Digital BW, The Print

In a message dated 12/22/2005 8:06:45 AM Pacific Standard Time,  
DigitalBlackandWhiteThePrint@yahoogroups.com writes:

Claude,


> ...
> The truth is that 30x40 landscapes  don't sell all that well ...Claude>>
<Yes, but they are fun to  make and look at.  Luckily, my "little" Bronica RF
645 with Tech Pan  in it can make nearly perfect 40 inch prints.>>
 
Yes, when tamed, Tech. Pan was the ultimate film. In 1980, I started using  
it (as SO-115) in 4x5, before they marketed it as "large format from 35mm" and  
finally, per my and others urgings, made it in 120. Great stuff. Highest res. 
 film they ever made in B&W and a real spoiler for slow ISO shooters.

<<Question:  What MP count is going to be needed for a  sharp 30 x 40 B&W 
print
from a digital camera?  I've been estimating  that at least 24 mp in a full
frame 35mm might be what I need to wait  for.  (I'm hoping my TP stash in the
freezer lasts long enough for such  to be affordable.)>>
 
My old Dicomed Scanback is equivalent to 5x7 film as a 42 Megapixel  capture 
(no mosaics or interpolation). Foveon showed 4x8 foot prints from a 16  
Megapixel monochrome, 22x22 mm chip shot by Greg Gorman back in 2001, at  Photokina 
that year. This cause quite a stir as most people thought it was as  good as 
4x5 fomat, but only represented 8 megapixels. I would say you can  stitch up 
3-6 shots from you Rebel and pretty much conclude that you are pretty  darn 
close at 24-48 Megapixels. The Anti Aliasing filter cuts your ultimate  sharpness 
in half, though, as opposed to the MF backs, which rely on  firmware/software 
tricks to remove color aliasing, and maintain full sharpness  in the luminance 
channel. Patience will reward all of us, but it will be  difficult and 
expensive to beat Tech Pan if that is your standard.
 
Claude


Paul
www.PaulRoark.com 





[Non-text portions of this message have been removed]

> ... it will be  difficult and
> expensive to beat Tech Pan if that is your standard.

Yes.

Along those lines, I'm trying to maximize the scanned quality of the closest
alternative films to Tech Pan.  Grain seems to be the hardest TP quality to
match.   

In enlarging, a diffusion enlarger was generally felt to hide the grain
better than a condenser (or worse, point source) enlarger.  Some have
claimed there is a similar effect in scanning.

So, the question is, will a diffusion screen in the light path of a Nikon
8000 tend to hike film grain.

I tried a standard Kodak diffusion sheet, but it may have been too course.
It at best did nothing.

In that past, when I've needed a high resolution substitute view screen for
a camera I've found frosted adhesive tape on glass makes an excellent view
screen.  It has a very fine frosting on it.

With a piece of tape over the glass carrier, I may be seeing the effect of
reducing the grain.  However, the tape itself and how it adheres to a
surface is too uneven to be sure.  Further experiments along these lines is
definitely warranted.  So ...

Question:  is there a source of super-fine-grained diffusion sheets that
might be tested for this application?

Paul
www.PaulRoark.com

The folks that make these view screens may be able to custom make  
something that will meet your needs:
http://www.satinsnowglass.com/index.html

On Dec 22, 2005, at 8:56 PM, Paul Roark wrote:

>
>
>
>
>> ... it will be  difficult and
>> expensive to beat Tech Pan if that is your standard.
>
> Yes.
>
> Along those lines, I'm trying to maximize the scanned quality of  
> the closest
> alternative films to Tech Pan.  Grain seems to be the hardest TP  
> quality to
> match.
>
> In enlarging, a diffusion enlarger was generally felt to hide the  
> grain
> better than a condenser (or worse, point source) enlarger.  Some have
> claimed there is a similar effect in scanning.
>
> So, the question is, will a diffusion screen in the light path of a  
> Nikon
> 8000 tend to hike film grain.
>
> I tried a standard Kodak diffusion sheet, but it may have been too  
> course.
> It at best did nothing.
>
> In that past, when I've needed a high resolution substitute view  
> screen for
> a camera I've found frosted adhesive tape on glass makes an  
> excellent view
> screen.  It has a very fine frosting on it.
>
> With a piece of tape over the glass carrier, I may be seeing the  
> effect of
> reducing the grain.  However, the tape itself and how it adheres to a
> surface is too uneven to be sure.  Further experiments along these  
> lines is
> definitely warranted.  So ...
>
> Question:  is there a source of super-fine-grained diffusion sheets  
> that
> might be tested for this application?
>
> Paul
> www.PaulRoark.com
>
>
>
>
>

You could also try some of the diffusion from Rosco:
http://www.rosco.com/us/video/cinegel.asp
or maybe something from Edmund optics:
http://www.edmundoptics.com/onlinecatalog/browse.cfm?
categoryid=453&level=2

I saw positive references to a specific diffuser in some places (see 
it http://www.scanhancer.com/), but never tried it myself. May be this 
guy has a diffuser for Nikon, not only for Minolta scanner...

