Digital BW, The Print

On 07/20/2012 03:13 PM, Peter Marquis-Kyle wrote:
> On 20/07/2012 6:10 PM, Ernst Dinkla wrote:
>> A blue LED that fits exactly the sensor's blue
>> filter spectrally may improve B&W film resolution.
>
> That's an interesting idea, Ernst. Could you explain why that would
> work, please?
>
> Peter Marquis-Kyle


I have used that on the Nikon 8000 scanner for B&W scans. Years ago I 
made my own wet mount filmholder and tweaked it to get the best overall 
sharpness around the edges for 6x9s. In Vuescan you can change the scan 
exposure per LED and take out either the Red + Green LED light or any 
other combo and use only the remaining LED, in this case the Blue one. I 
checked whether the scanner also does the auto-focusing with the Blue 
LED only and it does. I export the file as "DNG RAW" but the negative 
scan made positive so I can use the Photoshop ACR tools like the 
deconvolution sharpening easily. The yellow image is converted to a B&W 
image with ACR's filtering. Grain is half removed with Neat Image in 
Photoshop. The scans are the best I have made so far with that scanner.

My view on this is that with less LEDs involved and the ones used in a 
small spectral range with the shortest waves, like Blue LEDs are, plus 
an educated guess that Nikon used LEDs that fit the sensor's Blue filter 
best, I would get less flare and pixel blooming, the least light 
diffraction on the film grain and pull more information from the 
negative in total. With Tim Vitale I still think that I see aliased 
grain in the scan but now of a smaller size than with the other methods. 
The deconvolution sharpening does its work on the total of optical flaws 
from the camera lens film combination to the scan optics.

There have been more discussions on using only the Green separation of 
an RGB scan for B&W as it is the sharpest but I think that better 
sharpness is the compromise in focusing when all RGB channels are used. 
The suggestion that there is moire noise in the Blue separation is not 
what I have seen when I made separate scans with only the Blue or the 
Green LEDs used. I think this idea of less noise in the Green separation 
is exported from digital camera use where there are more Green pixels on 
the sensor than R or B. A linear CCD in a scanner has an equal number of 
sensor wells for R+G+B.

Given the last lines the use of a Green LED for B&W copying may be a 
better idea for DSLRs. The 2x number of Green pixels will outweigh the 
advantage of the shorter wavelength of Blue light.


-- 
Met vriendelijke groet, Ernst

Dinkla Grafische Techniek
Quad, pi\ufffdzografie, gicl\ufffde
www.pigment-print.com