I for one would consider trading some starting resolution for the 
degradation that inevitably occurs during even the the best analog 
inaging chain (film to enlarging or scanning). I expect that our 
Leica optics would still shine, paarticularly at wide apertures.

Alexey Merz
Department of Biochemistry
Dartmouth Medical School, Hanover NH 03755-3844
alexey@webcom.com | alexey@dartmouth.edu
603-650-1702 | http://www.webcom.com/alexey/

- --
To unsubscribe, see http://mejac.palo-alto.ca.us/leica-users/unsub.html