Vuescan - histograms

[MAW]

Currently using v8.2.36 with a DiMAGE Elite 5400 and I'm a bit confused
about the two histograms. When scanning a slide, the top histogram in
Preview is at first bunched up to the left hand end and the lower one
occupies the full width. I then lock the exposure and do another
preview. The top histogram is now more evenly distributed across the
full width and increasing the exposure has the expected effect.

With negatives, however, the top histogram just gets 'stuck' to the left
hand end. Increasing exposure reduces the black clipping and moves the
'hump' to the right, but the white end stays put about 1/3rd - half of
the way along. Adjusting black and white points has no effect on the top
half, as expected.

Can anyone explain why this is please?

Mark

 

[Bart van der Wolf]

Currently using v8.2.36 with a DiMAGE Elite 5400 and I'm a bit
confused about the two histograms. When scanning a slide, the top
histogram in Preview is at first bunched up to the left hand end and
the lower one occupies the full width.

In older versions of VueScan, the histogram at the top showed Raw data
and the one in the bottom showed the same data after Black/White point
and Gamma adjustment and color balancing. Kind of a before and after
situation.

I then lock the exposure and do another preview. The top histogram
is now more evenly distributed across the full width and increasing
the exposure has the expected effect.

The initial preview is acquired with a default exposure, good enough
for cropping, rotation, etc. It also allows to calculate the optimal
exposure, partly determined by the cropped area's content. Locking
exposure will then use the optimized exposure level for subsequent
previews and the final scan.

With negatives, however, the top histogram just gets 'stuck' to the
left hand end.

Negatives have a lower intrinsic density range than slides, and the
contrast is lower, so the Raw histogram will not be filled as much as
slides do. The image will also be inverted from negative to positive,
putting what is on the Raw darker end in the output highlight area of
the histogram/

Increasing exposure reduces the black clipping and moves the 'hump'
to the right, but the white end stays put about 1/3rd - half of the
way along. Adjusting black and white points has no effect on the top
half, as expected.

Underexposure will lead to the densest parts of the negative (dark in
Raw, light after inversion) being un-penetrated by the amount of light
accumulated during exposure. Increasing exposure will move the Raw
data to the right of the histogram because more data is accumulated,
the densest parts of film will be penetrated. There is however still
not enough density difference/contrast between the densest and most
transparent film areas to span the full width of the Raw histogram.

More recent versions of VueScan have a slightly different histogram
display, one selects a type of histogram on the Prefs tab.

Bart

 

[Noons]

Bart van der Wolf apparently said,on my timestamp of 16/12/2005 10:53 AM:

Underexposure will lead to the densest parts of the negative (dark in
Raw, light after inversion) being un-penetrated by the amount of light
accumulated during exposure. Increasing exposure will move the Raw data
to the right of the histogram because more data is accumulated, the
densest parts of film will be penetrated. There is however still not
enough density difference/contrast between the densest and most
transparent film areas to span the full width of the Raw histogram.

More recent versions of VueScan have a slightly different histogram
display, one selects a type of histogram on the Prefs tab.

I'm using the latest version. The histograms are now quite usable
as they are in the same screen as the preview images.

Epson 4990. One relaetd question to anyone who might know the answer:

In some colour negatives, I get a red histogram that is shifted to
the left by about one quarter. I scan at 4800@16-bit and then
downsize the output to 2400@8bit INSIDE vuescan.

What I'm trying to do is come up with a "rule" of vuescan settings
to change for "shifting" the red histogram to the right so it matches
the other histogram's curves and how to "stretch" the 16bit histograms
so the range of densities when output at 8bit covers the entire gamut
of the output image (256 values for each colour).

 

[Bart van der Wolf]

SNIP

Epson 4990. One relaetd question to anyone who might know the
answer:

In some colour negatives, I get a red histogram that is shifted to
the left by about one quarter.

I haven't studied the specific capabilities of that scanner, but in
general a red shifted histogram indicates that the scanner didn't
expose each channel at it's optimum. It should be obvious in the Raw
histogram only.

Whether the scanner is capable of exposing individual channels with
individual exposure times, or whether the scanner driver doesn't have
the right instruction to do that, is something you could ask Ed
Hamrick if nobody else in this group knows the answer.

