Nikon - Kodachrome mess
Don
I need a couple of sets from a Nikon film scanner claiming to do
Kodachrome slides (I believe 8000 is one of them).
Having initialized all Nikon Scan settings to neutral (no analog gain,
flat curves, etc) please save one of each:
1. a Positive set
2. a Positive - Kodachrome set
Here's a temporary email address to send it to:
It's a lot to ask, I know, but Nikon is no help and I will never buy a
Nikon product again...
Thanks very much in advance.
Don.
ThomasH
>
> Hi,
>
> I need a couple of sets from a Nikon film scanner claiming to do
> Kodachrome slides (I believe 8000 is one of them).
>
> Having initialized all Nikon Scan settings to neutral (no analog gain,
> flat curves, etc) please save one of each:
> 1. a Positive set
> 2. a Positive - Kodachrome set
>
> Here's a temporary email address to send it to:
>
> tempd...@yahoo.com
>
As much I would love to help, please explain more what are
the (1.) and (2.) settings to safe?? The setting file itself,
scan result? I am not sure what precisely is the problem with
your Kodachrome scan. Every scan with NikonScan is a problem
is you use NCM (Nikon Color Management,) including the brand
new NS 4.01. Disable NCM, or use ScannerRGB as output and use
Vuescan to process the data.
Thomas
Don
>Don wrote:
>>
>> Hi,
>>
>> I need a couple of sets from a Nikon film scanner claiming to do
>> Kodachrome slides (I believe 8000 is one of them).
>>
>> Having initialized all Nikon Scan settings to neutral (no analog gain,
>> flat curves, etc) please save one of each:
>> 1. a Positive set
>> 2. a Positive - Kodachrome set
>>
>> Here's a temporary email address to send it to:
>>
>> tempd...@yahoo.com
>>
>
>As much I would love to help, please explain more what are
>the (1.) and (2.) settings to safe?? The setting file itself,
>scan result? I am not sure what precisely is the problem with
>your Kodachrome scan. Every scan with NikonScan is a problem
>is you use NCM (Nikon Color Management,) including the brand
>new NS 4.01. Disable NCM, or use ScannerRGB as output and use
>Vuescan to process the data.
>
>Thomas
I'm referring to the setting file itself.
Basically, disable all color management, all analog gain at 0, flat
curves, etc...
Please exit and restart NikonScan so these settings take effect.
After that select Positive as the source and save the settings as
"Pos.set", for example.
Change source to Positive - Kodachrome and save that as "Kodak.set",
for example.
I don't have an 8000 so I don't know what the actual names are but
it's something like "Positive" and "Positive - Kodachrome".
I have an old LS-30 and it just doesn't do Kodachrome so I'm trying to
get the best results I can. You're right about NCM! Another hint is to
disable Digital ICE, which also doesn't work with Kodachrome!
Thanks very much in advance Thomas! Much appreciated!
Don.
degrub
severe problems with KC,i had one. i upgraded to a CS4000 as a result.
Frank
nathantw
works with Kodachrome.
"Don" <phoney...@yahoo.com> wrote in message
news:3fd402e2...@news.individual.net...
Erik Krause
you wrote...
> I have an old LS-30 and it just doesn't do Kodachrome so I'm trying to
> get the best results I can. You're right about NCM! Another hint is to
> disable Digital ICE, which also doesn't work with Kodachrome!
Did you try vuescan? It works very good with my LS 40 and Kodachromes.
Even infrared cleaning works to some extent.
--
Erik Krause
Digital contrast problems: http://www.erik-krause.de/contrast
Don
blue cast. I corrected them both but it's very time consuming as well
as hit and miss.
Instead, I want an objective, "official" compensation I can apply
across the board and then fine tune later. That's why I'm asking for
two sets from someone with an 8000 or 4000 (one vanilla positive, the
other Kodachrome positive) to compare them and then try to emulate it
in my LS-30.
Don.
On Tue, 09 Dec 2003 05:04:34 GMT, degrub <deg...@nospamflash.net>
wrote:
Don
wrote:
>Hello, Don
>you wrote...
>
>> I have an old LS-30 and it just doesn't do Kodachrome so I'm trying to
>> get the best results I can. You're right about NCM! Another hint is to
>> disable Digital ICE, which also doesn't work with Kodachrome!
>
>Did you try vuescan? It works very good with my LS 40 and Kodachromes.
>Even infrared cleaning works to some extent.
>
>--
Hi Erik,
Yes, I have, but I find VueScan's user interface very confusing. The
resulting scans are comparable to what I have achieved but I want
official values from Nikon.
From what I understand LS-40 has a Kodachrome mode so that's why
you're getting good results. That why I want a set from an LS-40 or
the 8000 with Kodachrome mode selected to see what Nikon does to
compensate for Kodachrome.
Granted, more recent scanners may draw on the hardware, but if they
don't I should be able to emulate (retrofit) the set to my LS-30.
Don.
Don
Anyway, none of my hard earned cash will end up at Nikon ever again
even if they advertise the 9000 will also do my laundry and vacuum my
apartment... ;o)
Don.
Mike Engles
Hello
It cannot be done. The LS 30 really struggles with Kodachromes.
The LS 40 is better,but ICE is not to be recommended. it is very hit and
miss withe Kodachromes
Mike Engles
Erik Krause
you wrote...
> Yes, I have, but I find VueScan's user interface very confusing.
You'll learn it if you have to. In the end it's quite logical ;-) I
found NikonScan rather confusing, especially since I did not find how
to disable any automatics...
> The
> resulting scans are comparable to what I have achieved but I want
> official values from Nikon.
>
> From what I understand LS-40 has a Kodachrome mode so that's why
> you're getting good results.
Kodachrome mode is pure software. Vuescan is able to save raw scan data
to disk and do any processing later. This includes Kodachrome typical
color correction. For the hardware there is only slide and negative.
I would prefer Vuescan, since ED Hamrick managed to get IR-cleaning
working with Kodachromes.
If your LS-30 is able to manage the high maximum density of Kodachrome
there is nothing against good scans from Kodachrome.
> That why I want a set from an LS-40 or
> the 8000 with Kodachrome mode selected to see what Nikon does to
> compensate for Kodachrome.
I'll unpack my scanner tomorrow (I have moved). If you don't have your
setting files then, ask me per mail.
Jim Norris
with no problems whatsoever. Well, it didn't clean up the fungus spots very
well..
I used Nikonscan for this purpose. While Nikonscan and Vuescan have quite
different user interfaces, I don't have any trouble getting either one to do
my bidding.
I used to Kodachrome slide setting most of the time, Once, I did by
accident use the E6 setting, but the result was a dark and greenish scan.
Jim
Don
blue cast. I corrected them both but it's very time consuming as well
as hit and miss.
Instead, I want an objective, "official" compensation I can apply
across the board and then fine tune later. That's why I'm asking for
two sets from someone with an 8000 or 4000 (one vanilla positive, the
other Kodachrome positive) to compare them and then try to emulate it
in my LS-30.
Don.
On Tue, 09 Dec 2003 05:04:34 GMT, degrub <deg...@nospamflash.net>
wrote:
>What problems are you having with Kodachromes ? i know the LS-20 had
Don
wrote:
>Hello, Don
>you wrote...
>
>> I have an old LS-30 and it just doesn't do Kodachrome so I'm trying to
>> get the best results I can. You're right about NCM! Another hint is to
>> disable Digital ICE, which also doesn't work with Kodachrome!
>
>Did you try vuescan? It works very good with my LS 40 and Kodachromes.
>Even infrared cleaning works to some extent.
>
>--
Hi Erik,
Yes, I have, but I find VueScan's user interface very confusing. The
resulting scans are comparable to what I have achieved but I want
official values from Nikon.
From what I understand LS-40 has a Kodachrome mode so that's why
you're getting good results. That why I want a set from an LS-40 or
the 8000 with Kodachrome mode selected to see what Nikon does to
compensate for Kodachrome.
Granted, more recent scanners may draw on the hardware, but if they
Don
Anyway, none of my hard earned cash will end up at Nikon ever again
even if they advertise the 9000 will also do my laundry and vacuum my
apartment... ;o)
Don.
>Hate to tell you, but the new 9000 that's coming out soon advertises it
Don
<mike.s...@btinternet.com> wrote:
>> >Don wrote:
>> >> I'm referring to the setting file itself.
>> >>
>> >> Basically, disable all color management, all analog gain at 0, flat
>> >> curves, etc...
>> >> Please exit and restart NikonScan so these settings take effect.
>> >>
>> >> After that select Positive as the source and save the settings as
>> >> "Pos.set", for example.
>> >> Change source to Positive - Kodachrome and save that as "Kodak.set",
>> >> for example.
>> >> I don't have an 8000 so I don't know what the actual names are but
>> >> it's something like "Positive" and "Positive - Kodachrome".
>> >>
>> >> I have an old LS-30 and it just doesn't do Kodachrome so I'm trying to
>> >> get the best results I can. You're right about NCM! Another hint is to
>> >> disable Digital ICE, which also doesn't work with Kodachrome!
>> >>
>> >> Thanks very much in advance Thomas! Much appreciated!
>> >>
>> >> Don.
>> >
>
>
>Hello
>It cannot be done. The LS 30 really struggles with Kodachromes.
>The LS 40 is better,but ICE is not to be recommended. it is very hit and
>miss withe Kodachromes
>
>Mike Engles
Hi Mike,
In my case it has been miss all the way...
Question: What is it about LS-30 specifically that inherently makes
Kodachrome scans impossible?
I would have thought correct analog gain setting could compensate for
both the dense film and, most of all, the putrid blue cast?
Sure, a 10-bit LS-30 can never compete with an 8000 but there exists
an optimum Kodachrome LS-30 setting given it's limitations. And a
company that makes it should know!!!
Instead Nikon blames everything from my version of Windows to a burned
out LED in spite of my saying up front that negatives scan just fine.
I will certainly never buy a Nikon product again!
Don.
degrub
Send me an email direct(take out the obvious) and i will send you a IT8
profile i use with my CS4000. It might help you get closer.
Frank
Don
>I found NikonScan rather confusing, especially since I did not find how
>to disable any automatics...
Yes, that's something I always do first (disable any automation).
Indeed, any time I read how wonderful and fantastic some new feature
is, my first question is inevitably "How do I disable it..." >:o)
>> From what I understand LS-40 has a Kodachrome mode so that's why
>> you're getting good results.
>
>Kodachrome mode is pure software.
Exactly!!! (Or at least, under software control, as in "analog gain".)
This is another reason why Nikon should be criticized for not
retrofitting Kodachrome mode on older scanners. I see no technical
reason, only marketing nonsense...
>If your LS-30 is able to manage the high maximum density of Kodachrome
>there is nothing against good scans from Kodachrome.
>
>> That why I want a set from an LS-40 or
>> the 8000 with Kodachrome mode selected to see what Nikon does to
>> compensate for Kodachrome.
>
>I'll unpack my scanner tomorrow (I have moved). If you don't have your
>setting files then, ask me per mail.
OK, will do.
Danke sehr! ;o)
Don.
Mike Engles
Hello
The Kodachrome emulsion is really very dense. If you have a slide that
is correctly exposed, or slightly overexposed, scanning it is not to
bad. There was a necessity for slide films to be slightly
underexposed,to maitain saturation. This kind of Kodachrome is so
variable, that it is almost impossible to get a generic setting.
Most of my slides are like this. My LS 30 really struggled. The LS 40 is
much better. The New LS 50 or LS 5000,should be much better again.
I did develope with the LS 30 a multiscanning technique. This involved
doing 16 scans and putting them together in Photoshop using layer
transparency and the Lighten and Darken modes. Very time consuming and
again the results very variable.
I then had a system of doing two scans, with and without ICE.
I used the clean one only where there was dirt.
Nikon is fine with Ektachromes. Kodachromes are very difficult.
I have fiddled with nearly all scanners for film. I always go back to
Nikon.
Mike Engles
Don
Hi Mike,
There are certain things which are linear, such as the notorious blue
cast and the overall darkness of the scan. For that there are generic
settings since Nikon implemented them on all scanners which do have
the Kodachrome option.
And that's all I want to do: Implement on my LS-30 whatever Kodachrome
setting does on all Nikon scanners which do have that setting! I can
then take it from there.
Now, is that too much to ask? Apparently it is...
So I'm relegated to begging on the Net for these set files because
Nikon is of no help whatsoever, preferring to obfuscate instead.
To my mind, that's just plain disgusting.
Any half-decent company would put that in a FAQ, offer the files on
their site and apologize for not providing them with the scanner.
What does Nikon do? Blames my version of Windows... :-/
Don.
Ed Hamrick
> And that's all I want to do: Implement on my LS-30 whatever Kodachrome
> setting does on all Nikon scanners which do have that setting! I can
> then take it from there.
I think you're looking for something that doesn't exist.
It sounds like you're looking for a sensitometric curve for
Kodachrome along with dye sensitivity curves as a function
of wavelength and dye colors as a function of wavelength.
If this is what you're looking for, you'll never find it
without producing it with specialized equipment.
It's not a simple matter of typing a few numbers into NikonScan
on the LS-30. If it were as simple as that, Nikon would have
done it years ago.