--Sergei

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Greg" 
<dfaprinting@y...> wrote:

>
> You could also try some of the diffusion from Rosco:
> http://www.rosco.com/us/video/cinegel.asp
> or maybe something from Edmund optics:
> http://www.edmundoptics.com/onlinecatalog/browse.cfm?
> categoryid=453&level=2
>

> 
> I saw positive references to a specific diffuser in some places (see
> it http://www.scanhancer.com/)  ...

It seems Minolta and this "scanhancer" had the same idea I did about
diffusing the scanner light source to reduce grain.  From our darkroom days,
it seems like it should work. 

So far, however, with the diffusers I already had and working with Tmax 100
and the Nikon 8000, I have not been able to produce any clear evidence of
grain reduction through this approach.  Perhaps a finer diffusion sheet
would work better, but maybe not.  

My latest experiment used 2 Kodak diffusers taped about 1/8 inch above the
glass carrier.  I would have thought any roughness from the diffusion
material itself would have been well out of focus at that distance, but
there was still no measurable evidence of grain reduction.  Sometimes I
think I see a reduction, but when I measure the Standard Deviation in the PS
Histogram tool, selecting a totally featureless middle gray "sky" shot,
there is no consistent difference between the scan with the diffuser and the
one with no diffusion sheets in the light path.

Note that I think it is also possilbe that Tmax 100 emulsion and tabular
grain technology may react differently to this approach that thicker,
old-technology films.

It may be worth a few more experiments, but I'm beginning to think there may
be no TMX grain reduction from this approach, at least with the Nikon 8000,
although surface blemishes, scratches, and dust may well be reduced.  

Reading the "scanhancer" page does not inspire confidence.  The Minolta
sample is also, perhaps not by coincidence, a "simulated" comparison.  (Hmm,
I wonder why they had to "simulate" a comparison?)

Paul
www.PaulRoark.com

You could always try wet mountind the film. Aztec makes an adapter 
for the 8/9000.

Where do you see that the samples are simulated? The page specifically says
the samples are NOT simulated:
http://www.scanhancer.com/index.php?art=10&men=10

I used the Scanhancer while I hade the Microtek Artixscan 120tf. It worked
really well on some films, on other films I could see no difference. If I
remember correctly tmx is one of the films where the Scanhancer had no
effect. 

In general the Scanhancer works well for films like Velvia that give you
pepper grain which shows up as black spots on the scan, or white spots in
the case of negatives. I remember it worked great for Delta 100 which had
exactly this problem.

Here's a sample I made from some Kodak PPN color negative scans:
http://daniel.staver.no/img/scanhancer01.jpg

--
Daniel Staver
http://daniel.staver.no

> Reading the "scanhancer" page does not inspire confidence.  
> The Minolta sample is also, perhaps not by coincidence, a 
> "simulated" comparison.  (Hmm, I wonder why they had to 
> "simulate" a comparison?)

Daniel,

> Where do you see that the samples are simulated?

The Minolta page with respect to their version of this.

> The page specifically says
> the samples are NOT simulated:
> http://www.scanhancer.com/index.php?art=10&men=10

I had not seen this page.  These are the types of results I expected, but
did not see.

> 
> I used the Scanhancer while I hade the Microtek Artixscan 120tf. It worked
> really well on some films, on other films I could see no difference. If I
> remember correctly tmx is one of the films where the Scanhancer had no
> effect.

I suspect that is a major part of the problem.  I was just using TMX.  I
wonder what it'd do with some of the older technology B&W films.  I might
test Agfa 25 (which I have on hand), although Efke 25 would be the currently
available film of interest that it might have more impact on.

> ... I remember it worked great for Delta 100 which had
> exactly this problem.

That's not a bad film to start with either.

It looks like the idea may warrant more experimenting -- with different
films.

Thanks for the information.

Paul
www.PaulRoark.com

Ah, you were talking about the Minolta version of the Scanhancer. I had the
Minolta 5400 before the Artixscan 120tf, and had the same results on that
one. Some films worked, others didn't. I don't really know why they would
have to simulate the results, they could just scan some Velvia with and
without the diffuser and you would clearly see the difference.

Try sending an email to Erik (the guy who's making the Scanhancer). From my
experience he's very friendly and very willing to work with you if you have
questions, ideas about how to use to Scanhancer or want to test it in a new
scanner. 

--
Daniel Staver
http://daniel.staver.no

> > The page specifically says
> > the samples are NOT simulated:
> > http://www.scanhancer.com/index.php?art=10&men=10
> I had not seen this page.  These are the types of results I 
> expected, but did not see.

> I suspect that is a major part of the problem.  I was just 
> using TMX.  I wonder what it'd do with some of the older 
> technology B&W films.  I might test Agfa 25 (which I have on 
> hand), although Efke 25 would be the currently available film 
> of interest that it might have more impact on.

