Problem scanning very dark Kodachrome

[Don]

As written in my former posting, I got a scan as good (or even better)
as your NikonScan in just 1-2 minutes. No need to tweak red/green
channels, almost the standard settings work.

The reason I didn't respond is because that's a subjective judgement.

If you examined the histograms objectively (like Erik did) you
would've observed that the dynamic range is just not there for a
meaningful edit without undesired artifacts (posterization).

But I'm also interested in Ed's comment, as I also have a LS-30 and
need to scan heaps of older family slides (Kodachrome) in the near
future.

If you're happy with the results you got above, then I'd say you'll be
OK.

For us "perfectionists" ;o) who examine the histograms and don't like
posterization, it's just not good enough.

Don.

 

[Don]

I got the file - thanks.

You're welcome!

The problem is that your scanner LED brightnesses
are significantly different from the LED brightness
on the scanner I used to calibrate VueScan.

I don't think that's the case because the "good" scan I posted is just
fine without *any* VueScan adjustment whatsoever:

http://members.aol.com/tempdon100164833/VS/VS-OK.jpg

That's precisely why I posted the scan to establish this baseline.

If there were significant differences in LED brightness (or LED
spectral characteristic, or whatever...) then the above scan would
also show some sort of aberration - and it doesn't.

You can fix this by getting an IT8 calibration target
and calibrating your scanner. This should solve the
problem entirely.

Again, I don't think so because calibration is done is software. In
theory, calibration can also be emulated manually in Photoshop by
saving a Curves or Levels setting. All calibration does is automate
this repetitive portion of the task since this constant bias is known.

So the problem remains, whether one calibrates or uses Photoshop,
there just isn't enough dynamic range for the edit without invoking
all sorts of ugly artifacts.

Note that using a raw scan file and setting
"Input|Scan mode" to "Transparency" produces a
perfectly normal looking image of snow, once you
click with the right mouse button on a neutral
color.

Can you give the specific X/Y coordinates where you clicked?

Also, could you post your results so we can examine them.

Don.

 

[Don]

Hello, Ed Hamrick
you wrote...


But did you see where the white point values go (or film base color, if
you used media type slide film)?


Sure? Does the IT8 profile change LED brightness? Than a raw scan would
look differently if I set IT8 profile in color tab? I thought profiles
are used for a software correction only?

I assume that you looked at the histogram of the raw file...

Yeah! What he said! ;o)

Don.

 

[Ed Hamrick]

Can you give the specific X/Y coordinates where you clicked?

I tried 3 or 4 positions - all worked fine. The key is
setting "Input|Scan mode" too "Transparency".

Also, could you post your results so we can examine them.

What did your last slave die of?

Regards,
Ed Hamrick

 

[Bart van der Wolf]

SNIP

I tried 3 or 4 positions - all worked fine. The key is
setting "Input|Scan mode" too "Transparency".

Yes, I had the same impression (the LS30 mode doesn't work for Don's
scanner,but Transparency on a Raw file is okay). With so little data to go
on, it will never be 'great', but it'll provide a decent basis for further
tweaking. Important to remember is that most slide films have aweful color
balance in the D-max area, so no matter how accurate the scan, the colors
will be a challenge on dense slides.

What did your last slave die of?

LOL

Bart

 

[Thomas F. Unke]

I tried 3 or 4 positions - all worked fine. The key is
setting "Input|Scan mode" too "Transparency".

I even didn't right click (learned something new here). But this way I
get the same or similar result as I have posted before.

A question: it seems that a direct scan (without intermediate raw
file) does not lead to that good results. I think that was the major
problem Don has had. So I assume as scan via raw files is *the* way to
go generally, not only in difficult cases.

 

[Ed Hamrick]

A question: it seems that a direct scan (without intermediate raw
file) does not lead to that good results. I think that was the major
problem Don has had. So I assume as scan via raw files is *the* way to
go generally, not only in difficult cases.

A better way to solve this would be calibrating the scanner
with an IT8 calibration slide.

Regards,
Ed Hamrick

 

[Charlie]

A better way to solve this would be calibrating the scanner
with an IT8 calibration slide.

Regards,
Ed Hamrick

Could someone post a source (and the cost) of such a calibration
slide?

Charlie Hoffpauir
http://freepages.genealogy.rootsweb.com/~charlieh/

 

[Gerhard Fuernkranz]

Could someone post a source (and the cost) of such a calibration slide?

For instance http://targets.coloraid.de/

Regards,
Gerhard

 

[Thomas F. Unke]

For instance http://targets.coloraid.de/

Perhaps a stupid question: Do you scanner experts here always
calibrate your scanner with a calibration slide?

 

[Don]

I tried 3 or 4 positions - all worked fine.

That is the canonical definition of "evasive".

Not to mention numerous other specific questions you failed to answer.