Regards,
Ed Hamrick
Mike Engles
Hello
The thing is that in my opinion it is not linear. It is very dependent
on the original exposure density of the image. There is something non
linear in the way transmitted light and Kodachromes react.
I have not found two similar Kodachromes.
Ekatchromes are not a problem. When I printed to Cibachrome, Kodachrome
was a pain, because of the density problem. I used it because there was
nothing sharper than Km25 for macro photography.
Mike Engles
Don
wrote:
>> And that's all I want to do: Implement on my LS-30 whatever Kodachrome
>> setting does on all Nikon scanners which do have that setting! I can
>> then take it from there.
>
>I think you're looking for something that doesn't exist.
>
>It sounds like you're looking for a sensitometric curve for
>Kodachrome along with dye sensitivity curves as a function
>of wavelength and dye colors as a function of wavelength.
I wouldn't have put it that way, but yes, I guess that's all I'm
asking.
Or to put it simply: "Implement on my LS-30 whatever Kodachrome
setting does on all Nikon scanners which do have that option!"
The fundamental technology is the same in both so the solution is
clearly in software. Fine, there may only be 10 bits on the LS-30 but
there exists an *objective* and *optimal* Kodachrome setting for LS-30
and Nikon knows what it is.
>If this is what you're looking for, you'll never find it
>without producing it with specialized equipment.
Exactly! Which is why I asked the manufacturer who does have all the
equipment and the know-how but who, instead of helping, obfuscated for
3 months at first denying that the problem exists and then produced a
"solution" by dragging the curve up by a notch. Blue cast still
there...
Gee, thanks! I produced far superior results myself with one LED tied
behind my back. Some support!
>It's not a simple matter of typing a few numbers into NikonScan
>on the LS-30. If it were as simple as that, Nikon would have
>done it years ago.
I beg to differ on two counts:
1. You have done it in ViewScan! (I take it you're "the" Ed Hamrick.)
If you can do it, surely a giga-conglomerate like Nikon can do it.
2. Nikon doesn't want to do it - for marketing reasons (to force
people to throw them more money by upgrading).
No thanks! I'll never buy a Nikon product again, and I will strongly
advise anyone who asks not to either.
Don.
Don
<mike.s...@btinternet.com> wrote:
Hi Mike,
Yes, my tests have also confirmed that some things are not linear. I
have scanned a Kodakchrome slide with LS-30 using neutral settings.
The result is a dark scan with a heavy blue cast. Unusable! I have
then taken a picture of the same slide with my digital camera. Perfect
color reproduction! I have then sampled one area of color in the
digicam picture and adjusted the same area in the LS-30 scan until it
was the same color as digicam color. However, this threw the rest of
the scan totally out of whack. So, the adjustment is not linear but a
bit more complicated.
However, it is undeniable that Nikon has found a solution, and it's
called the "Kodachrome" option. The solution is clearly in software
because subsequent scanners do not differ fundamentally from LS-30.
Their A/D converters may have a few extra bits but the technology is
essentially the same. VueScan also proves that it can be done.
So, the only logical conclusion is that Nikon doesn't want to do it.
I deeply resent the way Nikon goes about this, by first denying that
the problem exists and then not coming up with a solution (which they
have!) for marketing reasons.
Like I said, I will never buy Nikon again and strongly advise everyone
else not to. You may be happy now, but if there's a problem, you're
looking at months of frustration and expense, like I have, and you'll
be left hanging in the end.
Don.
Ed Hamrick
> Yes, my tests have also confirmed that some things are not linear. I
> have scanned a Kodakchrome slide with LS-30 using neutral settings.
> The result is a dark scan with a heavy blue cast. Unusable!
There's something seriously wrong with the way you have NikonScan
set up, since I've never seen this happen with NikonScan, Kodachrome
and my LS-30.
Try removing NikonScan and re-installing it.
> I deeply resent the way Nikon goes about this, by first denying that
> the problem exists and then not coming up with a solution (which they
> have!) for marketing reasons.
The problem doesn't exist.
The only Kodachrome problem with the LS-30 is when ICE is
used. The older version of ICE used with the LS-30 doesn't
properly handle Kodachrome. However, the image quality is
otherwise quite good.
> Like I said, I will never buy Nikon again and strongly advise everyone
> else not to. You may be happy now, but if there's a problem, you're
> looking at months of frustration and expense, like I have, and you'll
> be left hanging in the end.
Nikon makes excellent scanners, but like most consumer products
companies, they can't provide the kind of detailed technical
support that you're apparently looking for.
The kind of technical detail you're looking for could barely
be provided by the engineers that designed the scanner in the
first place, and there are probably only one or two people at
ASF who could provide this kind of technical support.
Regards,
Ed Hamrick
Mike Engles
Hello
I have some links for you.
http://www.btinternet.com/~mike.engles/mike/nikon.jpg
http://www.btinternet.com/~mike.engles/mike/nkn00003.set
http://www.btinternet.com/~mike.engles/mike/nkn00004.set
The first is a screen grab.
The other two are settings using NO Kodachrome and Kodachrome, in
Nikonscan 4. I just used the midtone sampler to set midtone curves for
neutral grey.
I am not sure how much use they will be to you.
I think that the Kodachrome setting is a non user retrievable setting.
It could even be built into the hardware of my LS 40.
It is not shown by a change in the analog gain settings, or in any other
setting that I can find.
The LS 40/LS4000 are much better scanners are will be the LS50 and LS
5000. The LS 30 is really working at it's limits. It just does not get
on well with Kodachromes. Very few do!
Good luck.
Kennedy McEwen
<phoney...@yahoo.com> writes
>On Fri, 12 Dec 2003 12:11:59 -0700, "Ed Hamrick" <use...@hamrick.com>
>wrote:
>
>>> And that's all I want to do: Implement on my LS-30 whatever Kodachrome
>>> setting does on all Nikon scanners which do have that setting! I can
>>> then take it from there.
>>
>>I think you're looking for something that doesn't exist.
>>
>>It sounds like you're looking for a sensitometric curve for
>>Kodachrome along with dye sensitivity curves as a function
>>of wavelength and dye colors as a function of wavelength.
>
>I wouldn't have put it that way, but yes, I guess that's all I'm
>asking.
>
>Or to put it simply: "Implement on my LS-30 whatever Kodachrome
>setting does on all Nikon scanners which do have that option!"
>
>The fundamental technology is the same in both so the solution is
>clearly in software.
So far, in this thread, you have not produced a single iota of evidence
which supports that claim.
The single biggest difference between how the the LS-30 and subsequent
scanners handle Kodachrome is with ICE. What evidence do you have that
this is entirely software? Do you know, for a fact, that Nikon have not
used a slightly different wavelength of IR LED to achieve better
discrimination between dirt and emulsion?
Similarly, do you know for a fact, that Nikon have not used red green
and blue LEDs with different spectral characteristics in their latter
scanners to more accurately reflect the eye response to the Kodachrome
dye set? For example, the highly monochrome output of a typical LED may
have a very different transmission through the dyes of differing film
with apparently the same colour, whereas a more polychromatic LED output
would produce more similar transmission.
As you say, the fundamental technology is the same, but the detailed
implementation can be significantly different. You are making a
significant assumption that your hardware is virtually the same without
any evidence whatsoever to back it up. It may well be the case that the
equivalent of the "Kodachrome setting" simply cannot be implemented on
the LS-30 because detailed hardware changes exist and no amount of
stamping your feet and bawling in the corner will change that.
Even if it was possible, why should Nikon permit you to upgrade your
scanner to the equivalent of a more advanced model? You will get a
similar response from Ford if you ask them to provide the upgraded code
for the engine management chip from an RS to fit into your L model
vehicle. At best, you might find a 3rd party performance chip supplier
who will provide their own code to improve the performance of both
models, but that is unlikely to be the same code as Ford use and will
come with its own set of limitations. Such a 3rd party exists for the
scanner world too - and he just told you, in the previous post, that
what you seek is not possible!
--
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)
Don
<mike.s...@btinternet.com> wrote:
>Hello
>
>I have some links for you.
>http://www.btinternet.com/~mike.engles/mike/nikon.jpg
>http://www.btinternet.com/~mike.engles/mike/nkn00003.set
>http://www.btinternet.com/~mike.engles/mike/nkn00004.set
>
>The first is a screen grab.
>The other two are settings using NO Kodachrome and Kodachrome, in
>Nikonscan 4. I just used the midtone sampler to set midtone curves for
>neutral grey.
>
>I am not sure how much use they will be to you.
>I think that the Kodachrome setting is a non user retrievable setting.
>It could even be built into the hardware of my LS 40.
>It is not shown by a change in the analog gain settings, or in any other
>setting that I can find.
>
>The LS 40/LS4000 are much better scanners are will be the LS50 and LS
>5000. The LS 30 is really working at it's limits. It just does not get
>on well with Kodachromes. Very few do!
>
>Good luck.
Thank you very much, Mike! I really appreciate it!
You may very well be right and Kodachrome set may not be of much use
without the rest of the program (i.e. the Kodachrome option) which is
disabled when LS-30 is connected. But I'm desperate so it's worth a
try.
Anyway, I'll have a go and see what happens. Maybe the only way to
solve this for good is to disassemble NikonScan and see what is Nikon
actually doing... ;o)
Thanks again! I really appreciate you going through the trouble of
setting this up for me. Thanks!
Don.
Don
wrote:
>"Don" <phoney...@yahoo.com> wrote:
>> Yes, my tests have also confirmed that some things are not linear. I
>> have scanned a Kodakchrome slide with LS-30 using neutral settings.
>> The result is a dark scan with a heavy blue cast. Unusable!
>
>There's something seriously wrong with the way you have NikonScan
>set up, since I've never seen this happen with NikonScan, Kodachrome
>and my LS-30.
>
>Try removing NikonScan and re-installing it.
You sound like the Nikon's so-called "support"... ;o)
And just as I wrote to them: The first thing to keep in mind is that
negatives scan perfectly! So, that eliminates pretty much all other
possible causes: hardware, software, any combination thereof, my
version of Windows, the phases of the moon... etc.
I've been at this for 9 months now and I have tried pretty much every
possible (and impossible!) permutation including multiple re-installs,
resets to factory defaults, etc, etc, etc, ad nauseaum....
LS-30 and NikonScan don't do Kodachrome out of the box!
>> I deeply resent the way Nikon goes about this, by first denying that
>> the problem exists and then not coming up with a solution (which they
>> have!) for marketing reasons.
>
>The problem doesn't exist.
I'm somewhat puzzled by this statement... Actually, I'm very puzzled!
Why do you have a separate Kodachrome setting in VueScan if there is
no problem?
Besides, if that were true, why do subsequent Nikon scanners have a
*separate* Kodachrome option in NikonScan, if there was no problem in
the first place?
Finally, a quick search on the Net will reveal plenty of others who
have the same problem with Nikon and Kodachrome. Here's one to wet
your appetite:
http://www.marginalsoftware.com/LS2000Notes/casestudyI/scanning_kodachrome_on_the_nikon_caseI.htm
Combing through *thousands* of others is left as an exercise for the
reader... ;o)
>> Like I said, I will never buy Nikon again and strongly advise everyone
>> else not to. You may be happy now, but if there's a problem, you're
>> looking at months of frustration and expense, like I have, and you'll
>> be left hanging in the end.
>
>Nikon makes excellent scanners, but like most consumer products
>companies, they can't provide the kind of detailed technical
>support that you're apparently looking for.
I'm sorry, but that just doesn't wash. It's a straw man. You're
confusing product liability regarding the most *basic* function of the
product with some exotic technical request.
The scanner does not perform the most fundamental function: Scan a
Kodachrome slide. I expect a solution from the company which produced
this hardware with such a major design flaw.
>The kind of technical detail you're looking for could barely
>be provided by the engineers that designed the scanner in the
>first place, and there are probably only one or two people at
>ASF who could provide this kind of technical support.
I'm not asking for anything special or exotic or arcane... It's the
bare essentials:
Nikon realized there was a problem with Kodachrome.
They implemented a *software* solution for subsequent scanners.
All I'm asking is a retrofit for my scanner.
Like I said, any half-decent company would provide one on their site
and apologize for the inconvenience.
After all this and the recent problem with upgrading the firmware on
my CoolPix (another excruciating example of Nikon's so-called support)
I will never purchase another Nikon product!
Don.
Ed Hamrick
> And just as I wrote to them: The first thing to keep in mind is that
> negatives scan perfectly! So, that eliminates pretty much all other
> possible causes: hardware, software, any combination thereof, my
> version of Windows, the phases of the moon... etc.
Yes, and negative scanning increases the CCD exposure time for
green and blue significantly. Slide scanning keeps the CCD exposure
time relatively constant for all three colors.
> Why do you have a separate Kodachrome setting in VueScan if there is
> no problem?
I've been meaning to remove it. It works best to set
"Input|Media type" to "Image".
> Besides, if that were true, why do subsequent Nikon scanners have a
> *separate* Kodachrome option in NikonScan, if there was no problem in
> the first place?
Because ASF improved ICE to work with Kodachrome.
