Minolta 5400 or Coolscan 5000

[DavidTT]

It's a difference in design and target audience.

Both VueScan and SilverFast are "point-and-shoot" programs similar to
disposable cameras. OK for casual tourist but not for people who care
for quality or have an inkling of what they are doing. Nothing wrong
with that if that's what they are after, but...

What you are saying above is essentially: Hasselblad is garbage, Kodak
Instamatic is far superior - fixed focus, auto-exposure...


Forgive me, but that is just plain nonsense. Exactly the opposite is
true.

By doing image editing at the scanning stage you irreparably *corrupt*
the image at the earliest possible stage and with the crudest possible
"tools".

If, instead, you scan as faithfully as possible you will have the
equivalent of a "digital negative". After that you can edit as much as
you want and still have the original untainted - should you change
your mind.

By "adjusting" the image at the scanning stage using the Preview
"keyhole" you are just irretrievably damaging data coming from the
scanner. Something which is often impossible to correct later and
certainly not without additional loss of quality.

Don.

Thanks Don. I couldn't have responded better.

Aside from capturing a raw scan, a scanner's sw (native or third party)
edits the information in the raw scan to produce a "good" scan. Its ease
of use can be a benefit for the novice as well as to the pros, while
editing a raw scan in PS takes a lot more skill and effort.

Editing a digital image with any sw means shuffling or losing
information in that image. Before being edited by a scanner sw, a raw
scan has all the image's information intact. When editing in PS, you can
preserve the original raw scan, and can track what information is being
lost as you go. When editing by a scanner sw, you don't know what
information is lost and can never recover it. PS' editing tools are far
superior than any scanner sw can offer, but you do need to learn how to
operate them.

Those who suggest to get a "best scan" from the scanner sw before
editing in PS fail to mention how a scanner sw achieves this. A
scanner's hardware offers little or no exposure control, much like a
disposable drugstore camera. The user interface has exposure control and
film type selection, etc., leading the uninitiated to believe that they
are operating a Nikon F5. In reality, all these controls are done by the
scanner's sw.

Real World Scanning and Halftones (3rd Edition) by David Blatner got
this right, and NOT Scantips, etc. One middle of the road recommendation
in the book is to make moderate edits in a scanner sw to ease the
process and minimize the information loss, and make final and critical
edits in PS.

 

[Kennedy McEwen]

It's a difference in design and target audience.

Both VueScan and SilverFast are "point-and-shoot" programs similar to
disposable cameras.

I disagree. They certainly *can* be point and shoot packages in that
they will give very acceptable results out of the box, but they both
have a full range of adjustments, in some cases exceeding the range that
the native software provides, or which can only be produced by the
native software in conjunction with another image processing package
(eg. PS) and some effort and skill on the part of the operator.

Forgive me, but that is just plain nonsense. Exactly the opposite is
true.

Not exactly - analogue gain, for example, is an adjustment which can
only be made at the scanning stage and, when necessary, improves the
quality of the scan in a way that can never be recovered in Photoshop.

In such cases, failing to make the exposure adjustment at the scan or
prescan stage results in irretrievably lost image information. You only
have a limited number of bits on that ADC - it can never be argued that
it is better not to use them all and expect PS to make good what it
never gets.

By "adjusting" the image at the scanning stage using the Preview
"keyhole" you are just irretrievably damaging data coming from the
scanner.

Not in the context that Bruce was suggesting when he wrote: "Have you
tried the long exposure option in Vuescan? - that usually (but not
always) works for me in reducing slide shadow noise."

ie. the very situation where prescan adjustment is important.

 

[Mark L Caton]

It's a difference in design and target audience.

Both VueScan and SilverFast are "point-and-shoot" programs similar to
disposable cameras. OK for casual tourist but not for people who care
for quality or have an inkling of what they are doing. Nothing wrong
with that if that's what they are after, but...