What did your last slave die of?

I beg your pardon?

After making all sorts of petty demands (all of which I fulfilled) you
can't even be bothered to post a single image (and name the neutral
X/Y point) to defend your unsubstantiated claims!?!?

It's a predictable pattern, though, repeated throughout. Every time it
comes to specifics (and you run out of places to hide) you have an
arrogant outburst - like the one above - and then go and stew
silently.

Therefore, the original assertion remains: VueScan's lack of
individual Analog Gain control (one of its many shortcomings) makes it
unsuitable for serious use under less than optimal conditions or by
discerning users.

Don.

 

[Charlie]

On Wed, 05 May 2004 14:29:13 GMT, (e-mail address removed) (Don) wrote:

Therefore, the original assertion remains: VueScan's lack of
individual Analog Gain control (one of its many shortcomings) makes it
unsuitable for serious use under less than optimal conditions or by
discerning users.

Don.

Enough.
<plonk>

Charlie Hoffpauir
http://freepages.genealogy.rootsweb.com/~charlieh/

 

[Ed Hamrick]

That is the canonical definition of "evasive".

I didn't write down the positions.

It's a predictable pattern, though, repeated throughout. Every time it
comes to specifics (and you run out of places to hide) you have an
arrogant outburst - like the one above - and then go and stew
silently.

I'm actually fairly busy getting VueScan 8.0 ready.

Regards,
Ed Hamrick

 

[RSD99]

To: "Don"
GET A LIFE ... one that involves something other than the baseless harassment of Ed
Hamrick and/or the senseless denigrating his software.

He has obviously done everything he can to try to "satisfy" you (a non-paying customer),
and you have done everything possible to argue with him and evade being satisfied! Strange
.... isn't it ... that everyone else seems to be able to make Viewscan work ...

To: "Ed Hamrick"
You have "gone the extra mile" ... now refund what "Don" has actually paid to you for your
software [ ;-)) ], and send him on his way as "An Unqualified Customer."

[S-CNR]

 

[Thomas F. Unke]

That is the canonical definition of "evasive".

Not to mention numerous other specific questions you failed to answer.

It is not evasive and there is no need to answer this. You could just
try it yourself if you'd be interested in a real solution. Ed and me
have posted simple ways of getting a proper scan from your raw
file. Ed has also suggested to calibrate your scanner.

Therefore, the original assertion remains: VueScan's lack of
individual Analog Gain control (one of its many shortcomings) makes it
unsuitable for serious use under less than optimal conditions or by
discerning users.

You are AFAIK the only person who has this assertion. There are
numerous people who find vuescan very suitable for serious use.

 

[David Blanchard]

A better way to solve this would be calibrating the scanner
with an IT8 calibration slide.

Regards,
Ed Hamrick

I continue to follow this thread because I have a Nikon LS-2000 and 20
years worth of Kodachromes...some of which are very difficult to scan
with either VueScan or NikonScan. By the way, newer versions of
VueScan no longer support Macintosh OS 9 (only OX X), and newer
versions of NikonScan do not work with my scanner or OS, so I'm
"stuck" with older versions. Having said all that in preface...

I have tried to understand the IT8 calibration concept and visited
many of the IT8 sites where targets can be purchased. What confuses
me is that they all mention a specific film type as if to imply that
each film type used will need its own specific target. Is this really
the case? If so, it could be a problem because I have not found any
sites that sell Kodachrome targets. If they do exist, pointers
appreciated.

Finally, I have used both VS and NS extensively on my slides and found
that NS gives me more control (i.e., analog gain for all three
channels) to scan the more difficult slides. For the easier to scan
slides, VS gives me better color results. My experience, of course,
may be different from yours...

David

 

[degrub]

Try B&H photo. THey have KC targets.
URL:
http://tinyurl.com/27hyd

Helped quite a bit to get the output closer to the real thing. You will
still have to tweak some of the images. Watch out for KC with varnish on
it, 1950s and earlier. It can be difficult to scan as it will not match
any target or Vuescan profile.

Frank

 

[Ed Hamrick]

send him on his way as "An Unqualified Customer."

I've found in the past that when a customer (or potential
customer) is adamant about something, they often have a
real point.

I've added separate control over the analog gain to my
list of things to do. I may even get this into 8.0.

Regards,
Ed Hamrick

 

[David J. Littleboy]

I've found in the past that when a customer (or potential
customer) is adamant about something, they often have a
real point.

Man, you've got the patience of a saint. I'd have gone postal ages ago.

I've added separate control over the analog gain to my
list of things to do. I may even get this into 8.0.

QED.

David J. Littleboy
Tokyo, Japan

 

[RSD99]

"David J. Littleboy" posted:
"... Man, you've got the patience of a saint. I'd have gone postal ages ago. ..."


Yes!