Vuescan Histo Questions

[Phil]

Hello all. Am enjoying all the controls with Vuescan 7.6.79 and it is
definitely doing a better job than Nikon Scan. However I am perplexed
by black point, white point and brightness. When I open the tiff in
Photoshop, and adjust the levels, the first blacks always appear
around 10 to 15 and the histo is always a sheer vertical drop. I've
adjusted the black point in Vuescan to .12% but the issue persists.
Also, why doesn't adjusting black point make a change in the preview?
White point is at zero and seems normal in the PS histogram.
In addition, in Photoshop I am finding that the gamma needs to be
pushed to 1.4, 1.6, or even higher. This seems rather radical. The
latest image scanned for instance, I set the brightness in Vuescan to
1.18 and the gamma still needed lots of pushing in PS.
BTW I am scanning trannies at 4000dpi, 64bit rgbi, outputting a 64
rgbi tiff. Would I be better at 48-bit?
Did I overlook something in the User Manual?
Phil

 

[Mendel Leisk]

Hello all. Am enjoying all the controls with Vuescan 7.6.79 and it is
definitely doing a better job than Nikon Scan. However I am perplexed
by black point, white point and brightness. When I open the tiff in
Photoshop, and adjust the levels, the first blacks always appear
around 10 to 15 and the histo is always a sheer vertical drop. I've
adjusted the black point in Vuescan to .12% but the issue persists.

..12% is not necessarily equiv. to 10~15 white point. Photoshop's range
is 1 through 255, whereas Vuescan uses percent. I usually leave black
point at 0. Are your shadows really noisy? I think I'd rather have a
little space than clipping. I would think it depends a LOT on slide
content, as well. I usually have the opposite problem, some clipping
of black end, with Tri-X scans. I employ "lock film base color" to
counter this. Though I don't follow the "advanced workflow
suggestions" in the html manual, I just switch lfbc on and set all 3
additional boxes in the color tab to 1. Works for me with Tri-X
scanned with Minolta Scan Dual II.

Also, why doesn't adjusting black point make a change in the preview?

Perhaps you have auto-refresh off. This is actually the way I prefer
it. You can make adjustments until you are done, then tell Vuescan to
refresh, either thru one of the pulldowns (forget which) or by typing

White point is at zero and seems normal in the PS histogram.

In addition, in Photoshop I am finding that the gamma needs to be
pushed to 1.4, 1.6, or even higher. This seems rather radical. The
latest image scanned for instance, I set the brightness in Vuescan to
1.18 and the gamma still needed lots of pushing in PS.

I use brightness of around .85 with Tri-X negs, with the FEW slides
I've scanned, 1 bright did seem a little dark. 1.5~ bright does seem a
little high. Are your shadows not getting washed out looking at that
level?

BTW I am scanning trannies at 4000dpi, 64bit rgbi, outputting a 64
rgbi tiff. Would I be better at 48-bit?

64 bit just means inclusion of the infrared detected defect info, 48
bit does not.

 

[Erik Krause]

Hello, Phil
you wrote...

When I open the tiff in
Photoshop, and adjust the levels, the first blacks always appear
around 10 to 15 and the histo is always a sheer vertical drop. I've
adjusted the black point in Vuescan to .12% but the issue persists.

Vuescan uses percentual black and white point setting as long as you
don't lock image color. The percantage relates to the amount of pixels
that should be clipped to black or white, just like in Photoshop the
percent setting in the options dialog of levels adjustment.

If your image consits of 100,000 pixels and you set black point to
0.12% vuescan sets black point in order to clip the 120 darkest pixels
to black. If your selection area contains the film or scanning frame
border there might well be 120 pixels of total black.

Also, why doesn't adjusting black point make a change in the preview?

As a result the histogram does not change at all, since the 120 darkes
pixels are already black. Please read the manual on Border and Buffer
options in the crop tab.

If you lock image color, black and white point are displayed as
absolute values between 0.0 and 1.0

My personal preference is to set black point to 0.0% always (and white
point to 0.0% most of the time) since I consider vuescan black point
processing as far to coarse compared to Photoshop and since I don't
want to loose image data.

 

[Bart van der Wolf]

Hello all. Am enjoying all the controls with Vuescan 7.6.79 and it is
definitely doing a better job than Nikon Scan.

That seems to be the general impression, although milage may vary...

However I am perplexed
by black point, white point and brightness. When I open the tiff in
Photoshop, and adjust the levels, the first blacks always appear
around 10 to 15 and the histo is always a sheer vertical drop. I've
adjusted the black point in Vuescan to .12% but the issue persists.
Also, why doesn't adjusting black point make a change in the preview?

The (VueScan) percentages depend on image content, and the result thus will
differ between images. Clipping in the Black point will make a difference to
the overall color balance.