What you are saying above is essentially: Hasselblad is garbage, Kodak
Instamatic is far superior - fixed focus, auto-exposure...


Forgive me, but that is just plain nonsense. Exactly the opposite is
true.

By doing image editing at the scanning stage you irreparably *corrupt*
the image at the earliest possible stage and with the crudest possible
"tools".

If, instead, you scan as faithfully as possible you will have the
equivalent of a "digital negative". After that you can edit as much as
you want and still have the original untainted - should you change
your mind.

By "adjusting" the image at the scanning stage using the Preview
"keyhole" you are just irretrievably damaging data coming from the
scanner. Something which is often impossible to correct later and
certainly not without additional loss of quality.

Don.

Don,

You do like to argue don't you? But you haven't seriously worked thru
Vuescan have you?
What you say at the beginning of this post is pure opinion, and then you say
some sensible afterwards which contradicts.

"Point and shoot"? Get real - whats your agenda?

Vuescan is exactly the tool to take the raw negative. Follow the basic
workflow and you can edit in your tool of choice later.

For some one using the 5400, its the Minolta scan utility which tries to
corrupt the image at scan time (if you allow it), not Vuescan. I don't know
about Nikonscan, perhaps you don't know about Vuescan, but don't get sold on
the name.

Vuescan allows for the maximum capture of dynamic range, if you control it
properly - then edit to get appropriate balance. I don't see this as point
and shoot.

Mark

 

[Toby]

It's a difference in design and target audience.

Both VueScan and SilverFast are "point-and-shoot" programs similar to
disposable cameras. OK for casual tourist but not for people who care
for quality or have an inkling of what they are doing. Nothing wrong
with that if that's what they are after, but...

What you are saying above is essentially: Hasselblad is garbage, Kodak
Instamatic is far superior - fixed focus, auto-exposure...

Not at all. Silverfast has extensive color correction capabilities if you
don't wish to use "negafix" for instance, or use it only as a place to start
further color corrections. The prescan adjustments I was referring to had to
do with optimizing luminance values. Adjusting exposure can get you a good
histogram going into PS. Adjusting afterwards, especially in 8 bit, will get
you a gappy histogram where many tonal values are missing. The key is having
scanning software that can optimize the performance of the hardware, to get
the most information possible into the final scan.

Forgive me, but that is just plain nonsense. Exactly the opposite is
true.

By doing image editing at the scanning stage you irreparably *corrupt*
the image at the earliest possible stage and with the crudest possible
"tools".

If, instead, you scan as faithfully as possible you will have the
equivalent of a "digital negative". After that you can edit as much as
you want and still have the original untainted - should you change
your mind.

The key phrase here is "as faithfully as possible". I'm not saying that one
should apply a lot of sharpening, for instance during the scan (although
Silverfast's sharpening is quite interesting and gives results unlike its PS
counterpart)--this clearly leads to loss of information. I'm saying, as you
are, that it is best to get the max digital information into the scan
itself, so that you have more to work with in PS. You wouldn't suggest not
adjusting scanner exposure and then trying to compensate for
under/overexposure in PS would you?

For instance, using "Curves" in Silverfast I can extract much more tonal
information than I can doing a "straight" scan. I get information in places
where no information would exist in a "raw" scan, and which no amount of
manipulation in PS is going to restore. I find your argument akin to saying
that I would get better quality by shooting with a 50 mm lens and enlarging
6x than by shooting with a 300 mm lens and not having to enlarge in the
darkroom.

Toby

 

[Hecate]

It's a difference in design and target audience.

Both VueScan and SilverFast are "point-and-shoot" programs similar to
disposable cameras. OK for casual tourist but not for people who care
for quality or have an inkling of what they are doing. Nothing wrong
with that if that's what they are after, but...

I'#d disagree with that. Silverfast SE may be for the casual user, but
Ai certainly isn't. And from what I saw of Vuescan, even though it
doesn't suit me, I wouldn't have called it simple.