> I'm sorry, but that just doesn't wash. It's a straw man. You're
> confusing product liability regarding the most *basic* function of the
> product with some exotic technical request.
Kodachrome scans ok with the LS-30, except the dark areas don't
scan well (since NikonScan only does 8-bit scans on the LS-30) and
you have to do color balancing yourself (VueScan does automatic
white balance, which is why it doesn't have the problem).
> The scanner does not perform the most fundamental function: Scan a
> Kodachrome slide. I expect a solution from the company which produced
> this hardware with such a major design flaw.
You keep repeating yourself - the LS-30 works fine with Kodachrome.
You just don't like the default color balance. Once you figure out
how to do color balancing, you'll find that the problem goes away.
Regards,
Ed Hamrick
Don
<r...@nospam.demon.co.uk> wrote:
>>The fundamental technology is the same in both so the solution is
>>clearly in software.
>
>So far, in this thread, you have not produced a single iota of evidence
>which supports that claim.
If you read the thread carefully you would've found plenty of proof.
Here's a sampling:
Is the fact that VueScan has a Kodachrome setting for use with LS-30
proof enough for you?
If not, take a garden variety LS-30, put a Kodachrome slide in. Play
with analog gain (decrease blue by about .25%, increase RGB by about
50%). Then do some heavy post processing.
Still don't believe me? OK, then try this:
http://www.marginalsoftware.com/LS2000Notes/casestudyI/scanning_kodachrome_on_the_nikon_caseI.htm
Yes, I can get reasonable results (like others who have contributed
here) but I don't want this labor-intensive hit-and-miss approach.
I want (and I'm entitled to!) *recommended* values by the company
which produced this sub-standard product! Kodachrome is not some weird
and unusual product. It's a staple in many a photographer's archive.
>The single biggest difference between how the the LS-30 and subsequent
>scanners handle Kodachrome is with ICE. What evidence do you have that
>this is entirely software? Do you know, for a fact, that Nikon have not
>used a slightly different wavelength of IR LED to achieve better
>discrimination between dirt and emulsion?
Do you know for a fact that they have?
But, this is not about ICE.
Still, if you insist, the above mentioned VueScan manages ICE with
LS-30 as well. Are you convinced now?
If not, re-read the thread carefully. As others have pointed out, only
some really dense Kodachrome slides with lots of silver left in have
problems with ICE. Many do not.
>Similarly, do you know for a fact, that Nikon have not used red green
>and blue LEDs with different spectral characteristics in their latter
>scanners to more accurately reflect the eye response to the Kodachrome
>dye set? For example, the highly monochrome output of a typical LED may
>have a very different transmission through the dyes of differing film
>with apparently the same colour, whereas a more polychromatic LED output
>would produce more similar transmission.
A whole bunch of "may have" in there...
Do you know for a fact that Nikon has used different LEDs? No? So you
are just speculating.
Normally, one does not respond to speculations, but all my evidence
shows there is no substantial difference.
>As you say, the fundamental technology is the same, but the detailed
>implementation can be significantly different. You are making a
>significant assumption that your hardware is virtually the same without
>any evidence whatsoever to back it up. It may well be the case that the
>equivalent of the "Kodachrome setting" simply cannot be implemented on
>the LS-30 because detailed hardware changes exist and no amount of
>stamping your feet and bawling in the corner will change that.
All speculations already addressed above.
>Even if it was possible, why should Nikon permit you to upgrade your
>scanner to the equivalent of a more advanced model? You will get a
>similar response from Ford if you ask them to provide the upgraded code
>for the engine management chip from an RS to fit into your L model
>vehicle.
OK, let's go automotive...
Ford tells me this car can drive everywhere. I buy it. I then find out
it doesn't turn left, only right.
According to you, my asking for a left turn is a feature of a more
advanced model!?
What you are ignoring is that scanning Kodachrome slides is the most
basic functionality of a scanner. A scanner which does not, has a
major design flaw because Kodachrome is not some rare eccentricity.
> At best, you might find a 3rd party performance chip supplier
>who will provide their own code to improve the performance of both
>models, but that is unlikely to be the same code as Ford use and will
>come with its own set of limitations. Such a 3rd party exists for the
>scanner world too - and he just told you, in the previous post, that
>what you seek is not possible!
Well, he (and you) should then check his own software which provides
just such an "impossible" functionality.
I hope I've addressed all of your speculations. I find it curious that
you are ignoring all of my (and other contributor's) evidence and yet
have not produced anything concrete yourself, basing your whole
premise on "can", "may" and the like. If you have any hard facts I
would be genuinely interested in reading them.
Don.
Don
wrote:
>"Don" <phoney...@yahoo.com> wrote:
>> And just as I wrote to them: The first thing to keep in mind is that
>> negatives scan perfectly! So, that eliminates pretty much all other
>> possible causes: hardware, software, any combination thereof, my
>> version of Windows, the phases of the moon... etc.
>
>Yes, and negative scanning increases the CCD exposure time for
>green and blue significantly. Slide scanning keeps the CCD exposure
>time relatively constant for all three colors.
Which in case of Kodachrome is exactly the problem! Just like the
negative orange cast is removed so should the Kodachrome blue cast be
removed. Failing that, the very least Nikon should do is provide this
correction as a retrofit.
Don't you find it inconsistent and illogical that Nikon automatically
removes the negative orange cast, yet (according to you) expects LS-30
users to remove the Kodachrome blue cast themselves?
>> Besides, if that were true, why do subsequent Nikon scanners have a
>> *separate* Kodachrome option in NikonScan, if there was no problem in
>> the first place?
>
>Because ASF improved ICE to work with Kodachrome.
Following that logic, NikonScan should also have B&W option removed
for LS-30 because ICE doesn't work with B&W negatives either?
The truth is, ICE has nothing to do with the Kodachrome blue cast or
dark scans which are both due to different nature of Kodachrome film.
That's why NikonScan needs a special Kodachrome setting for LS-30.
>> I'm sorry, but that just doesn't wash. It's a straw man. You're
>> confusing product liability regarding the most *basic* function of the
>> product with some exotic technical request.
>
>Kodachrome scans ok with the LS-30, except the dark areas don't
>scan well (since NikonScan only does 8-bit scans on the LS-30) and
>you have to do color balancing yourself (VueScan does automatic
>white balance, which is why it doesn't have the problem).
In other words: NikonScan produces a blue cast at factory defaults,
VueScan doesn't because you correct it.
That's what I've been saying all along! NikonScan produces dark scans
and a heavy blue cast on an LS-30 resulting in unusable Kodachrome
scans. And Nikon should correct this just like you do, and just like
they correct the negative orange cast.
>> The scanner does not perform the most fundamental function: Scan a
>> Kodachrome slide. I expect a solution from the company which produced
>> this hardware with such a major design flaw.
>
>You keep repeating yourself - the LS-30 works fine with Kodachrome.
Are you saying that with settings at factory defaults your LS-30
NikonScan Kodachrome scans don't have a blue cast and are not dark?
>You just don't like the default color balance.
Using the same logic, would you call the heavy orange cast on an
uncorrected negative "the default color balance"?
There is *no* default color balance with Kodachrome on LS-30 under
NikonScan! That's exactly the problem!
>Once you figure out
>how to do color balancing, you'll find that the problem goes away.
That's not only condescending but it misses the point. You have also
conveniently ignored all of my references.
If I made a mistake shooting my slides then, indeed, it would be up to
me to fix it in post processing.
However, if there's an intrinsic and inherent aspect of the media
(like the negative orange mask or the Kodachrome blue mask) that's the
job of the device manufacturer. And, indeed, Nikon does correct the
negatives, but fails miserably with Kodachrome on LS-30.
So, the facts remain: NikonScan can not do Kodachrome out of the box
on an LS-30 and Nikon refuses to fix it which is quite despicable.
Don.
Ed Hamrick
> In other words: NikonScan produces a blue cast at factory defaults,
Exactly. You need to adjust the color balance manually.
> So, the facts remain: NikonScan can not do Kodachrome out of the box
> on an LS-30 and Nikon refuses to fix it which is quite despicable.
Despicable is a bit strong. NikonScan just doesn't include
automatic color balancing. You have to color balance these scans
yourself. If you don't want to do manual color balancing, use
VueScan.
Regards,
Ed Hamrick
Don
wrote:
>"Don" <phoney...@yahoo.com> wrote:
>> In other words: NikonScan produces a blue cast at factory defaults,
>
>Exactly. You need to adjust the color balance manually.
And I was prepared to! Only if Nikon would provide me with the
*recommended* values they use when the Kodachrome setting is selected.
They have my money and it's the least they can do!
Considering that it's their failure to omit/retrofit Kodachrome in the
first place, I think I was being very reasonable.
Instead, the incompetent Nikon's so-called "support" blames my version
of Windows...
>> So, the facts remain: NikonScan can not do Kodachrome out of the box
>> on an LS-30 and Nikon refuses to fix it which is quite despicable.
>
>Despicable is a bit strong. NikonScan just doesn't include
>automatic color balancing. You have to color balance these scans
>yourself. If you don't want to do manual color balancing, use
>VueScan.
If you have been given the run-around and the nonsense from Nikon as I
have, you'd consider "despicable" quite mild.
Nikon is clearly at fault here. First, by failing to provide this
essential feature and then for totally incompetent "support" even
though I was being very patient and willing to do as much as
necessary. For three months!!!
The bottom line is I will never purchase another Nikon product and I
will certainly advise anyone else not to either. Like as I wrote
already, you may be happy now, but if there's a problem, you're
looking at months of frustration and expense, like I have, and you'll
be left hanging in the end.
Don.
Kennedy McEwen
<phoney...@yahoo.com> writes
>
>LS-30 and NikonScan don't do Kodachrome out of the box!
>
"Yes, I can get reasonable results (like others who have contributed
The LS-30 will scan KC out of the box - you have already established
that. Like all emulsions, optimum performance requires additional post
scan adjustment.
That subsequent scanners provide a capability to reduce this post scan
work is of no relevance to this discussion other than your individual
envy of them.
Kennedy McEwen
<phoney...@yahoo.com> writes
>On Sat, 13 Dec 2003 16:46:10 +0000, Kennedy McEwen
><r...@nospam.demon.co.uk> wrote:
>
>>>The fundamental technology is the same in both so the solution is
>>>clearly in software.
>>
>>So far, in this thread, you have not produced a single iota of evidence
>>which supports that claim.
>
>If you read the thread carefully you would've found plenty of proof.
>Here's a sampling:
>
>Is the fact that VueScan has a Kodachrome setting for use with LS-30
>proof enough for you?
>
Ford proves that the engine hardware for a Cosworth is the same as that
of an Escort L. Neither is the case, should you still be confused.
>If not, take a garden variety LS-30, put a Kodachrome slide in. Play
>with analog gain (decrease blue by about .25%, increase RGB by about
>50%). Then do some heavy post processing.
>
Again, that doesn't prove squat about the hardware of an LS-30 or any
comparison of it to subsequent scanners.
>Still don't believe me? OK, then try this:
>
>http://www.marginalsoftware.com/LS2000Notes/casestudyI/scanning_kodachro
>me_on_the_nikon_caseI.htm
Again, read my comments above - Nikon scanners have been able to get
acceptable results from Kodachrome for several generations, with some
level of post scan work as in the LS-2000 example you reference. Just
because a specific Kodachrome option exists in more recent scanners does
not prove that this is a "software only" modification, as per your
allegation and continual foot stamping in this thread.
>
>Yes, I can get reasonable results (like others who have contributed
>here) but I don't want this labor-intensive hit-and-miss approach.
>
So buy a scanner which has the modifications, whether hardware or
software or both, built into it!
>I want (and I'm entitled to!) *recommended* values by the company
>which produced this sub-standard product!
No, you are NOT entitled to it - no more than you are entitled to a
grain dissolver option such as that provided by the latest Minolta
scanner, or the Dynamic Exposure Extender provided in the latest issue
of Nikon scanners. You are only entitled to the features, hardware AND
software, which were provided with the scanner you paid for!
> Kodachrome is not some weird
>and unusual product. It's a staple in many a photographer's archive.
>
And the LS-30 scanner you have produces perfectly adequate results from
the Kodachrome emulsions - by your own admission - you just have
additional work to do! You have no right to demand that Nikon provide
you with labour saving features just because they have provided these in
subsequent scanners - even if they are only software features - no more
than you have to GEM and ROC features, for example, which specifically
are software only!
>>The single biggest difference between how the the LS-30 and subsequent
>>scanners handle Kodachrome is with ICE. What evidence do you have that
>>this is entirely software? Do you know, for a fact, that Nikon have not
>>used a slightly different wavelength of IR LED to achieve better
>>discrimination between dirt and emulsion?
>
>Do you know for a fact that they have?
>
No, but I am not ASSUMING that they haven't, nor doing so without
evidence!
>But, this is not about ICE.
>
>Still, if you insist, the above mentioned VueScan manages ICE with
>LS-30 as well. Are you convinced now?
No - I refer you to my previous comparison with 3rd party aftermarket
suppliers.
>
>If not, re-read the thread carefully. As others have pointed out, only
>some really dense Kodachrome slides with lots of silver left in have
>problems with ICE. Many do not.
