Digital BW, The Print

Djon wrote:

>  But I have seen many reports 
>  
>
>>of other methods like your's. 
>>    
>>
>
>Yes...we report that because the simple method works.
>  
>
With some 3200s it might work because the focus is at the glass level, 
it isn't working on my 3200.

>In my opinion Epson QC tries to keep the 
>  
>
>>focus at 1 mm above the glass as the film in the carriers is at that 
>>height. 
>>    
>>
>
>I don't know how "opinion" is relevant. If you have Epson
>documentation about "1mm", please share it. In MY "opinion," its 
>urban legend :-)  As likely the APPROXIMATE 1mm is just a convenience
> for injection molding...nothing to do with focus, since focus isn't
>critical, as you point out here:
>  
>
Norman Koren has a nice page on focus finding on the 2450, the later 
models do not differ on that aspect. Focus height differs per scanner 
due to QC.  The DOF of approx 2-3 mm is there to get a good focus for 
reflectance originals on the glass and on the rim around the glass when 
the lamphouse is taken away, that's about 2,5 mm above the glass. The 
film holders keep the film at 1 mm above the glass so in a concept like 
that you try to keep best focus at 1 mm above the glass as you want more 
resolution with film than with reflecting originals. Whether it is 
written down by Epson isn't the point, the whole concept of that range 
of scanners shows that they are designed like that. That my scanner has 
best focus at 1.5 mm above the glass and your's at the glass is just one 
indication that the normal best focus could be at 1 mm above the glass 
but QC isn't that tight.

> 
>  
>
>>in practice, it is hard to notice where it exactly is because the
>>    
>>
>DOF is 
>  
>
>>so extreme.
>>    
>>
>
>Yes..."in practice" its not significant in large film (unlike 35,where
>we work with sparse film information, large film allows a multitude of
>sins).  
>  
>
I think with that approach you are throwing away a lot more potential 
quality with that scanner.

>A scanner isn't like an enlarger or contact printer. It's a reader: it
>reads and reports densities. When a point source is used an optical
>compromise is evidently introduced...the reason Minolta scanners do so
>much better when manually focused than autofocused.
>  
>
There's much more analogy between enlargers and scanners than that there 
are differences. Even the matrix CCD sensors with multi sampling of the 
Epsons and Nikons need optics to get light focused on the sensor lines. 
That there's a compromise on that in the Epson concept to have larger 
wells in the sensors for other properties like dynamic range and Dmax 
doesn't make a lens obsolete. The half pixel offset of more sensor lines 
is another way to get to the resolutions needed, following optical laws 
as well.
The Nikon 8000 that I also have doesn't create a shift in quality 
between manual or automatic focusing. The wet mount carrier that I made 
and tweaked is very plane and with the film wetmounted it doesn't matter 
where the focus point is set, the lens will focus to one plane and the 
rest falls well within the very short depth of focus. That fix focus + 
more DOF can not equal active focusing isn't news. I can check that 
every day with the scanners I have. That a lot of DOF and fix focus plus 
half a pixel offset, large well, sensor lines still can deliver good 
scans is good news but doesn't make the rest of the optical laws a thing 
of the past.
Drum scanners are still considered to be the best around and they rely 
on a point light source, small beam etc. There's no optical compromise 
found it that concept.
There are more roads to Rome, all quided by Nyquist.

>  
>
>>I wouldn't dare to claim that you can do without a mask. 
>>    
>>
>
>Try it.  There may be an effect around the edge, but even that
>probably won't be noticable unless that edge is buckled or torn. 
>
>Precise location of Silverfast's dotted frame is what's critical
>...light frome elsewhere, such as around the edges, is simply rejected
>digitally.
>  
>
I have done it on both the Epson and the Nikon. Put a 35 mm slide on the 
wetmount carrier of the Nikon and keep the same exposure between masked 
and unmasked, the shadows definitely have a loss, starting from the 
borders of the frame and into to center. Vuescan measures at a given 
percentage from the cropline, the default is 15% and that was used. 
After long threads on how to improve the scan quality on the Epsons and 
trials done by list members of the Epson 3200-4870 lists there was one 
thing that everyone agreed on, masking helps a lot.

>Especially not 
>  
>
>>on the 3200 that has a fixed lightsource over the entire length of the 
>>scan area. 
>>    
>>
>
>The moving source may somehow account for the 4870's slight sharpness
>advantage over the 3200 with 35mm negs, but it's proportionately less
>significant with large, old negatives and common enlargement sizes. 
>
>For enlargement to a given size, a 4X5 (5") negative would get only .3
>(30%) of the benefit that a 35mm (1.5") negative would get, switching
>from a 3200 to a 4870. Might be visible in a mural, if that's relevant. 
>  
>
The moving light source gives less light spilling into the scanners 
interior, allows more (focused) light while keeping overall heat lower. 
Especially with larger films that heat is a problem for good scanning. 
That's why the wet mounting with a sheet of 2 mm floatglass above the 
film is so effective, it gives a better light distribution by the 
wetmount and the film doesn't expand with the heat from the lamp.
If we all printed A5 size I absolutely agree that it doesn't matter. In 
that case I wouldn't scan 4x5.

>How a densitometer can be 
>  
>
>>used for that purpose is a mystery to me.
>>    
>>
>
>Understood. Its a mystery due to careless word usage by marketers. A
>densitometer is simply a device that quantifies density, measuring
>transmitted or reflected light. Many devices can do that...even common
>spotmeters, if used with a stable light source.
>
>Incredibly accurate densitometers were in use fifty years ago by the
>expert lithographers. Good photolabs started using them in the
>Sixties. Until their recent sloppy decline, minilabs routinely used
>them to monitor process strips.  
>
>The practical use of a densitometer for our discussion would be in
>print evaluation (Forum centers on prints, visual objects, not digital
>readouts) .. we'd begin by measuring blacks around edges, since
>contrast and curves are so easily tweaked in Photoshop.
>  
>
It is cheaper, more revealing, preciser and easier to check the scans in 
Photoshop after scanning than measuring a print of that scan with a 
densitometer.  The last would probably not detect much difference in the 
prints of the mask/unmasked scans of that 35 mm slide. Even incredibly 
accurate densitometers  in use fifty years ago by the expert 
lithographers would have a hard time then. I'm in the graphic industry 
for 30 years meanwhile. Still learning though.


Ernst