 

[Toby]

Sorry for my very inaccurate language. Kennedy said it much better than I
did. David seems to assume that maker's s/w is the equal of some other
packages such as Silverfast in controlling the scanner hardware with the aim
of extracting the maximum information possible, or at least choosing the
range from which to extract maximum information. My experience (at least
insofar as Nikonscan goes) is that this is not the case.

I love my Hassy, but I also appreciate the autofocus and autoexposue options
of my Nikons in some situations. I wouldn't give up manual control for
strictly automatic, but having both is better yet.

Toby

I disagree. They certainly *can* be point and shoot packages in that
they will give very acceptable results out of the box, but they both
have a full range of adjustments, in some cases exceeding the range that
the native software provides, or which can only be produced by the
native software in conjunction with another image processing package
(eg. PS) and some effort and skill on the part of the operator.
Not exactly - analogue gain, for example, is an adjustment which can
only be made at the scanning stage and, when necessary, improves the
quality of the scan in a way that can never be recovered in Photoshop.

In such cases, failing to make the exposure adjustment at the scan or
prescan stage results in irretrievably lost image information. You only
have a limited number of bits on that ADC - it can never be argued that
it is better not to use them all and expect PS to make good what it
never gets.


Not in the context that Bruce was suggesting when he wrote: "Have you
tried the long exposure option in Vuescan? - that usually (but not
always) works for me in reducing slide shadow noise."

ie. the very situation where prescan adjustment is important.
--
Kennedy
Yes, Socrates himself is particularly missed;
A lovely little thinker, but a bugger when he's pissed.
Python Philosophers (replace 'nospam' with 'kennedym' when

replying)

 

[Toby]

I have been using AI. It is overkill for what I have been doing so far, and
I could have certainly gotten away with SE so far, although the CMYK output
could prove important for me. The main difference seems to be in AI's very
advanced color correction controls, batch control, CMYK and LAB output, and
a few other odds and ends, mostly relating to professional users.

Have a look here:

http://www.lasersoft-imaging.com/show/compare-ai-vs-se/en.html

I've never tried VueScan but I've heard very good things about it.

Toby

 

[Don]

I disagree. They certainly *can* be point and shoot packages in that
they will give very acceptable results out of the box, but they both
have a full range of adjustments, in some cases exceeding the range that
the native software provides, or which can only be produced by the
native software in conjunction with another image processing package
(eg. PS) and some effort and skill on the part of the operator.

That was not the context of the statement. It was the
"point-and-shoot" aspect which was claimed as superior. Besides,
saying "You lose a lot of quality" when scanning raw and editing in PS
is just sheer nonsense and I'm sure you must agree with that.

Not exactly - analogue gain, for example, is an adjustment which can
only be made at the scanning stage and, when necessary, improves the
quality of the scan in a way that can never be recovered in Photoshop.

Again, that is not the context. It's automated workflow with minimum
user intervention which is at discussion. The usual "point-and-shoot"
user can't even spell analog(ue) gain, let alone use it!

Hey, as a Canadian I should actually be using "analogue" spelling but
"analog" is shorter. ;-)

Not in the context that Bruce was suggesting when he wrote: "Have you
tried the long exposure option in Vuescan? - that usually (but not
always) works for me in reducing slide shadow noise."

ie. the very situation where prescan adjustment is important.

Yes, but that only proves my point. Not only is "long exposure" yet
another "point-and-shoot" feature but relying on auto-exposure based
on Preview is false security because it's based on a very small
sample.

Indeed, reputable software documentation warns that auto-exposure is
not totally reliable and - for optimum results - the user should
examine the scan and then fine tune auto-exposure with Analog Gain.

Don.

 

[Don]

Don,

You do like to argue don't you?

Huh!? Let's for the sake of the argument (sic) assume that to be
correct (which it is not) have you considered what would that make
your response?

But you haven't seriously worked thru
Vuescan have you?