I scan at 4800@16-bit and then downsize the output to 2400@8bit
INSIDE vuescan.

What I'm trying to do is come up with a "rule" of vuescan settings
to change for "shifting" the red histogram to the right so it
matches
the other histogram's curves and how to "stretch" the 16bit
histograms
so the range of densities when output at 8bit covers the entire
gamut
of the output image (256 values for each colour).

Just to make sure. Do you see that in the Raw histogram only, and
after locking exposure and a new prescan?

Bart

 

[Noons]

Bart van der Wolf apparently said,on my timestamp of 16/12/2005 11:59 PM:

Just to make sure. Do you see that in the Raw histogram only, and after
locking exposure and a new prescan?

Only the raw histogram, after locking exposure and doing another preview.
And only with one type of colour negative film. All others are fine.

Now that vuescan has got the histograms so much easier to use,
what I'd really like to do is scan at 16bit and save at 8bit, while
using vuescan to set boundaries so that whatever is scanned in 16bit
is "expanded" into a full gamut 8bit scale. Haven't succeeded yet.

I can get the 16bit - 8 bit thing happening, that's easy. Already
had it sorted out a few releases back. But making sure the start and
end points in the 16bit histogram (whatever they may be in a given scan)
are mapped to 0 and 255 in the 8bit scale with all others in between
proportionaly spread is the bit I'm striking out on. It's probably
a dead simple white/black point adjustment or something like that, but
I just can't work it out...

 

[Bart van der Wolf]

SNIP

It's probably a dead simple white/black point adjustment or
something like that, but I just can't work it out...

After locking exposure, and a new Preview (to rescan at final exposure
level), choose lock image color (also on the input tab). You can now
adjust the channel Black/White points on the B/W points graph, and/or
on the Color tab.

One word of caution, channels do not necessarily all have a full
coverage of the histogram range. The "advanced workflow"
recommendation for (color) negative film will make sure that exposure
is practically optimal, shadows are neutral if the shadow subjects are
neutral, and a right Mouse-button click on a light (non-clipped) image
area will set correct overall color balance. There is little need to
fiddle with the channel B/W points anymore.

Bart

 

[Noons]

Bart van der Wolf apparently said,on my timestamp of 17/12/2005 8:48 AM:

After locking exposure, and a new Preview (to rescan at final exposure
level), choose lock image color (also on the input tab). You can now
adjust the channel Black/White points on the B/W points graph, and/or on
the Color tab.

One word of caution, channels do not necessarily all have a full
coverage of the histogram range. The "advanced workflow" recommendation
for (color) negative film will make sure that exposure is practically
optimal, shadows are neutral if the shadow subjects are neutral, and a
right Mouse-button click on a light (non-clipped) image area will set
correct overall color balance. There is little need to fiddle with the
channel B/W points anymore.

Bart

Ta muchly. Will give this a try.

 

[MAW]

[snip]
More recent versions of VueScan have a slightly different histogram
display, one selects a type of histogram on the Prefs tab.

Many thanks, very useful explanation. I thought, however, that negatives
had a wider latitude than slides, or is that something else?

Cheers,

Mark

 

[Bart van der Wolf]

SNIP

I thought, however, that negatives had a wider latitude than slides,
or is that something else?

Negatives have a wider *exposure* latitude. Their density range (as
seen by the scanner) is lower.

Bart

 

[MAW]

Negatives have a wider *exposure* latitude. Their density range (as seen
by the scanner) is lower.

Ah! Thanks. I've now got version 8.3.17 and find the new histogram
system much better.

Mark

 

[Robert Feinman]

Ah! Thanks. I've now got version 8.3.17 and find the new histogram
system much better.

Mark

To all enjoying Vuescan's newest features why not send a note to
Ed Hamrick telling him so. I'm sure the positive feedback can only
serve to encourage him to keep making new improvements (and at
no additional cost either).

 

[MAW]

To all enjoying Vuescan's newest features why not send a note to
Ed Hamrick telling him so. I'm sure the positive feedback can only
serve to encourage him to keep making new improvements (and at
no additional cost either).

I thought he read this news group. How does one contact him otherwise?

 

[Roger S.]

He may or may not read it but you can email him. See hamrick.com for
details.