I'm now showing about a 5% reduction in midtone grain with Acros 100 and at
least that much with Agfapan 25, where in some areas it looks more like a 9%
improvement.  These numbers are the change in the Standard Deviation measure
of featureless areas of the negatives.  I used 2 Kodak diffusion sheets
taped onto the top of the Nikon 8000 glass carrier, about 1/8 inch above the
glass.  

So, it's not dramatic, but I am seeing some positive effects from the
diffusion.

Paul
www.PaulRoark.com

Paul, I've got a Scanhancer, have used it with my Nikon V and slides.

Until I got Nikon's FH-3 carrier I couldn't use Scanhancer with
negatives (FH-3 uses Nikon's slide slot and light chamber which works
with Scanhancer but the SA-21 strip film carrier doesn't cooperate).

I have no observations as yet about silver B&W because I've seen no
reason for it, have been very happy with Vuescan's "slight grain
reduction" which does reduce grain while leaving it sharp.

With color slides I'm certain that there's a faint advantage to
Scanhancer Vs careful use of gaussian blur and USM, but I don't
believe it's often worth the effort because few images have ever been
improved meaningfully by diffusion enlargers or cold lamp heads,
despite Ansel's advice...Scanhancer does simulate that effect.

In additon, Scanhancer does defeat popcorn grain, which you'll get 
with large enlargements from C41 B&W films...but I find C41 B&W
plastic-looking, charmless, so I stick to my knitting with silver films. 

Djon

> That's not a bad film to start with either.
> 
> It looks like the idea may warrant more experimenting -- with different
> films.
> 
> Thanks for the information.
> 
> Paul
> www.PaulRoark.com
>

I said, unclearly, "I have no observations as yet about silver B&W
because I've seen no reason for it"

I meant that I "see no reason" for Scanhancer with silver film. I do
see a reason for it with C41 B&W, but I prefer the real thing.

Paul...you're actually measuring this? How? With what tool? 

Are you a statistician? What methodology did you use?

Also, have you used Vuescan's "slight grain reduction" ?  Its very
effective in practical application, and seems not to erode silver
grain sharpness or lose detail.

> 
> So, it's not dramatic, but I am seeing some positive effects from the
> diffusion.
> 
> Paul
> www.PaulRoark.com
>

> 
> Paul...you're actually measuring this? How? With what tool?

I scan a featureless film -- like a clear blue sky shot.  (Actually I use a
ground glass over the lens to make frames for comparing grain.  Also, I
shoot a step wedge type target with plain patches for calibrating my
development of film.)  Then in Photoshop the Histogram gives information
regarding the standard deviation within a selected area.  If the film and
scan were perfect, this would be 0 -- all pixels would have the same
density.  However, due to grain and other sources of density differences,
not all pixels are the same value.  The more they deviate from each other,
the higher the measure of "standard deviation."  Most of the density
differences appears to be due to grain.

I think it's a valid and objective way to make comparisons.  Perhaps we have
people more into the math that could comment further and explain the measure
better.

> 
> Are you a statistician? 

No


> Also, have you used Vuescan's "slight grain reduction"?  

No, I don't use Vuescan.  It sounds like it's a software algorithm like the
GEM that Nikon has -- which I find quite useful and better than I seem to be
able to do with Photoshop tools.  

Paul
www.PaulRoark.com

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Paul Roark" 
<paul.roark@v...> wrote:
>
> > 
> > Paul...you're actually measuring this? How? With what tool?
> 
> I scan a featureless film -- like a clear blue sky shot.  (Actually 
I use a
> ground glass over the lens to make frames for comparing grain.  
Also, I
> shoot a step wedge type target with plain patches for calibrating my
> development of film.)  Then in Photoshop the Histogram gives 
information
> regarding the standard deviation within a selected area.  If the 
film and
> scan were perfect, this would be 0 -- all pixels would have the same
> density.  However, due to grain and other sources of density 
differences,
> not all pixels are the same value.  The more they deviate from each 
other,
> the higher the measure of "standard deviation."  Most of the density
> differences appears to be due to grain.
> 

Well, if the use of a diffuser results in any measureable reduction
in (the standard deviation of) the distribution of the pixel values
for a uniform target, would not this imply that the actual
spread is due to the aliasing introduced by the scanner rather than
the graininess of the negative ?  I  did a similar set of 
measurements with a Nikon LS4000 and Neopan Acros as well as NP-400
(using CoolScan 4.12) over a year ago but the variation from one 
scan to the next were large enough that I gave up on the idea of 
getting smooth tones from a film scanner.

It would be great, though, if you could post your (raw) readings 
(mean and standard deviations) for the film base+fog, fully exposed 
film leader and the zone V for the films you mentioned (I am 
particularly interested in the Acros 100 and Neopan 400 either 135 or 
120 films.)  


Thanks !