>
And, with later scanners the number of Kodachrome slides which ICE has
problems with is even less - indeed I have not found a single example
yet on my LS-4000. That merely indicates a step change in handling that
emulsion, it does NOT indicate that you should, or even can, achieve the
same results with earlier hardware or that you are entitled to expect
it.
>>Similarly, do you know for a fact, that Nikon have not used red green
>>and blue LEDs with different spectral characteristics in their latter
>>scanners to more accurately reflect the eye response to the Kodachrome
>>dye set? For example, the highly monochrome output of a typical LED may
>>have a very different transmission through the dyes of differing film
>>with apparently the same colour, whereas a more polychromatic LED output
>>would produce more similar transmission.
>
>Do you know for a fact that Nikon has used different LEDs? No? So you
>are just speculating.
>
No, I am recognising that hardware changes are not only likely between
models but could have significant effects on the performance in the area
of concern to you AND explaining to you WHY that would be so.
>Normally, one does not respond to speculations, but all my evidence
>shows there is no substantial difference.
None of the evidence you have so far provided suggests anything of the
sort. In fact, the evidence provided indicates that such changes are
actually quite likely!
>
>>As you say, the fundamental technology is the same, but the detailed
>>implementation can be significantly different. You are making a
>>significant assumption that your hardware is virtually the same without
>>any evidence whatsoever to back it up. It may well be the case that the
>>equivalent of the "Kodachrome setting" simply cannot be implemented on
>>the LS-30 because detailed hardware changes exist and no amount of
>>stamping your feet and bawling in the corner will change that.
>
>All speculations already addressed above.
On the contrary, you are clearly bawling and stamping your feet
demanding something that you are not entitled to - there is no
speculation in that whatsoever, it is a statement which is wholly
supported by the evidence in the thread.
>
>>Even if it was possible, why should Nikon permit you to upgrade your
>>scanner to the equivalent of a more advanced model? You will get a
>>similar response from Ford if you ask them to provide the upgraded code
>>for the engine management chip from an RS to fit into your L model
>>vehicle.
>
>OK, let's go automotive...
>
>Ford tells me this car can drive everywhere. I buy it. I then find out
>it doesn't turn left, only right.
>
But you have already stated that you can get good results from your
LS-30 with KC film - you just need to make the effort. Thus, the actual
analogy is that your car steers to the left perfectly well if you are
prepared to make the effort to turn the steering wheel in that
direction.
>According to you, my asking for a left turn is a feature of a more
>advanced model!?
>
Cars which steer in either direction without any effort expended by the
"driver" are from a different generation, yet to be released to the
general public, and you have no entitlement to expect autosteering to be
retro-fitted into your older model when it does become available, even
if it becomes a standard feature of subsequent models.
>What you are ignoring is that scanning Kodachrome slides is the most
>basic functionality of a scanner.
No, I am not ignoring it - neither am I ASSUMING it, because
demonstrably that is NOT the case. On the contrary, it is well known
that KC is one of, if not absolutely, the most difficult emulsion for a
scanner to cope with - certainly not the most basic one.
>A scanner which does not, has a
>major design flaw because Kodachrome is not some rare eccentricity.
>
Neither rare nor eccentric, but certainly difficult - and as such it is
not unexpected that you should need to apply additional effort to
achieve optimal results, especially with older, less developed
equipment. Even the output from the latest LS-5000 will require post
scan processing to achieve the optimum results in some cases, whether
with KC or any other emulsion.
>
>
>I hope I've addressed all of your speculations.
No, all you have done is emphasised that you are making a significant
assumption which none of the evidence you provide even begins to address
but which is critical to the question of whether your desire CAN be
implemented. In addition, you are making a further significant
assumption, again without evidence, that you are entitled to features
made available in more advanced models - even if it were technically
possible to do that.
Don
<r...@nospam.demon.co.uk> wrote:
>In article <3fdb5126...@news.individual.net>, Don
><phoney...@yahoo.com> writes
>>
>>LS-30 and NikonScan don't do Kodachrome out of the box!
>>
>Contradicted by your earlier statement:
>"Yes, I can get reasonable results (like others who have contributed
>here) but I don't want this labour-intensive hit-and-miss approach."
We must not be using the same alphabet...
Please re-read carefully what I wrote. Two key phrases to pay
particular attention to:
labor-intensive
hit-and-miss
Ergo: LS-30 and NikonScan don't do Kodachrome out of the box!
Key phrase:
out of the box
Don
Don
<r...@nospam.demon.co.uk> wrote:
>In article <3fdb645a...@news.individual.net>, Don
><phoney...@yahoo.com> writes
>>On Sat, 13 Dec 2003 16:46:10 +0000, Kennedy McEwen
>><r...@nospam.demon.co.uk> wrote:
>>
>>>>The fundamental technology is the same in both so the solution is
>>>>clearly in software.
>>>
>>>So far, in this thread, you have not produced a single iota of evidence
>>>which supports that claim.
>>
>>If you read the thread carefully you would've found plenty of proof.
>>Here's a sampling:
>>
>>Is the fact that VueScan has a Kodachrome setting for use with LS-30
>>proof enough for you?
>>
>No - no more than the supply of 3rd party engine management chips for
>Ford proves that the engine hardware for a Cosworth is the same as that
>of an Escort L. Neither is the case, should you still be confused.
Me: ...the solution is clearly in software.
You: ...you have not produced a single iota of evidence...
Me: ...VueScan has a Kodachrome setting for use with LS-30... etc
Ipso facto, the software does it, even though it may be indirect - as
in modifying analog gain.
... repetition skipped...
>>Yes, I can get reasonable results (like others who have contributed
>>here) but I don't want this labor-intensive hit-and-miss approach.
>>
>So buy a scanner which has the modifications, whether hardware or
>software or both, built into it!
I have! It's called LS-30. However, what I did not count on is:
1. False advertising by Nikon.
2. Nikon's inane so-called "support".
>> Kodachrome is not some weird
>>and unusual product. It's a staple in many a photographer's archive.
>>
>And the LS-30 scanner you have produces perfectly adequate results from
>the Kodachrome emulsions - by your own admission -
You've conveniently ignored those two pesky key phrases:
labor-intensive
hit-and-miss
>>>The single biggest difference between how the the LS-30 and subsequent
>>>scanners handle Kodachrome is with ICE. What evidence do you have that
>>>this is entirely software? Do you know, for a fact, that Nikon have not
>>>used a slightly different wavelength of IR LED to achieve better
>>>discrimination between dirt and emulsion?
>>
>>Do you know for a fact that they have?
>>
>No, but I am not ASSUMING that they haven't, nor doing so without
>evidence!
Sophistry. There's ample evidence, as I've illustrated.
...lots of repetition skipped...
>>A scanner which does not, has a
>>major design flaw because Kodachrome is not some rare eccentricity.
>>
>Neither rare nor eccentric, but certainly difficult
So why didn't Nikon state this clearly and unambiguously? After I
encountered this design flaw I scanned (pun intended) all the LS-30
documentation carefully and there is absolutely no mention anywhere of
Kodachrome being "difficult" or a problem.
There was one vague phrase (probably concocted by lawyers) that "some
film may not scan properly" or words to that effect. However,
Kodachrome is not "some film". It's a whole category. And a major one
at that!
Anyway, I don't mean any disrespect, but I really have no time for
this. I now have to disassemble NikonScan and that will keep me busy
for a while.
If what I wrote really bothers you that much, please feel free to put
me in your kill file.
Just like I have put Nikon in my kill file...
>Socrates himself is particularly missed
Indeed!
Don.
Rafe B.
>Hi,
>
>I need a couple of sets from a Nikon film scanner claiming to do
>Kodachrome slides (I believe 8000 is one of them).
>
>Having initialized all Nikon Scan settings to neutral (no analog gain,
>flat curves, etc) please save one of each:
>1. a Positive set
>2. a Positive - Kodachrome set
>
>Here's a temporary email address to send it to:
>
> tempd...@yahoo.com
>
>It's a lot to ask, I know, but Nikon is no help and I will never buy a
>Nikon product again...
>
>Thanks very much in advance.
>
>Don.
I've posted three Kodachrome scans straight out
of the LS-8000, downsized to 1000 pixels wide, at
the URLs below
http://www.terrapinphoto.com/kchrome1.jpg
http://www.terrapinphoto.com/kchrome2.jpg
http://www.terrapinphoto.com/kchrome3.jpg
Totally neutral settings in NikonScan. ICE off.
Film dates;
kchrome1: October 1968
kchrome2: November 1979
kchrome3: November 1980
rafe b.
http://www.terrapinphoto.com
Rafe B.
Erik Krause
you wrote...
> > Why do you have a separate Kodachrome setting in VueScan if there is
> > no problem?
>
> I've been meaning to remove it.
Please don't! You can remove the slide film types since thea are of no
worth (at least with my LS 40). But I urge you not to remove Media Type
Slide and Generic Slide!
> It works best to set
> "Input|Media type" to "Image".
It works best with advanced workflow!
--
Erik Krause
Digital contrast problems: http://www.erik-krause.de/contrast
Don
wrote:
>On Mon, 08 Dec 2003 04:52:18 GMT, phoney...@yahoo.com (Don) wrote:
>
>>Hi,
>>
>>I need a couple of sets from a Nikon film scanner claiming to do
>>Kodachrome slides (I believe 8000 is one of them).
...
>I've posted three Kodachrome scans straight out
>of the LS-8000, downsized to 1000 pixels wide, at
>the URLs below
>
>http://www.terrapinphoto.com/kchrome1.jpg
>http://www.terrapinphoto.com/kchrome2.jpg
>http://www.terrapinphoto.com/kchrome3.jpg
>
>
>Totally neutral settings in NikonScan. ICE off.
>
>Film dates;
>
>kchrome1: October 1968
>kchrome2: November 1979
>kchrome3: November 1980
Thanks very much, Rafe!
So far no luck with a couple of sets I received (thanks again Erik and
Mike!) and I'm starting to suspect the correction is actually done
internally within NikonScan.
Don.
Kennedy McEwen
<phoney...@yahoo.com> writes
>
>Me: ...the solution is clearly in software.
>You: ...you have not produced a single iota of evidence...
>Me: ...VueScan has a Kodachrome setting for use with LS-30... etc
>
>Ipso facto, the software does it, even though it may be indirect - as
>in modifying analog gain.
>
nothing in this context - AS THE VUESCAN AUTHOR HAS ALREADY TOLD YOU!
>It's called LS-30. However, what I did not count on is:
>1. False advertising by Nikon.
>2. Nikon's inane so-called "support".
3. Your incompetence.
4. Your ignorance of the amount of work most people put into getting the
scanned images they want.
>
>You've conveniently ignored those two pesky key phrases:
> labor-intensive
> hit-and-miss
>
No, not ignored - simply put into the context of the topic. As many
have mentioned the scanner works perfectly well out of the box and even
produces acceptable Kodachrome results.
>
>Sophistry. There's ample evidence, as I've illustrated.
>
As I have been at pains to point out - not a single iota of the evidence
that you have presented supports the claim that the same KC capability
of more recent scanners can be implemented on the LS-30 with a software
only modification. It may be the case that it is, but that is
coincidental - NOTHING you have presented proves, or even indicates,
that it is! (Vuescan is a completely different application which
specifically includes modes, such as automatic white balance, that
NikonScan does not, so any comparison between VS with the LS-30 and
options provided in NS are completely irrelevant!)
>>Neither rare nor eccentric, but certainly difficult
>
>So why didn't Nikon state this clearly and unambiguously? After I
>encountered this design flaw I scanned (pun intended) all the LS-30
>documentation carefully and there is absolutely no mention anywhere of
>Kodachrome being "difficult" or a problem.
>
>There was one vague phrase (probably concocted by lawyers) that "some
>film may not scan properly" or words to that effect. However,
>Kodachrome is not "some film". It's a whole category. And a major one
>at that!
>
But it is still "some film" and, amongst the highest density films
available, together with a completely unique dye structure which is
significantly different from every other colour slide film that has ever
reached the market. Neither rare nor eccentric, but certainly different
from typical film and, as a consequence of these differences, likely to
require additional effort to achieve ideal results - however, as others
have demonstrated, acceptable results are regularly achieved with KC
straight out of the box.
>
>Anyway, I don't mean any disrespect, but I really have no time for
>this. I now have to disassemble NikonScan and that will keep me busy
>for a while.
>
I think you will find that such action, and the use of any such
information you may find in there, constitutes a serious breach of the
NikonScan User Licence. Should Nikon choose to sue you as a
consequence, it has not been a wise move on your part to openly state
your intention to conduct this illegal act on a public forum.
>If what I wrote really bothers you that much, please feel free to put
>me in your kill file.
>
Doesn't bother me at all - you have yet to demonstrate any logic in your
argument, but have demonstrated a strong desire to get more than you
paid for, even to the point of undertaking illegal activities.
Kennedy McEwen
<phoney...@yahoo.com> writes
>
>Please re-read carefully what I wrote. Two key phrases to pay
>particular attention to:
> labor-intensive
> hit-and-miss
>
>Ergo: LS-30 and NikonScan don't do Kodachrome out of the box!