White point is at zero and seems normal in the PS histogram.
In addition, in Photoshop I am finding that the gamma needs to be
pushed to 1.4, 1.6, or even higher. This seems rather radical. The
latest image scanned for instance, I set the brightness in Vuescan to
1.18 and the gamma still needed lots of pushing in PS.

This seems to indicate a Color Management mis-match somewhere in your
imaging chain.

BTW I am scanning trannies at 4000dpi, 64bit rgbi, outputting a 64
rgbi tiff. Would I be better at 48-bit?

The additional 16 bpc just indicate that your scanner is (?) capable of ICE.

Bart

 

[Phil]

Erik and Mendel,
Thanks for the helpful tips. Forgot to say that I'm scanning color
trannies (set to Generic Vendor) on a SuperCoolscan 4000. This is only
my 2nd day, I haven't fiddled with B&W negs yet. Refresh is turned on
but still no change in preview image when adjusting BP and WP.
Brightness adjustments however do change the image in preview. I am
cropping out the black border and buffer is set to 15%. WP is at 0.
and Exposure Clipping is at 1%. It doesn't seem that setting BP to .12
and Brightness to 1.18 is hurting anything. When I open Levels in PS,
there is no clipping of the blacks and the shadows are looking good.
Could this be a characteristic of the scanner?
Also I'm a bit confused about the IR clean. I have it turned on to
medium in the Filter tab. Do I have to set bits per pixel in Input and
TIFF file type in Output both to 64rgbi? Right now I have them both
set to 48, thinking that 64 was more than I needed. When would you
want to output a different figure than you input?
Phil

 

[Erik Krause]

Hello, Phil
you wrote...

Brightness adjustments however do change the image in preview. I am
cropping out the black border and buffer is set to 15%. WP is at 0.
and Exposure Clipping is at 1%. It doesn't seem that setting BP to .12
and Brightness to 1.18 is hurting anything.

A BP of 0.12% is not much. There can be a bit of dust that causes some
pixels to be entirely black.

When I open Levels in PS,
there is no clipping of the blacks and the shadows are looking good.

You have better control if you use histogram dialog in PS. You get
pixel count if you select a region of the histogram with the mouse.

Could this be a characteristic of the scanner?

Yes, it is. The Nikon scanners do a high density range. They 'look'
even through the darkest parts of slides. Thats why slide black is not
RGB 0,0,0

Also I'm a bit confused about the IR clean. I have it turned on to
medium in the Filter tab. Do I have to set bits per pixel in Input and
TIFF file type in Output both to 64rgbi? Right now I have them both
set to 48, thinking that 64 was more than I needed.

At the moment I don't know whether vuescan scans the IR channel if you
set input to 48 bit and IR cleaning on. If you want to use IR clean
there must be the IR channel available, so it is save to set input to
64 bit (16 bit RGB ech channel + 16 bit IR channel).

When would you
want to output a different figure than you input?

Normally you don't need the IR channel in the output image, so it is
save to set 48 bit output. However, if you scan Raw to disk you should
set 'Raw file type' to '64 bit RGBI' to have the IR channel available
for further processing.

There is another scarcely documented point: If you set 'Raw output with
Scan', Raw data is written as it comes from the Scanner. If you set
'Raw output with Save' IR-cleaning and grain reduction filter are
applied before Raw data is written to disk. In this second case you
won't need to save 64 bit RGBI Raw.

 

[Wilfred van der Vegte]

Erik Krause wrote:

Normally you don't need the IR channel in the output image, so it is
save to set 48 bit output. However, if you scan Raw to disk you should
set 'Raw file type' to '64 bit RGBI' to have the IR channel available
for further processing.

It can also be a good idea to save the infrared channel to remove dust
and scratches in Photoshop. I created a custom action that gives me far
better results than ICE or VueScan. The disadvantage is that it takes a
lot of computing time, but then again, it can be executed in batch mode.

 

[Markus Plail]

It can also be a good idea to save the infrared channel to remove dust
and scratches in Photoshop. I created a custom action that gives me
far better results than ICE or VueScan. The disadvantage is that it
takes a lot of computing time, but then again, it can be executed in
batch mode.

Would it be possible to share this action with us?

regards
Markus

 

[Wilfred van der Vegte]

Would it be possible to share this action with us?

I could, but I'm not sure how much of it is scanner-dependent. I will
have to write some additional documentation but I'll see what I can do.

 

[Bart van der Wolf]

SNIP

and Exposure Clipping is at 1%.

That is a rather high value. As the help file says: "A nominal value of .01
works well for most images".

Bart

 

[Phil]

Bart,
Hmmm. Hadn't considered that. My monitor is calibrated to a Photocal
Profile but Vuescan wouldn't let me set this as the profile so I set
the Monitor color space to Adobe RGB, which is my default (98) in
Photoshop. Output color space is set to Adobe RGB. Am I on the right
track?
Thanks,
Phil

 

[Erik Krause]

Hello, Wilfred van der Vegte
you wrote...