Far more than I ever wanted to in order to prove challenges like this
wrong. Indeed, the reason you have individual Analog Gain control now
is, to a large extent, thanks to my objective and thorough analysis of
VueScan's output.

You're welcome... ;o)

"Point and shoot"? Get real - whats your agenda?

As a careful reader would have observed: None. May I reciprocate?

For some one using the 5400, its the Minolta scan utility which tries to
corrupt the image at scan time (if you allow it), not Vuescan. I don't know
about Nikonscan, perhaps you don't know about Vuescan, but don't get sold on
the name.

Well, if you followed this group carefully over the last year you
would've noticed that I have many issues with Nikon and pick no
favorites. Indeed, I myself often play a devil's advocate to my own
statements in order to get to the bottom of the matter.

Vuescan allows for the maximum capture of dynamic range, if you control it
properly - then edit to get appropriate balance. I don't see this as point
and shoot.

If you tabulate responses to questions about VueScan (and this
includes the author himself) you'd notice that it almost inevitably
involves a point-and-shoot "workflow":

1. Delete VueScan.ini (i.e. reset to factory default)
2. Set media to Image (i.e. set to neutral)
3. Preview
4. Right click on neutral area (i.e. set the gray point)
5. Scan

A very basic procedure, BTW, which applies to *all* scanning software
I'm aware of - even NikonScan - which means if that's how you want to
scan then *any* software will do!

Indeed, the only time I ever see raw scan mentioned in the context of
VueScan is for debugging purposes.

Sure, there are a few individuals who do use raw "properly" but that's
not VueScan's target audience. Its biggest selling point is exactly
the point-and-shoot aspect - as Toby's message clearly illustrates -
and the same goes for SilverFast.

Again, there's nothing wrong with that. But to go from that to "You
lose a lot of quality when scanning raw and editing in PS" is just
plain nonsense.

Don.

 

[Don]

Not at all. Silverfast has extensive color correction capabilities if you
don't wish to use "negafix" for instance, or use it only as a place to start
further color corrections. The prescan adjustments I was referring to had to
do with optimizing luminance values. Adjusting exposure can get you a good
histogram going into PS. Adjusting afterwards, especially in 8 bit, will get
you a gappy histogram where many tonal values are missing. The key is having
scanning software that can optimize the performance of the hardware, to get
the most information possible into the final scan.

What you don't seem to realize is that the output you're getting after
making color corrections in the scanning software is *not* the most
information possible. Indeed, it is information *corrupted* by the
scanning software color correction. This corruption simply *masks*
things which you consider a minus i.e. gappy histogram a.k.a. "comb
histogram".

The reason you get comb histograms is because that *is* the "most
data" *if* you let the scanning software do the gamma conversion. Set
gamma to 1.0 and comb histograms will disappear. That's the purest and
most unadulterated scan you can get!!

Now to prove to you that what you consider "good" (smooth histograms
at gamma 2.2) is actually *bad* try this in NikonScan:

1. All settings at neutral (gamma at 2.2!)
2. Crop as required
3. Scan

The result is a comb histogram image.

4. Go back to Crop
5. Set Scale to 99%
6. Scan

You'll notice that the scan takes longer but the result is a smooth
histogram! Oh, joy!

Not really... That smooth histogram is a *corruption*. The gaps were
filled *artificially* due to image interpolation by the scanning
software - which is why the scan took longer. The data used to fill
the gaps is *fantasy*, it's not real, it's a *guess* by the
interpolation algorithm.

Finally, take the first scan (the comb histogram one) and load it into
Photoshop:

7. Go to Image / Image Size...
8. Change size by a few pixels, set Resample to *Bicubic*
9. Click on OK

Bingo! Smooth histogram again. A miracle! Yeah, right.... But at what
costs? At the cost of image *corruption*! By scaling up or down you
have *corrupted* image data by *imaginary* data as explained above.

I'm sure even you will agree that scaling an image up or down in
Photoshop is a corruption. The same goes for scanner software color
correction. I hope that makes it clear now.