>
>Key phrase:
> out of the box
>
optimal results. You do need to press the "Scan" button, you know!
Mike Engles
Hello
I think that you are correct.
Mike Engles
Mike Engles
Hello
I realise that you are really fed up. Unfortunately the LS 30 is not
really up to scanning Kodachromes well. I had one. I bought a LS 40. It
is much better, but still has trouble with dense images. I am hoping
that the LS 50 or LS 5000 will do Kodachromes,out of the box, as you
say. I think Nikon scanners are the best. I have tried them all.
I also only use Nikon Scan.
Here are a couple of links to Kodachrome images on the LS 40.
I stll had to do quite a lot of work on them.
Mike Engles
http://www.btinternet.com/~mike.engles/mike/Bumblebee.jpg
http://www.btinternet.com/~mike.engles/mike/AntsAphids.jpg
Mike Engles
Hello
With my Ls 40 and Nikonscan 4 set to 'positive' I get a neutral well
exposed slide with these settings.
Analog gain
Master +.66
Red +.9
Green 0
Blue-.20
I have to assume that your monitor is calibrated.
Choose a slide that know has areas of grey tone.
Do a preview
Next use the midtone dropper and click on any area that should be
neutral grey. This should fine tune the tone for that area and with any
luck give you a image without the cyan/blue cast.
I use the Apple compensated colour space and have my monitor at inverse
gamma 1.8. I use Adobe gamma.
I hope this is some help.
Mike Engles
Hils
and miss" sums it up perfectly. KC scans take a lot more work than any
other film I've used, and some I can do nothing with: the most
consistent element I've found is a lot of red noise in shadows.
Thankfully it's becoming moot, I no longer use Kodachrome and I'm
planning a move to all-digital working.
>Instead, the incompetent Nikon's so-called "support" blames my version
>of Windows...
That would probably drive me crazy too. (As did Nikon taking 9 weeks and
then 6 weeks to repair the scanner.)
--
Hil
Don
<mike.s...@btinternet.com> wrote:
>> So far no luck with a couple of sets I received (thanks again Erik and
>> Mike!) and I'm starting to suspect the correction is actually done
>> internally within NikonScan.
>>
>> Don.
>
>
>Hello
>
>I think that you are correct.
>
>Mike Engles
Hi Mike!
I'm now busy disassembling NikonScan 3.1 (version 4 doesn't work with
LS-30 if Nikon's site is to be believed and I'm just too drained to
try yet another futile install.).
Disassembly is very time consuming, of course, and so far I only
located where LS2K, 4K and 8K drivers are loaded, as well as TWAIN
loading code where Nikon's own "Nikon Scan.ds" gets sucked in. SCSI
code (which is how LS-30 connects) is still hiding.
Of course, all this may be a waste of time, if the Kodachrome
compensation is done in firmware (I seriously doubt it, though). But,
I may have to continue wasting more time by disassembling the
firmware... all other things being equal.
At least I'm doing something constructive and learning about the inner
workings of NikonScan, rather than the total waste of time while being
led around by the nose by Nikon's so-called "support".
Don.
Don
<mike.s...@btinternet.com> wrote:
>Hello
>
>With my Ls 40 and Nikonscan 4 set to 'positive' I get a neutral well
>exposed slide with these settings.
>
>Analog gain
>
>Master +.66
>Red +.9
>Green 0
>Blue-.20
>
>I have to assume that your monitor is calibrated.
>Choose a slide that know has areas of grey tone.
>Do a preview
>Next use the midtone dropper and click on any area that should be
>neutral grey. This should fine tune the tone for that area and with any
>luck give you a image without the cyan/blue cast.
>
>I use the Apple compensated colour space and have my monitor at inverse
>gamma 1.8. I use Adobe gamma.
>
>I hope this is some help.
>
>Mike Engles
Hi Mike!
Thanks again for your constructive messages. I really appreciate you
taking the time and the trouble to help me out! I only wish there's
something I can do in return. I'll keep you posted on anything I
"discover" if ever get anywhere with my snooping around NikonScan.
Your settings are very close to what I arrived at. In particular the
slight boost of the red! Most people miss that one and only attenuate
the blue channel (the usual advice on the Net)! Well spotted!
Isn't it unbelievable that Nikon could settle all this in a nanosecond
by giving us the recommended values, instead of our guesswork?
Anyway, as I think I mentioned elsewhere, one of the tests I did was
to take a picture of a slide with my CoolPix camera against the light
source of my flatbed scanner. This produced excellent results
(regarding the color balance) and the pic looks identical to the slide
itself. Of course, the resolution is another story...
Now, scanning the same slide with the scanner produced areas with
absolute absence of red, even though red is quite prominent in the
same areas of the pic taken with the camera!
I also tried to extend the range by making two scans at between .5 and
1 difference in master analog gain setting. The idea was to combine
them in Photoshop with a large feathering value. That brightens up the
dark areas without burning out the highlights - but it's very time
consuming. Best results were achieved by making the scans in
succession without removing the slide. Otherwise, the two images are
not aligned as accurately.
Don.
Don
wrote:
Thanks very much, Hil!
With Ed and Kennedy going at me like there's no tomorrow, it's good to
know there are others who have had problems with both Kodachrome as
well as trouble with Nikon.
So, it's not all in my head... ;o)
Don.
Don
restate the 'baseline", which has gotten lost, and leave it at that:
* All I wanted from Nikon was to produce a reasonable facsimile of the
scanned item. At least not make it substantially worse.
* I don't expect (and, indeed, don't want!) Nikon to do anything more.
* However, I'm not going to settle for anything less.
In particular:
1. A Kodachrome slide does *not* have a visible blue cast. A neutral
Nikon LS-30 scan does. This cast is therefore a known, fixed and
predictable entity. It is also a clear case of image degradation
caused by hardware. Nikon *must* compensate for this.
2. A negative film *has* a visible orange cast. This is also a known,
fixed and predictable entity. Logically, Nikon is not required to
compensate for this because the orange cast is visible in the film
itself. For commercial reasons, however, they do and that's nice.
3. Both of the above assume a reasonable source image. However, if the
image is under or overexposed, or has some other flaw, I don't expect
Nikon to do anything. If they do, it's a "cherry on top" which is much
appreciated, but I consider that "color correction" and that's my job.
For some reason (I can't fathom) these facts seems to irritate you.
I'm sorry, and like I suggested last time, maybe you should just skip
my messages.
However, if you fail to acknowledge these objective facts then there's
nothing else I can write and we'll just be going around in circles.
Finally, maybe some pause for thought should be taken from Mike's,
Erik's and Hil's reactions, who did not find my statements excessive
and have, indeed, been very helpful and supportive, just like you seem
to be to others which is why I find your responses even more puzzling.
With that, I consider the matter closed. Live long and prosper!
Don.
Hils
>LS-30 if Nikon's site is to be believed and I'm just too drained to
>try yet another futile install.).
Nikon's European website says that NikonScan 4 "does not support SCSI
computer connections". Surely it would have been less trouble to have
left the SCSI in, but then, this is a company which thinks that buying a
new operating system will improve scan quality, and that no-one wants
(or they can't build) full-frame DSLRs. >:-|
--
Hil
Kennedy McEwen
<phoney...@yahoo.com> writes
>
>1. A Kodachrome slide does *not* have a visible blue cast. A neutral
>Nikon LS-30 scan does. This cast is therefore a known, fixed and
>predictable entity. It is also a clear case of image degradation
>caused by hardware. Nikon *must* compensate for this.
>
why there may not even be a possible generic solution to what you
request.
The difference that you are experiencing between KC and other emulsions,
in terms of the colour your scanner produces and the colour your eyes
see is simply due to the different spectral characteristics of the dyes
used in the emulsion and how they map onto the spectral response of your
eye and the scanner. The very fact that you can take an E-6 slide and a
KC slide with very similar colours to the eye and produce a
significantly different colour balance scan between the two proves that
the dyes used in the two film emulsions have very different detailed
spectral characteristics. The difference is a form of metamerism.
Nikon use LEDs as their illumination source, which has the advantage
that they have a very controlled spectral characteristic, but this is
also much narrower than the response of, say, the gelatine filters on a
tri-linear CCD. In either case, the scanner sees a much narrower band
of the dye spectrum than your eye does. However, if the emulsion
spectrum is reasonably constant across a range of film types then the
response within the narrow band will be a fixed proportion of the
response produced in the cones of your eye, and can thus be balanced by
selection of the exposure or the brightness of the LEDs.
Nikon have chosen to balance the response of each channel to produce a
reasonable colour for the majority of colour slide film in use today,
which are chromogenic emulsions. Kodachrome, whilst popular, is
significantly different. Furthermore, Kodachrome has modified the
spectral characteristics of the dyes depending on when and where it was
processed - the dyes are in the development process, not embedded in the
emulsion like other colour film types. Check the files that Rafe posted
of differing KC variants for evidence of this - all on the same scanner,
all at the same setting, only the version of KC changed, but a different
colour cast on each - and that from a much more capable scanner than
yours WITH Kodachrome capability.
Nikon have only provided a separate default balance for KC emulsions on
their later scanners - and you have assumed that this is purely a
function of the software without any evidence to support the assumption.
Because of the variability of KC spectral characteristics, it is quite
likely that a single unique and acceptable default balance for KC has
required selection of a particular spectral LED characteristic, in at
least one colour and possibly all four (including IR). This in itself
would prevent the generic profile to be produced for an older scanner
type such as you request, even if Nikon wanted to continue support for a
long obsolete scanner.
>2. A negative film *has* a visible orange cast. This is also a known,
>fixed and predictable entity. Logically, Nikon is not required to
>compensate for this because the orange cast is visible in the film
>itself. For commercial reasons, however, they do and that's nice.
>
Not the same issue at all. The orange cast in the negative determines
the black point of the final image. The spectral mapping of the
emulsion and the sensor/LED determines the white point in a slide image.
I hope you understand the difference here, because it specifically
identifies that some of your alleged evidence in support of your
assumption is wrong. In addition, the orange mask REQUIRES different
exposures of all three channels when using negatives, specifically to
provide a usable dynamic range at the ADC input - the orange mask swamps
the image contrast recorded by the emulsion. That is NOT necessary with
slides - or even desirable with a limited dynamic range ADC, due to the
high contrast present in the original material.
>3. Both of the above assume a reasonable source image. However, if the
>image is under or overexposed, or has some other flaw, I don't expect
>Nikon to do anything. If they do, it's a "cherry on top" which is much
>appreciated, but I consider that "color correction" and that's my job.
>
Indeed it is, but you seem to consider colour correction caused by
Kodachrome's unique dye characteristics to not be your job, simply
because Nikon have addressed that issue with later scanner models.
>For some reason (I can't fathom) these facts seems to irritate you.
The facts don't irritate me at all, those that actually are facts.
However I am trying to explain to you that a) you have made a
fundamental assumption without any evidence and b) even if that
assumption proved to be correct, Nikon are under no obligation to
provide it, no matter how loudly you stamp your feet. I have no
objection to you attempting to achieve your desire, and wish you well in
your quest, however it is, as others have pointed out, likely to be
futile with the equipment you have. Neither is Nikon's failure to
provide the information you require valid reason for criticism of them,
whether it is possible or not. I would, however, accept that Nikon
should have informed you that this option was not supported for your
scanner early in your quest.
I assume that you have downloaded and installed the firmware upgrade for
the LS-30 that Nikon HAVE provided.
Don
Don
<r...@nospam.demon.co.uk> wrote:
>>1. A Kodachrome slide does *not* have a visible blue cast. A neutral
>>Nikon LS-30 scan does. This cast is therefore a known, fixed and
>>predictable entity. It is also a clear case of image degradation
>>caused by hardware. Nikon *must* compensate for this.
>>
>There is no "must" about it.
I would dispute that because Kodachrome is a staple film stock and
Nikon can't just dismiss it out of hand. But then we'll get back into
a circular argument, so we better leave it...
>In fact, I have tried to explain to you
>why there may not even be a possible generic solution to what you
>request.
Every generic setting is a compromise by definition.
I'm sure you'll agree that some sort of optimum setting for Kodachrome
on LS-30 is possible because people have been doing it themselves and
there is loose consensus on a rule-of-thumb setting. (Roughly, about
25-30% attenuation of the blue with a slight 5-10% boost to the red,
plus .5 to 1 EV boost to RGB analog gain.) My own settings do produce
consistent results although some minor tweaking may still be needed
but it's nothing to do with color balance.
So, everything seems to point to the fact that such a generic setting
is possible. It will at least give a solid base. Sure, "your mileage
may vary" but it would be better than starting from scratch every
time. I mean there are certain things which are common to all
Kodachrome slides: the darkness, the blue cast. Even if they vary
among individual slides there is at least some commonality. And isn't
it better to have Nikon give recommended settings instead of us
"civilians" stabbing in the dark?
Anyway, in the absence of Nikon's cooperation, the only other way to
settle this is if I ever manage to turn on Kodachrome option for LS-30
and get the definitive answer.
>Nikon use LEDs as their illumination source, which has the advantage
>that they have a very controlled spectral characteristic, but this is
...