I could, but I'm not sure how much of it is scanner-dependent. I will
have to write some additional documentation but I'll see what I can do.

I'm pretty used to actions and willing to read and understand them.
Could you send them to me by mail (without additional effort)? Reply
adress should work...

 

[Bart van der Wolf]

Bart,
Hmmm. Hadn't considered that. My monitor is calibrated to a Photocal
Profile but Vuescan wouldn't let me set this as the profile so I set
the Monitor color space to Adobe RGB, which is my default (98) in
Photoshop. Output color space is set to Adobe RGB. Am I on the right
track?

Possibly. I can't tell whether your Photocal profile differs a lot from
Adobe RGB, but it may be partially responsible for some deviations.
It is a bit odd that the Photocal profile can't be selected as a VueScan
monitor profile though.

Bart

 

[wim wiskerke]

Erik Krause wrote:



It can also be a good idea to save the infrared channel to remove dust
and scratches in Photoshop. I created a custom action that gives me far
better results than ICE or VueScan.

What is the difference? In wich respect is it better?

The disadvantage is that it takes a
lot of computing time, but then again, it can be executed in batch mode.

If you execute this in batch mode, does this mean you do not correct
or alter any setting by hand?

regards, wim

 

[Wilfred van der Vegte]

Hello, Wilfred van der Vegte
you wrote...




I'm pretty used to actions and willing to read and understand them.
Could you send them to me by mail (without additional effort)? Reply
adress should work...

Hi Erik,

I was planning to write the documentation and make them public anyway. I
think I'll have time to do it this weekend. I have some ideas about
which steps may be scanner dependent, resolution dependent or film
dependent.

The basic idea behind the action is as I described before in this newsgroup:

Select the IR
channel, adjust the levels so that the defects are black and the rest of
the channel is white. Then, select the combined RGB channel, choose
Select > Load Selection and select the IR channel; check 'invert'. Then,
there's several things you can do to this selection. Expanding it by 1
pixel and applying a one-pixel feather radius always seems to be a good
idea. Then you could apply a Gaussian Blur to the selection (several
pixels radius), or the Dust & Scratches filter, and you're done. Of
course, there are more sophisticated ways involving layer blendings,
etc., but what I described here is pretty effective and it's no big
deal. You can even record it as an action.

In the improved version, I use the Dust & Scratches filter to remove the
defects on a separate layer containing the full image, with the defects
and a small feathered area around them selected. Then, on this layer I
delete everything but the defects and the feathered selection around
them. And before flattening the image, I set the blending mode of this
layer to Lighten (for slides) or Darken (for negs).
Especially the D&S filter takes a lot of computing time when used on a
16-bits full scan from my DSE 5400.

 

[Bernard Leverd]

Bart,
Hmmm. Hadn't considered that. My monitor is calibrated to a Photocal
Profile but Vuescan wouldn't let me set this as the profile so I set

ICC profile names may have 2 extentions: icc and icm , and Vuescan
only recognizes one of them. Just duplicate your Photocal profile with
the other extention, and Vuescan will see it.
I also have a Photocal/Optical profile, and it worked like that for
me.
In case it would'nt work, IMO you should rather use sRGB as monitor
profile instead of Adobe RGB.
Bernard

 

[Bart van der Wolf]

(e-mail address removed) (Phil) wrote in message

ICC profile names may have 2 extentions: icc and icm , and Vuescan
only recognizes one of them.

In fact, you can change the extension VueScan looks for (pull down filename
mask in Windows version anyway), so you don't have to duplicate or rename
the file.

Bart

 

[Phil]

Bernard,
Thanks, you're right on the money with the .icc extension.
Unfortunately this doesn't make a difference in the scanning numbers.
I've re-articulated the problem in a new post. Am going to try sRGB.
Phil

 

[Wilfred van der Vegte]

On Thu, 04 Mar 2004 10:57:56 +0100, Wilfred van der Vegte



What is the difference? In wich respect is it better?

I wrote the action because VueScan seems to misinterpret the IR data
from my DSE 5400 in some cases. In those cases where VueScan works, IMHO
my action leaves less visible artefatcs than VueScan, or the ICE
implementation in the Minolta software. Still, in all cases, if you know
where the defects were, you can still see some remaining artefacts where
the heaviest defects were, regardless of how they were removed. Except,
perhaps, for careful manual clone-stamping.

If you execute this in batch mode, does this mean you do not correct
or alter any setting by hand?

Up till now I haven't found this necessary. For other scanners, some
tweaking might be needed for optimization.
I'm hoping to put the actions on line one of these days. I'll keep this
newsgroup posted.

Wilfred van der Vegte.