I won't reply to the rest because it's based on the same mistaken
notion as above.

Don.

 

[Don]

I'd disagree with that. Silverfast SE may be for the casual user, but
Ai certainly isn't. And from what I saw of Vuescan, even though it
doesn't suit me, I wouldn't have called it simple.

Actually, the messy user interface is one of VueScan's major problems,
which is why the author's panacea advice is to ignore everything and
just set the gray point.

But even VueScan can be used by a knowledgeable user to get good
results (again, by ignoring everything and just scanning raw).

However, all that was not the gist of the discussion. It's the notion
that point-and-shoot delivers purest data, which is patent nonsense.

Don.

 

[Don]

Thanks Don. I couldn't have responded better.

No, thank you! You obviously understand what I was saying.

Let's hope my subsequent explanation in a parallel message helps Toby
understand too. In the end, understanding this can only benefit even
if he choses to continue using the same workflow as before. At least
it will be a decision based on being fully aware of what is going on.

When editing by a scanner sw, you don't know what
information is lost and can never recover it. PS' editing tools are far
superior than any scanner sw can offer, but you do need to learn how to
operate them.

Yes, that's indeed *the* key and what I was trying to convey. Again,
let's hope this clear and concise paragraph can throw some more light.

Don.

 

[DavidTT]

No, thank you! You obviously understand what I was saying.

Let's hope my subsequent explanation in a parallel message helps Toby
understand too. In the end, understanding this can only benefit even
if he choses to continue using the same workflow as before. At least
it will be a decision based on being fully aware of what is going on.


Yes, that's indeed *the* key and what I was trying to convey. Again,
let's hope this clear and concise paragraph can throw some more light.

Don.

So it's you and I (one Canadian and one Yank) against the world. <g>
Actually, there are a few others:

This author ran a simple test. He concluded that tutorials like Scantips
to be incorrect and even recorded the great Ed Hamrick in agreement.
http://www.lumika.org/gear_nikon_scan_vs_photoshop.htm

Here's another author preferring raw scans for the same reasons we are
discussing. He pointed out that those who capture raw scan need to able
to stand ugly looking scans, and then be able to edit them in PS.
http://www.naturephotographers.net/mh0202-1.pdf (See "Input Method 2:
Manage Color After the Scan")

If the other side can only pause and think about what hardware is inside
a scanner, they may come to the conclusion that the limit hardware can
only do so much in capturing a scan and it is the scanner software that
is doing most of the work in creating a "best" scan. Here's one simple
way to realize this. To produce the first scan preview, note that the
scanner takes some time and makes some noise stepping through the scan.
Display the preview scan and its histogram. Make some corrections
(exposure, levels, curves, etc.). Note that the preview and the
histogram changes instantly and the scanner makes no noise. The software
is making the correction and not the hardware. As Don pointed out, the
better looking histogram after a correction is not from real pixels
rescanned by hardware, but is a result of the software's interpolation
of the original pixels. On my Polaroid ss4000, this is even more
obvious. When scanning a raw scan without any corrections, there is a
status window showing the scan progress. If I make corrections and then
scan, the same scan status window is followed by another status window
showing correction progress.

 

[openshutter]

While I have found the discussion of the merits of silverfast and
vuescan very interesting and informative, I'd like to bring this back to
the topic I originally posted:

I have been trying to gather information from people to help me decide
between the Minolta 5400 and the Nikon Coolscan 5000. I am interested
in knowing any pros or cons that people may like to share about these
scanners.

Thanks!

 

[Kennedy McEwen]

That was not the context of the statement.

Its certainly how I read the statement "it offers advanced controls..."

It was the
"point-and-shoot" aspect which was claimed as superior.

As you well know, for many users a "point-and-shoot" capability *will*
always produce superior results. That is why the "straight out of the
box" results from Vuescan and Silverfast have earned those packages a
substantial user base.