>Nikon have chosen to balance the response of each channel to produce a
>reasonable colour for the majority of colour slide film in use today,
...
>Nikon have only provided a separate default balance for KC emulsions on
>their later scanners - and you have assumed that this is purely a
...
Now, this is the type of constructive response I expected to get!
Thanks!
>>3. Both of the above assume a reasonable source image. However, if the
>>image is under or overexposed, or has some other flaw, I don't expect
>>Nikon to do anything. If they do, it's a "cherry on top" which is much
>>appreciated, but I consider that "color correction" and that's my job.
>>
>Indeed it is, but you seem to consider colour correction caused by
>Kodachrome's unique dye characteristics to not be your job, simply
>because Nikon have addressed that issue with later scanner models.
The other is product liability. In spite of it being difficult
Kodachrome is still a major brand, especially when it comes to
archives. Like they say: "If I knew then what I know..." But I'm stuck
with my Kodachrome archive.
> I would, however, accept that Nikon
>should have informed you that this option was not supported for your
>scanner early in your quest.
Indeed! Instead, 1st-level support seems totally unaware of the
Kodachrome problem trying to change the subject in all sorts of
irrelevant directions. 2nd-level support then continued to give me a
run-around denying the problem. Finally, they had me send in a slide
(which itself is ridiculous because they must have Kodachrome stock)
only to produce a mediocre "solution" worse than what I myself have
arrived at! Their solution just brought the levels a tad up and didn't
address the blue cast (the major problem) at all!
Throughout all this I was very patient and jumped through all the
hoops (no matter how ridiculous) hoping for a decent result in the
end. Now that I didn't get it after everything they've put me through
I have earned the right to stomp my feet!
I mean how would you have reacted when you tell them up front:
1. negatives scan fine
2. a blank slide frame comes up white, as it should
3. Kodachrome is dark and blue
and their response is : It's your version of Windows! No, it's a
burned out LED! No, you don't know how to scan! Etc.
>I assume that you have downloaded and installed the firmware upgrade for
>the LS-30 that Nikon HAVE provided.
That's the first thing I did! After I bought the scanner 4 years ago I
ended up working in a different country and the scanner remained
unopened. So when I finally got back for good the first thing I did is
upgrade the firmware and NikonScan.
After that, it all went downhill.
Should have left it unopened... ;o)
Don.
Kennedy McEwen
<phoney...@yahoo.com> writes
>
>I'm sure you'll agree that some sort of optimum setting for Kodachrome
>on LS-30 is possible because people have been doing it themselves and
>there is loose consensus on a rule-of-thumb setting. (Roughly, about
>25-30% attenuation of the blue with a slight 5-10% boost to the red,
>plus .5 to 1 EV boost to RGB analog gain.) My own settings do produce
>consistent results although some minor tweaking may still be needed
>but it's nothing to do with color balance.
>
results from some Kodachrome stock you may have - but not all, as the
results that Rafe posted from his LS-8000 demonstrate. And there is no
evidence that this generic setting that you have derived for your own
satisfaction is at all similar to the settings that Nikon use in their
later scanners due to changes in those scanner's design.
Don
<r...@nospam.demon.co.uk> wrote:
>In article <3fe09251...@news.individual.net>, Don
><phoney...@yahoo.com> writes
>>
>>I'm sure you'll agree that some sort of optimum setting for Kodachrome
>>on LS-30 is possible because people have been doing it themselves and
>>there is loose consensus on a rule-of-thumb setting. (Roughly, about
>>25-30% attenuation of the blue with a slight 5-10% boost to the red,
>>plus .5 to 1 EV boost to RGB analog gain.) My own settings do produce
>>consistent results although some minor tweaking may still be needed
>>but it's nothing to do with color balance.
>>
>Yes, a generic setting is possible which gives reasonably acceptable
>results from some Kodachrome stock you may have - but not all, as the
>results that Rafe posted from his LS-8000 demonstrate. And there is no
>evidence that this generic setting that you have derived for your own
>satisfaction is at all similar to the settings that Nikon use in their
>later scanners due to changes in those scanner's design.
Considering what I get now, *all* of Rafe's scans look *fantastic* by
comparison!
First a CoolPix 950 picture of the slide, indistinguishable from the
way the slide actually looks i.e. this is the baseline:
http://members.aol.com/tempdon100164833/nikon/CoolPix.jpg
And then the LS30's despicable mutilation, all settings at neutral -
i.e. "factory defaults":
http://members.aol.com/tempdon100164833/nikon/LS30.jpg
According to Ed this is the "default color balance".
Yes, right... Well, I don't think so...
So, it's clearly up to Nikon's so-called "support" to provide
recommended Kodachrome settings. Since it's their hardware, they
should know and it would certainly be better than any of my (or
anybody else's) *guesses*.
Don.
P.S. For any parachutists with an urge to comment: ***Please read the
whole thread first!***
Kennedy McEwen
<phoney...@yahoo.com> writes
>
>First a CoolPix 950 picture of the slide, indistinguishable from the
>way the slide actually looks i.e. this is the baseline:
>
> http://members.aol.com/tempdon100164833/nikon/CoolPix.jpg
>
>And then the LS30's despicable mutilation, all settings at neutral -
>i.e. "factory defaults":
>
> http://members.aol.com/tempdon100164833/nikon/LS30.jpg
>
>According to Ed this is the "default color balance".
>Yes, right... Well, I don't think so...
>
between these two images as level adjustment in Photoshop readily
demonstrates!
Don
<r...@nospam.demon.co.uk> wrote:
>In article <3fe2a0cd...@news.individual.net>, Don
><phoney...@yahoo.com> writes
>>
>>First a CoolPix 950 picture of the slide, indistinguishable from the
>>way the slide actually looks i.e. this is the baseline:
>>
>> http://members.aol.com/tempdon100164833/nikon/CoolPix.jpg
>>
>>And then the LS30's despicable mutilation, all settings at neutral -
>>i.e. "factory defaults":
>>
>> http://members.aol.com/tempdon100164833/nikon/LS30.jpg
>>
>>According to Ed this is the "default color balance".
>>Yes, right... Well, I don't think so...
>>
>Well INCREASE THE EXPOSURE THEN because that is the only real difference
>between these two images as level adjustment in Photoshop readily
>demonstrates!
THAT DOESN'T REMOVE THE BLUE CAST! Can we stop shouting now? ;-)
But seriously, if it were only as simple as that...
Haven't you noticed the putrid blue cast in the LS-30 scan at all!?
Increasing the exposure blindly does nothing for the blue cast. All it
does is burn out the highlights. A hint... Go to the sky and you'll
notice the near total absence of red in the Nikon scan. Compare that
to the CoolPix sky. The ratio of colors is completely different.
What's needed is a complex boost of certain colors in analog gain, and
reduction of others, together with a boost to the combined RGB. And
still, there will be need for some post-processing. All that is a long
and convoluted hit-and-miss procedure for a civilian like me, but it's
a nanosecond for Nikon with all the equipment they have.
And since both the blue cast and the darkness are *constant* Nikon has
no excuse for not providing this for a major film type like
Kodachrome.
Don.
Kennedy McEwen
<phoney...@yahoo.com> writes
>On Fri, 19 Dec 2003 14:40:20 +0000, Kennedy McEwen
><r...@nospam.demon.co.uk> wrote:
>
>>In article <3fe2a0cd...@news.individual.net>, Don
>><phoney...@yahoo.com> writes
>>>
>>>First a CoolPix 950 picture of the slide, indistinguishable from the
>>>way the slide actually looks i.e. this is the baseline:
>>>
>>> http://members.aol.com/tempdon100164833/nikon/CoolPix.jpg
>>>
>>>And then the LS30's despicable mutilation, all settings at neutral -
>>>i.e. "factory defaults":
>>>
>>> http://members.aol.com/tempdon100164833/nikon/LS30.jpg
>>>
>>>According to Ed this is the "default color balance".
>>>Yes, right... Well, I don't think so...
>>>
>>Well INCREASE THE EXPOSURE THEN because that is the only real difference
>>between these two images as level adjustment in Photoshop readily
>>demonstrates!
>
>THAT DOESN'T REMOVE THE BLUE CAST! Can we stop shouting now? ;-)
>
>But seriously, if it were only as simple as that...
>
>Haven't you noticed the putrid blue cast in the LS-30 scan at all!?
>
is density, with the blue appearing more objectionable on the denser
image! Level adjustment of the LS-30 scan in PS results in an image
which is very similar to the Coolpix image apart from the obvious
quantisation, hence exposure is really all that differentiates them. You
can take any residual blue cast out by reducing the exposure increase of
the blue channel, but then your image will look distinctly less blue
than your Coolpix image and I have no doubt you'll be complaining about
that too!
Don
<r...@nospam.demon.co.uk> wrote:
We must not be looking at the same images...
You've conveniently avoided my remark about the red so here's a
sampling of colors of the sky:
LS30: R=1, G=29, B=89
CoolPix: R=94, G=140, B=232
As can be seen, quite unambiguously, the ratio of colors is completely
different. LS30 blue is 89x the red, while CoolPix blue is barely 2.5
the red.
No amount of exposure increase is going to bring this red up!
I know, I've done it! Many times! And here it is, the same image, LS30
scan with RGB analog gain boosted by 2 eV values:
http://members.aol.com/tempdon100164833/nikon/LS302EV.jpg
And here's a sampling of the same area of the sky:
LS30: R=1, G=56, B=173
You've, sadly, stopped being objective and are just sticking to your
story.
Don.
Kennedy McEwen
<phoney...@yahoo.com> writes
than one issue at a time in your posts! Your first complaint was about
the blue cast - that is what I addressed. I am sure that you are fully
aware that measuring colour balance on a fully saturated area of the
image is, at best misleading. However it does illustrate the point I
have been making: the spectral response of the LEDs in the LS-30 is NOT
the same as your digicam filters, and unlikely to be the same as later
scanners.
I really don't know what else you expect because you are now stating
that no amount of exposure adjustment will give you the results you want
but, in essence, with an 8-bit output and a 10-bit ADC, you don't really
have much more than the exposure of the individual channels to play
with.
Don
<r...@nospam.demon.co.uk> wrote:
Fair enough.
>Your first complaint was about
>the blue cast - that is what I addressed.
True. However, I don't see this blue cast in the CoolPix image.
But let me accept for a moment that there is one. You must agree that
the LS-30 blue cast is orders of magnitude greater/worse (let's ignore
the "why", for a moment). Can we at least agree on that?
By stating there is a blue cast in both images, you seem to minimize
(perhaps unintentionally) the magnitude of the LS-30 blue cast.
I want to settle this because, early on, both you and Ed seem to
indicate that there was no blue cast at all, and that what I was
seeing was really the "default color balance". And this view seems to
permeate still.
To avoid further confusion: This may be the "default color balance" in
the sense that the gray amorphous mess of a (true) raw digicam image
is the "default". However, when I request a raw image from a digicam I
know I'm going to get a raw image with all the warts. But when I do a
regular scan on an LS-30 I expect something far better. If this blue
and dark image were the result of a *raw* LS-30 scan, I wouldn't be
complaining at all. I hope that may clear things up a bit more.
>I am sure that you are fully
>aware that measuring colour balance on a fully saturated area of the
>image is, at best misleading. However it does illustrate the point I
>have been making: the spectral response of the LEDs in the LS-30 is NOT
>the same as your digicam filters, and unlikely to be the same as later
>scanners.
Granted, there are bound to be some differences between the 1D LS-30
array and the 2D CoolPix array. And, yes, I know that some digicams
may have two green elements per pixel to compensate for human's high
sensitivity in the green part of the spectrum, as well as other bells
and whistles, etc, etc. but the difference between LS-30 (a film
scanner with semi-pro aspirations - at least back when I bought it)
and a consumer digicam (bought around the same time) must be clearly
in the scanner's favor. And yet (ignoring the resolution) this
consumer camera beats this dedicated (semi-pro) film scanner!
Now, as you know, I very much doubt there are significant LED
differences between various LS models. In my view the ADC bit depth
between various LS models is far more significant than any LED
response differences which I still maintain are minimal, if any (we
have no proof either way, so let's leave it to one side).
Nevertheless, and I still keep coming back to this because it's the
essence on my original question: There is an optimum setting for LS-30
and Kodachrome, no matter how general. After all, the Kodachrome
setting on all subsequent models is also a compromise (even if they
have superior LEDs).
And that's all I'm after. An optimum, recommended setting for my
scanner, with all the limitations it may have, to give me a reasonable
starting point. Nikon's refusal to even acknowledge the problem, let
alone provide a solution, is in my view plain despicable.
>I really don't know what else you expect because you are now stating
>that no amount of exposure adjustment will give you the results you want
That's not what I said. You stated (I'm paraphrasing, *correct me if
I'm wrong*) that the only reason I find the blue cast objectionable is
because of the darkness (density) of the LS-30 image.
So, what I said in response is that no amount of extra exposure, i.e.
eV boost (i.e. lightening the image) will eliminate the LS-30 blue
cast. If anything, after I boosted exposure by 2 eV the blue cast
became even more pronounced.
>but, in essence, with an 8-bit output and a 10-bit ADC, you don't really
>have much more than the exposure of the individual channels to play
>with.
