Digital BW, The Print

On 08/07/2011 05:54 PM, Lew Schwartz wrote:
> If I understand the problem of aliasing correctly, it results from a
> phenomenon similar to the one encountered in the darkroom with diffuser vs
> condenser light sources on the enlarger and, if this is the case, a diffuser
> may well be the answer. Since the 5000 has an auto frame advance mechanism
> to deal with, however, I don't think this would be very easy.

The enlarger light sources differ on the Callier effect in the film 
emulsion which is a kind of diffraction on the spaces between the film 
grain. There might be something of that Callier effect happening in the 
Nikon scanner too but it is likely more complex. There is little 
oversampling done in the Nikon scanner, the sampling rate comes close to 
the delivered resolution. In flatbed models like the Epson V700 the 
sampling rate is much higher than the actual resolution the scanner can 
deliver (based on the sensor well sizes, well pitch). Several sensors 
are shifted half the pitch width in both directions. More samples made 
with bigger sensor well sizes average the readings and noise is reduced 
that way.  Do not expect the resolution to be equal to the sampling 
rate. That oversampling has an analogy to the use of a wider aperture in 
drum scanning for grain reduction.

In practice I can not avoid aliased grain on my Nikon 8000. Nevertheless 
I  do not go the route to reduce grain by making the light more diffuse 
or with a small focus shift. Software like Neat Image can do a good job 
in suppressing aliased grain if I can define the aliased grain as good 
as possible over the frame size. Wet mounting of the film is one step, 
tweaking the film holders to get exact focus on all spots the next step. 
I checked using only the green or blue LEDs for the B&W scan. Vuescan's 
analogue gain sliders allow that:  the channel used at 2x, the other two 
taken out, RAW histogram to check whether there is enough exposure. DNG 
export. Further processing in ARC and Photoshop + Neat Image.
The blue light shifts the grain aliasing a bit on the tone range 
compared to green light but is not noisier in my experience. In theory 
the shorter wavelength blue light should show less diffraction between 
the grain particles if the grain aliasing is partly caused by the 
Callier effect. So far I used ARC's noise reduction very modest to 
remove noise from the positive highlights (actually then still the dense 
part of the digital negative) and in Photoshop Neat Image to reduce it 
in the positive shadows. Still not sure about sharpening, I think 
Photoshop's smart lens sharpening does it better and I should not use 
ARC's tools for it. ARC is typically more made for digital imagers while 
Photoshop still is a mix of tools suitable for both. All in all I can 
reduce noise and keep detail better than with the methods before.

In contrast with selecting one channel from a normally full RGB LED 
lighted scan the focusing in the scanner is also done with the single 
LED color so optimal for that light. I also think that it should reduce 
flare in the optical path and more optical artefacts that an RGB 
lightsource could create in the channel that is selected from a normal 
B&W scan. This method has some analogy to the use of single band/UV rich 
enlarger light in large format B&W photography. The neon A4+ cold light 
that I had made for graphic arts films must be on the attic somewhere 
but I did not forget the principles. Of course I checked what kind of 9" 
CCFL tubes are available with a blue light to make a B&W flatbed. There 
are some with a spectral peak around 457 NM, most likely right in the 
middle of the blue sensor sensitivity.

-- 
Met vriendelijke groeten,   Ernst

Try: http://tech.groups.yahoo.com/group/Wide_Inkjet_Printers/

|      Dinkla Grafische Techniek      |
|         www.pigment-print.com        |
|                 ( unvollendet )                 |