However I don't read the issue as one of point and shoot superiority,
but one of the assumption that 3rd party software is inferior - as Toby
pointed out, it isn't. Point and shoot facilities merely enhance their
ease of use, but there is no reason to assume that without this they
would be inferior to native software.

Also note that there is a difference between an optimised scan, the term
used by Toby, and a processed image - an optimised scan can still be
raw, for example.

I seem to recall only a few months back a long tirade from your very
good self about the inability of native scanner software to achieve just
such an optimised scan - so what makes it so different now?

Again, that is not the context. It's automated workflow with minimum
user intervention which is at discussion.

That is why I said "not exactly" in response to your comment that
*exactly* the opposite was true. In general, the opposite is true, but
not in every case, as my example demonstrates, so "not exactly the
opposite" is true. I consider this to be as "in context" as any other
comment in the thread and more than many.

 

[WD]

I have used a Coolscan 5000. Excellent results (barring software
issues).
I too faced the decision between the Minolta 5400 and Coolscan 5000.
The best way to decide would be to get your hands on both and try a
number
of different scans. I did not do that. I chose the Nikon for the
following reasons:

- Very fast scanning
- This line of scanners historically serves the professional
market at
the level below drum scanning.
- Had a Minolta Dual Scan III and was not happy with
DMax (I would GUESS that the 5400 to beat the Dual Scan
in these respects)

While I do envy the higher resolution of the Minolta, I felt that I
could
not go wrong with the Nikon given its lineage. The Minolta does not
have this kind of track record. It may be superior, but I did not want
to 'risk' it.

After using the Coolscan 5000 for a while I can say that I am quite
satisfied.
I can also say that it is still beat by a good drum scan. One must
expect
this given the price and technology difference.
I scan mostly negatives. I had a negative that I had had drummed
scanned that
I scanned with the Nikon. The Nikon scan was a bit more noisy/grainy.
But more than that, it did not have quite the punch or vibrancy of
the drum scan. It was close however.

After having scanned the same negatives I had scanned with the Dual
Scan III,
The Nikon is clearly superior in resolution (the Dual scan ~ 2800 dpi)
as well as color and 'punch'.

I must also say, I still wonder how I would have made out if I chose
the Minolta 5400. It may well be better than the Coolscan. Although
the
long scan times would have been a real bummer. Like anything else,
when it takes a very long time to complete a scan one tends not to
want
to spend alot of time experimenting to get better scans. That is sort
of a 'non-tangible' disadvantage of long scan times.

Also the GEM grain removal can be very useful. My take is that if you
have a negative that is very sharp and pushes the film to the limit,
than GEM will also soften detail. However, if your negative is
'average', then GEM can seem to reduce grain with virtually no loss
of detail allowing further sharpening. So GEM can be at times VERY
useful.

ICE works very well.

 

[false_dmitrii]

It may be what you have picked up, but it doesn't appear to hold any
water. If anything, the near collimated light source of the Nikon

<snipped explanation>

I can't argue with that.

I really can't! Glad I stopped where I did.

I don't know whether the error was in the post I read or my
misinterpretation of it. I'm still pretty sure the post claimed the
LS-5000 showed more of the halos than the SE5400...but that's all I
can contribute.

Regards,
false_dmitrii

 

[false_dmitrii]

It did happen, with early versions of the software (1.1 and 1.2) as
far as I know it was cured from version 1.3 onwards.

I'd read that, but doesn't the newsgroup still get complaints about
bands in shadows from time to time? Or is this just a false
impression I've developed?

false_dmitrii

 

[Don]

As you well know, for many users a "point-and-shoot" capability *will*
always produce superior results. That is why the "straight out of the
box" results from Vuescan and Silverfast have earned those packages a
substantial user base.