Which is why at the very beginning of this thread I acknowledge the
10-bit ADC limitations of the LS-30 and was looking for recommended
values to achieve optimum results given this limitation.
Don.
Kennedy McEwen
<phoney...@yahoo.com> writes
>
>True. However, I don't see this blue cast in the CoolPix image.
>
>But let me accept for a moment that there is one. You must agree that
>the LS-30 blue cast is orders of magnitude greater/worse (let's ignore
>the "why", for a moment). Can we at least agree on that?
>
of each image we get the Coolpix giving RGB of 186:200:200 while the
LS-30 image is 50:60:85 - that is hardly orders of magnitude, in fact,
it corresponds to a hue difference of around 27%.
>By stating there is a blue cast in both images, you seem to minimize
>(perhaps unintentionally) the magnitude of the LS-30 blue cast.
>
Certainly not deliberately, however the colours of the LS-30 are indeed
more saturated - even after levels adjustments to align the overall
brightness of both images.
>I want to settle this because, early on, both you and Ed seem to
>indicate that there was no blue cast at all, and that what I was
>seeing was really the "default color balance". And this view seems to
>permeate still.
>
I am sure that neither Ed nor myself have stated that there is no excess
blue in the LS-30 scan, though I know Ed is more than capable of
speaking for himself. However we have both stated that what you get
with Nikonscan is the default colour balance - the software has no
automatic white balance process, which Vuescan has. In the case of
conventional chromogenic emulsions that default produces a relatively
neutral colour balance. The spectral characteristics of the colour
couplers used in Kodachrome are very different and produce a different
colour characteristic when seen by the narrow spectral illumination of
the LS-30 LED system, which you assume has remained unchanged through
several generations of scanner design.
>
>>I am sure that you are fully
>>aware that measuring colour balance on a fully saturated area of the
>>image is, at best misleading. However it does illustrate the point I
>>have been making: the spectral response of the LEDs in the LS-30 is NOT
>>the same as your digicam filters, and unlikely to be the same as later
>>scanners.
>
>Granted, there are bound to be some differences between the 1D LS-30
>array and the 2D CoolPix array. And, yes, I know that some digicams
>may have two green elements per pixel to compensate for human's high
>sensitivity in the green part of the spectrum, as well as other bells
>and whistles, etc, etc. but the difference between LS-30 (a film
>scanner with semi-pro aspirations - at least back when I bought it)
>and a consumer digicam (bought around the same time) must be clearly
>in the scanner's favor. And yet (ignoring the resolution) this
>consumer camera beats this dedicated (semi-pro) film scanner!
>
Up till now I thought I was debating with someone who understood some of
the essence of colour imaging, but clearly I have been mistaken. To
begin with, lets consider a typical digicam CCD - not the one in your
Coolpix, rather that of the latest Olympus camera, but typical of the
genre. You can see its specification (watching for the line wrap) at:
http://www.kodak.com/global/plugins/acrobat/en/digital/ccd/products/fullf
rame/KAF-5101CELongSpec.pdf
I draw your attention to Figure 5, the spectral response of the device,
on page 12. Note the peaks and response widths of the red, green and
blue channels, and the amount of overlap that exists between them. (The
two, slightly different, green spectra are due to the mixing of dyes
depending on whether the green pixels are interleaved with red or blue
pixels in a given row.) Due to the smaller pixel footprint, the mixing
of the dyes in your Coolpix can be expected to exceed that of the Kodak
chip shown here, but it gives you an idea of the response.
Now lets look at the spectrum of some LEDs.
A red typical LED at http://ledlights.home.att.net/spectra/660.gif
A typical green LED at http://ledlights.home.att.net/spectra/etggrn.gif
A typical blue LED at http://ledlights.home.att.net/spectra/470.gif
Now, I hope that you can see the difference in the spectral
characteristics of the filters on a typical digicam CCD and the light
emitted by typical LEDs (not necessarily the LEDs in your LS-30, or any
Nikon scanner). For example, the width of the CCD blue, as defined
between 10% points, is around 180nm compared to the LED spectral band of
40nm. Similarly the greens on the CCD are around 140nm compared to the
LED 70nm whilst the red on the CCD compares over 240nm to the LED's 80nm
- all ignoring the infrared response of the CCD which causes further
colour mixing.
Now, whilst the colour response of these systems can be normalised to
match that of the eye, that can only be achieved for one particular
balance of light - a nominal white. But a good visual white can be
produced from many different spectral characteristics. A traditional
white is a black body emission corresponding to a particular
temperature, with all of the visible colours and wavelengths present.
For example, mean daylight is generally accepted as around the spectrum
emitted by a black body at 5600K, or around 5330degC. However, this
same white coloured light can also be produced for the eye by taking
three coloured LEDs, or the phosphors on your CRT screen, and mixing
their emission in the correct proportion. However, look at those three
LED spectra again - no matter how well you choose to mix them, not all
of the wavelengths of light will be present. In particular, there are
complete gaps between red and green from 570nm to 620nm, and then again
between green and blue from 480nm to 500nm, not to mention the peaks and
troughs where emissions exist - yet a visible white, indistinguishable
to either the eye, or the digicam's CCD for that matter, from natural
sunlit white can readily be produced.
It should not be difficult for you to guess which of these systems, the
filtered CCD or the LED illuminated CCD, will be much more susceptible
to changes in the detailed spectrum of the dyes used to produce the
image on a slide. The LED system may produce purer colours by only
sampling a section of the dye spectrum, but that benefit also means that
it is much more susceptible to changes in the detailed dye spectrum
being manifest as a change in colour balance which would be
imperceptible to a wider response system - metamerism.
Once again I return to the comments that you misconstrued from Ed and
myself previously - only one default white balance is provided on the
LS-30. That is the balance of the red green and blue LED illumination
which produces a neutral grey when exposed through the red green and
blue layers of dyes in the neutral greys in the most popular colour
emulsion, the chromogenic type. A different spectral characteristic for
the dyes will result in a different default colour balance in the image
- and this will be far more dominant in the LED illuminated system of
the LS-30 than it will appear to either a filtered digicam CCD or the
human eye, whose colour response you can find at:
http://www.photo.net/photo/edscott/vis00010.htm - note the amount of
cross-coupling between the nominal wavelengths that your eye has!
So Nikon had a choice for their scanners, which emulsion should they
normalise the CCD exposure for (they change the colour balance by
exposure differences to each LED, not changing the LED intensity which
also changes the emission spectra!). Now you wish they had selected a
neutral balance for Kodachrome but, despite its popularity, it is a
relatively minor player in the scheme of things, especially in the past
decade, and is completely different from the emulsions typified by the
E-6 process. So Nikon chose the E-6 emulsion as their basis for the
default colour balance and Kodachrome, because of the detailed spectral
characteristics of the dyes used, results in an off neutral colour
balance. Not only that, but the changes in the various dyes used in the
Kodachrome process over the years means that the degree off neutral is
highly variable. Later, when the LS-30 was superceded, Nikon addressed
the Kodachrome issue - and whilst you have suggested that the this is
only software, the very limitations of the LED process suggest that
there was more to it than that.
>Now, as you know, I very much doubt there are significant LED
>differences between various LS models.
Let's just examine some LED examples, just taking some red examples,
each quite a popular device in its time. Fortunately, someone has taken
the trouble to measure these for a different application, but the
results remain valid:
LED Technology L05R5000E1
http://cc.joensuu.fi/photobio/pdf/spectra/L05R5000E1.pdf
Quantum Devices QDDH68002
http://cc.joensuu.fi/photobio/pdf/spectra/QDDH68002.pdf
Quantum Devices QDDH66002
http://cc.joensuu.fi/photobio/pdf/spectra/QDDH66002.pdf
Hamamatsu L3882
http://cc.joensuu.fi/photobio/pdf/spectra/L3882.pdf
Kingbright L-793SRC-E
http://cc.joensuu.fi/photobio/pdf/spectra/L793SRCE.pdf
Now, do you still think that all LEDs are the same?
In order to achieve a more robust and repeatable KC colour balance what
do you THINK is required in terms of the LED spectrum?
Why would a KC profile produced for, for example, an illumination system
based on the QDDH66002 red LED bear any relevance to the profile of an
L05R5000E1?
Why should a correction profile produced for one LED scanner bear any
relevance to that of another?
No doubt it will come as a major surprise to you that the profile of the
LS-50 is completely different from that of the LS-4000 despite these
scanners being nominally identical in all other respects to the user and
merely a single generation apart! The only difference that would give
rise to a different colour profile (as can be demonstrated by viewing a
hex dump of their profiles) is the different spectral characteristics of
the light source and the CCD. Yet, comparing your LS-30 to these two
scanners is like comparing Concorde to a Cayley flyer - they operate on
the same principles (LED Illumination and aerodynamics) but beyond that
there is little in common.
>Nevertheless, and I still keep coming back to this because it's the
>essence on my original question: There is an optimum setting for LS-30
>and Kodachrome, no matter how general. After all, the Kodachrome
>setting on all subsequent models is also a compromise (even if they
>have superior LEDs).
>
Not necessarily superior, just different - and not obsolete, which is
the main issue why semiconductor users change models reasonably
regularly. Nevertheless, I keep coming back to the same question - that
last one in the list above!
>And that's all I'm after. An optimum, recommended setting for my
>scanner, with all the limitations it may have, to give me a reasonable
>starting point.
But there is no optimum recommended for your scanner - that is it, pure
and simple, down to the end user. Too many variables to cope with! As
I pointed out to you way back in the thread, even the images that Rafe
posted from a much more recent scanner show variability between KC
emulsions - and that is for a scanner where Nikon have gone to some
trouble to produce something like an acceptable medium, something like a
profile which varies a limited amount across a range of dye spectra.
> Nikon's refusal to even acknowledge the problem, let
>alone provide a solution, is in my view plain despicable.
>
In your view it may be. In reality it is not so. How many
manufacturers provide a free upgrade, or the information to achieve
that, from one standard of device to match the capability of a later
device? Would you expect them to upgrade the CCD in your Coolpix 950
with that of a D1x? In fairness, as I suggested earlier, they might
have told you that it was not possible - however their refusal to be
drawn on the issue simply reflects the evolution of the technology. They
cannot provide what you want because it simply does not exist - and even
if to provide you with the profile differences of a more recent scanner
for you to experiment with would only lead to further embroilment in an
issue which there is no value in them getting into - they no longer
support your scanner. Plain, simple - whether you like it or not is
irrelevant.
Don
<r...@nospam.demon.co.uk> wrote:
>>True. However, I don't see this blue cast in the CoolPix image.
>>
>>But let me accept for a moment that there is one. You must agree that
>>the LS-30 blue cast is orders of magnitude greater/worse (let's ignore
>>the "why", for a moment). Can we at least agree on that?
>>
>No, it wouldn't appear that we can. Taking a relatively mid-toned area
>of each image we get the Coolpix giving RGB of 186:200:200 while the
>LS-30 image is 50:60:85 - that is hardly orders of magnitude, in fact,
>it corresponds to a hue difference of around 27%.
That to me is a major difference. If it were only a couple of %
points, I would agree with you, but over a *quarter* more blue is a
major cast.
And, indeed, stepping back and looking at the images (I know about
inaccuracies of perception but a head-to-head comparison is still
legitimate in this case - see below) there is no perceived blue cast
in the CoolPix image, while it's quite apparent in the +2 eV LS-30
one.
>>By stating there is a blue cast in both images, you seem to minimize
>>(perhaps unintentionally) the magnitude of the LS-30 blue cast.
>>
>Certainly not deliberately, however the colours of the LS-30 are indeed
>more saturated - even after levels adjustments to align the overall
>brightness of both images.
Saturation is the other side of the coin but it can be ameliorated
(relatively easily) by modifying exposure. The blue cast problem is
much more unwieldy which is why I expect Nikon to provide recommended
values based on objective measurements.
>Now, I hope that you can see the difference in the spectral
>characteristics of the filters on a typical digicam CCD and the light
>emitted by typical LEDs
...
>Now, whilst the colour response of these systems can be normalised to
>match that of the eye, that can only be achieved for one particular
>balance of light - a nominal white.
...
>It should not be difficult for you to guess which of these systems, the
>filtered CCD or the LED illuminated CCD, will be much more susceptible
>to changes in the detailed spectrum of the dyes used to produce the
>image on a slide.
I think you misunderstood the purpose for posting the CoolPix image.
I may not have the actual measurements ready at hand as you do but I'm
certainly well aware of the big picture regarding differences between
filtered CCD and LED illuminated CCD response and have absolutely no
problem following what you write.
All that notwithstanding, however, the purpose of the CoolPix image
was to show what the slide actually looks like. It was *not* to
compare relative *technical* merits of the two devices. It other
words, the CoolPix image was just *control*. I can't show you the
slide itself so the CoolPix image is the next best thing. That's all!
Otherwise, I could be accused of using an underexposed slide with a
blue cast. (You went on about me not showing any evidence at the start
of all this.)
By producing a CoolPix image I establish a *baseline* against which
the scanner image is to be judged. Or, more accurately, the scanner
image Nikon chose to present from an LS-30 Kodachrome scan. That's
why, if you read the thread back, you'll see I stress the word
*baseline*.