Actually, I don't agree with that on at least two grounds. In order to
achieve these results (new) users are advised to ignore all features
and just set the gray point. (Witness Ed's advice to *turn-off* the
Kodachrome setting in VueScan and use the gray point instead!!!) Well,
you can do that just as easily with virtually *any* scanner software
and NikonScan's user interface is vastly superior to VueScan's.
Actually, to be accurate, (unlike VueScan) NikonScan *has* a user
interface.

Second, superior to what? To their inability to use the software
properly, *not* superior to what the scanner can produce. Indeed,
their "point-and-shoot" meddling will only corrupt what the scanner
could deliver under optimal conditions. A new coat of paint does not
improve the structural weaknesses of a building, it only masks it to a
lesser or greater degree. Would you call that coat of paint superior
to the actual building work?

Which neatly gets us off this digression and brings us back to the
subject matter...

The key is his mistaken notion:

I prefer to scan raw and edit in Photoshop.

You lose a lot of quality this way.

You don't! You're just getting what the scanner is capable of, and you
get it unadulterated.

Until he comprehends this all else will go over his head.

However I don't read the issue as one of point and shoot superiority,
but one of the assumption that 3rd party software is inferior - as Toby
pointed out, it isn't. Point and shoot facilities merely enhance their
ease of use, but there is no reason to assume that without this they
would be inferior to native software.

I never said it was inferior and that's not at discussion. The problem
is all of his "conclusions" are based on a wrong premise (see above).

Furthermore, he can't tell the difference between "better" as in
getting more out of hardware (his original claim above to which I
responded) and "better" in the sense of making it easier for the user
at the expense of real quality (your parallel point which just
confuses him into thinking that, somehow, user-friendliness directly
translates into superior hardware output quality).

So, before we descent into such esoterics (which will just confuse him
even more) we have to make it clear to him that the above notion as it
relates to getting the most out of hardware is just plain wrong.

Otherwise, he will (and does!) misinterpret your statements (relating
to usage) as tacit support of his mistaken notion (relating to
hardware output quality) and that will just reinforce it and make it
that much more difficult to correct later.

So, for starters, you should maybe distance yourself clearly from his
above statement (without equivocation!) to help him understand that.
Once he does, and is willing to learn more, *then* it's the time for
"yes, but...", "except on Friday 13th and full moon..." etc.

I seem to recall only a few months back a long tirade from your very
good self about the inability of native scanner software to achieve just
such an optimised scan - so what makes it so different now?

Nothing! No amount of curves adjustments will ever bring back the
missing dynamic range. It will just make it worse. (The case of coat
of paint hiding structural deficiencies.) This was clear to me even
before I started. So...

I relied on facts, I ran tests, I read, I learned (all good advice to
him as well) and then achieved exactly what I said was possible all
along with Analog Gain because it was a simple yet irrefutable axiom.
There *is* a global setting for scanning Kodachromes with LS-30.
However, it's a relative one. Specifically, it's a function of
exposure.

I never said native scanner software was incapable. Quite the
contrary! I said Nikon was *unwilling*. The software *is* capable
(using Analog Gain) but Nikon refused to retrofit or, failing that,
supply the correct settings - i.e. workflow. That was the problem.

Don.

 

[Don]

So it's you and I (one Canadian and one Yank) against the world. <g>

Yeah! Bring it on!

Juuust kidding... ;o)

Actually, there are a few others:

Yes, pretty much everybody agrees because it's such a basic and
elementary notion.

Here's one simple
way to realize this. To produce the first scan preview, note that the
scanner takes some time and makes some noise stepping through the scan.
Display the preview scan and its histogram. Make some corrections
(exposure, levels, curves, etc.). Note that the preview and the
histogram changes instantly and the scanner makes no noise. The software
is making the correction and not the hardware.

Exactly! What's more, these corrections are based on the Preview which
is only a small "keyhole" sample of the full image.

Nikon actually warns about this (hidden deep within the manual) as it
relates to auto-exposure, where they indicate that for best results
the full scan should be examined and optional fine adjustments made
using Analog Gain.

Don.