Now, I understand you may wish to bring up technical differences
between the two devices, but that's not at discussion here. What is,
is Nikon's failure to provide recommended values to produce a balanced
LS-30 Kodachrome scan. I'm not asking for an identical image to the
CoolPix one, but a reasonable image given LS-30 limitations. And I
stress that repeatedly in the thread.
You already agreed that such a general setting is possible (with all
the caveats attached). Just like Nikon provides Kodachrome settings
for subsequent scanners (with all appropriate caveats also attached).
>Once again I return to the comments that you misconstrued from Ed and
>myself previously - only one default white balance is provided on the
>LS-30.
And therein lies the problem! Since this is all under software control
there is no excuse for Nikon not providing a Kodachrome balance. It
should have been there from the start, but I'll take a retrofit
gladly. Instead, Nikon refuses to even acknowledge the problem.
>Now you wish they had selected a
>neutral balance for Kodachrome but, despite its popularity, it is a
>relatively minor player in the scheme of things, especially in the past
>decade, and is completely different from the emulsions typified by the
>E-6 process.
Actually, I think the reasons are probably more mundane. Nikon is a
Japanese company and Kodak is essentially non-existent in Japan. (One
of the bones of contention in trade talks involving Japan's
protectionism.) And since Japan (in audio-video and mobile
communications, to name two) is so far ahead it wasn't until the rest
of the world caught on and started using Nikon film scanners that the
Kodachrome problem became an issue.
This creates the situation where a consumer product (CoolPix) is
superior to a semi-pro piece of equipment (LS-30) - all technical
differences notwithstanding - and that just seems wrong. That was the
gist of my statement when I made the (non-technical) "comparison"
(actually, established a baseline).
Your point was that I can't expect Nikon to support an obsolete
product and I'm simply saying this product never worked to start with!
>>Now, as you know, I very much doubt there are significant LED
>>differences between various LS models.
>
>Let's just examine some LED examples
...
>Now, do you still think that all LEDs are the same?
...
>No doubt it will come as a major surprise to you that the profile of the
>LS-50 is completely different from that of the LS-4000 despite these
>scanners being nominally identical in all other respects to the user and
>merely a single generation apart!
That is not relevant on two counts:
1. Even if the differences exist, they are minor. The basic technology
is still the same. It's still the same paradigm.
2. The settings are all under software control. This means it's really
up to Nikon. If they so chose, they could (in theory) produce a
balance for every batch of Kodachrome out there. Of course, that would
be silly, but it is possible. However, given that, there is absolutely
no excuse for leaving Kodachrome out completely on the LS-30.
There may be another reason: My disassembly of NikonScan shows that
there are basically 3 modules to handle the setup. These are based on
*connection* type! FireWire and USB modules have a Kodachrome setting,
SCSI one doesn't. So, my guess (and it's only a guess at this point as
the disassembly is a long term project) is that Nikon was simply lazy
and didn't feel like messing with a module (SCSI) which had most bugs
ironed out. In NikonScan4 they, apparently, abandoned it completely.
>>Nevertheless, and I still keep coming back to this because it's the
>>essence on my original question: There is an optimum setting for LS-30
>>and Kodachrome, no matter how general. After all, the Kodachrome
>>setting on all subsequent models is also a compromise (even if they
>>have superior LEDs).
>>
>Not necessarily superior, just different - and not obsolete, which is
>the main issue why semiconductor users change models reasonably
>regularly. Nevertheless, I keep coming back to the same question - that
>last one in the list above!
(Why should a correction profile produced for one LED scanner bear any
relevance to that of another?)
The relevance is in the basic underlying technology. You used an
automotive metaphor early on, so let me use one too. It's like the
difference between driving different car models. The controls may be
slightly different, but the paradigm is still the same.
>But there is no optimum recommended for your scanner - that is it, pure
>and simple, down to the end user. Too many variables to cope with! As
>I pointed out to you way back in the thread, even the images that Rafe
>posted from a much more recent scanner show variability between KC
>emulsions - and that is for a scanner where Nikon have gone to some
>trouble to produce something like an acceptable medium, something like a
>profile which varies a limited amount across a range of dye spectra.
But that's exactly what I'm entitled to and it proves that it is
possible. With all the warts and caveats it, nevertheless, gives a
decent starting point. That's all I'm asking. The lion share of the
(known and mundane) work is done and I can take care of fine tuning.
Compare that to the LS-30 image where I always have to start from the
bottom of the hill.
>> Nikon's refusal to even acknowledge the problem, let
>>alone provide a solution, is in my view plain despicable.
>>
>In your view it may be. In reality it is not so. How many
>manufacturers provide a free upgrade, or the information to achieve
>that, from one standard of device to match the capability of a later
>device?
But that's not what I'm asking. Retrofitting Kodachrome on the LS-30
is not a "free upgrade" but to *fix* a flaw!
The problem with your argument is twofold:
1. The early model (LS-30 in this case) has a flaw (Kodachrome).
2. There is a generic setting to correct this flaw (at least in part,
but at the very least produce acceptable results).
That's what I hold Nikon responsible for.
And, of course, Nikon still don't acknowledge there is a problem in
the first place. When I mentioned the well documented problem LS-30
(and other Nikon scanners) had with Kodachrome demonstrated on various
Web sites, Nikon "support" became very testy claiming "those people
don't know how to scan Kodachrome". *Months* later Nikon came with a
"solution" considerably inferior to what those "ignoramuses" on the
Net or me, a civilian, came up with.
That I squarely blame Nikon for!
Don.
P.S. I'm very grateful to you for taking the trouble and the time to
illustrate your points in such detail. Even though I may not think
parts of it relevant for the particular item of discussion at hand I
really appreciate the effort! Be assured that I not only follow it all
but I read it with interest. Thanks!
Kennedy McEwen
<phoney...@yahoo.com> writes
>On Mon, 22 Dec 2003 22:54:21 +0000, Kennedy McEwen
><r...@nospam.demon.co.uk> wrote:
>
>>>True. However, I don't see this blue cast in the CoolPix image.
>>>
>>>But let me accept for a moment that there is one. You must agree that
>>>the LS-30 blue cast is orders of magnitude greater/worse (let's ignore
>>>the "why", for a moment). Can we at least agree on that?
>>>
>>No, it wouldn't appear that we can. Taking a relatively mid-toned area
>>of each image we get the Coolpix giving RGB of 186:200:200 while the
>>LS-30 image is 50:60:85 - that is hardly orders of magnitude, in fact,
>>it corresponds to a hue difference of around 27%.
>
>That to me is a major difference. If it were only a couple of %
>points, I would agree with you, but over a *quarter* more blue is a
>major cast.
>
"orders of magnitude greater/worse" - it is exactly 27% difference in
hue!
>
>Saturation is the other side of the coin but it can be ameliorated
>(relatively easily) by modifying exposure.
No, saturation CANNOT be ameliorated by modifying exposure, at least not
the master exposure - and the equations which translate RGB to HSI prove
that quite clearly:
S= (Max(R,G,B) - Min(R,G,B)) / Max(R,G,B)
Thus with R=8, G=10, B=20, S=(20 - 8)/20 = 0.6
Increasing the exposure on all three channels equally, by two EV, will
result in values around R=32, G=40, B=80, and again S=0.6 (the
approximation is only one of rounding since the additional exposure
permits more precision of the data than simply multiplicative factors on
the original quantisation, however this only affects the precision of
the saturation, not its magnitude).
Saturation CAN be affected by brightness (an overall offset of the data
up or down in level), contrast (if the datum is non-zero) or by gamma
modifications, but NOT by exposure (which is essentially a contrast
modification with a zero datum.
Low noise, high saturation is one of the key features that LED
illumination provides since the waveband of each channel does not leak
into the other two colours, thus artificially increasing the value of
Min(R,G,B) and hence reducing saturation. Without this clear separation
of the sensor colours, the actual R, G & B values of the image must be
estimated by removing the crosstalk that occurs between the channels for
a given balance, generally one of the standard blackbody temperature
whites. Such differential processing, simply to produce the primary
colours, inevitably reduces the signal to noise ratio of the resulting
data.
>I think you misunderstood the purpose for posting the CoolPix image.
>
>All that notwithstanding, however, the purpose of the CoolPix image
>was to show what the slide actually looks like. It was *not* to
>compare relative *technical* merits of the two devices. It other
>words, the CoolPix image was just *control*. I can't show you the
>slide itself so the CoolPix image is the next best thing. That's all!
>
That may well be what you originally intended it for, however you then
went on to write about how the Coolpix could cope readily with
Kodachrome characteristics whilst the Coolscan could not. In short,
your argument exceeded the purpose for which a control is intended. That
is why I posted the explanation of WHY the Coolpix and Coolscan systems
respond completely differently to spectral variations. One has a much
more similar spectral characteristic (although still significantly
different) to the human eye than the other BECAUSE one is intended to
operate under a range of spectra present in the real world, whilst the
other only has to deal with very specific dye spectra and produce the
highest signal to noise image under those limited range of
circumstances. Indeed, it could do so equally well with only three
carefully chosen specific wavelengths, such as laser scanning systems
achieve.
>Otherwise, I could be accused of using an underexposed slide with a
>blue cast. (You went on about me not showing any evidence at the start
>of all this.)
>
You are deliberately twisting my comments now - I specifically referred
to evidence supporting your claim that the change is entirely software,
not about whether the image you used was any particular colour! There
has never been any suggestion, other than from yourself, that you should
get a KC image which has no colour cast from an LS-30 scanner using the
standard settings! I have, at no time, suggested that your problem was
associated with a colour cast in the original material!
>By producing a CoolPix image I establish a *baseline* against which
>the scanner image is to be judged.
However, whilst your baseline shows the effect of the scan you achieve,
you then proceeded to extend that comparison to how the two imaging
systems perform - which goes well BEYOND a mere control, and is totally
unsuitable for that purpose for the reasons I explained - it is not
possible for the LED system in the LS-30 to cope with the spectral
differences between KC and conventional films with one default profile.
However your statement "this consumer camera beats this dedicated
(semi-pro) film scanner!" indicates that you expect it to.
Furthermore, the entire paragraph of which that is merely your final
sentence indicates a complete misunderstanding (or at least lack of
knowledge, whether misunderstood or not) of the significance of the
difference between the LED spectral purity and the bandwidth of the
filters used in digicams - or conventional tri-linear CCD scanners for
that matter!
>
>Now, I understand you may wish to bring up technical differences
>between the two devices, but that's not at discussion here. What is,
>is Nikon's failure to provide recommended values to produce a balanced
>LS-30 Kodachrome scan. I'm not asking for an identical image to the
>CoolPix one, but a reasonable image given LS-30 limitations. And I
>stress that repeatedly in the thread.
>
But you assume that Nikon have, or are prepared to invest time and money
in providing that for an obsolete scanner, in fact one of the least
capable scanners for which they do not provide KC support.
You also assume that only ONE such profile exists.
It is clear that neither assumption is valid and the evidence, which
should have ended the discussion at the time has been made available to
you much earlier in the thread - as well as being repeated many times
for your benefit.
>You already agreed that such a general setting is possible (with all
>the caveats attached).
No, I have NOT!! I have been at pains to point out the reasons why a
GENERAL setting is NOT possible given all of the variations in
Kodachrome spectral characteristics! One generic setting MAY suit some
of the KC stock that YOU personally have, if you are very fortunate it
might even suit all of it - but are Nikon expected to determine what
that is through some process of ESP? Even if you provided your local
Nikon store or service centre a specific example of YOUR emulsion, and
they passed this back to the Nikon engineering team, what value is there
in them determining a SPECIFIC balance for it? The sample you provide
may not even be close to a mid-range KC image from your own stock, let
alone the entire KC production that has been processed over the years!
> Just like Nikon provides Kodachrome settings
>for subsequent scanners (with all appropriate caveats also attached).
>
Scanners which have been designed to be capable of doing so, not
scanners that you WISH were, but which obviously were not!
>>Once again I return to the comments that you misconstrued from Ed and
>>myself previously - only one default white balance is provided on the
>>LS-30.
>
>And therein lies the problem! Since this is all under software control
>there is no excuse for Nikon not providing a Kodachrome balance. It
>should have been there from the start, but I'll take a retrofit
>gladly.
How many white balances are you prepared to pay for? Indeed, have you
even offered to pay for any of these additional balances which you
require over and above the default balance provided?
You assume that one will do the job, however the variation of KC
spectral characteristics and the narrow spectrum of the LEDs used in the
LS-30 means this is impossible. Indeed, if only one modified white
balance did the job then I am sure that Nikon would have addressed the
problem long ago (as Ed pointed out to you very early in this thread!)
> Instead, Nikon refuses to even acknowledge the problem.
NO, Nikon HAVE acknowledged the problem - and provided a solution. The
problem is that you are not prepared to pay for it and expect some
special treatment.
Similarly, 3rd party software providers have also provided solutions for
the scanner that you have, but yet again, you are not prepared to use
those.
In short, there are several options available to address the problem you
have, including buying a tri-linear CCD based scanner from another
manufacturer.
>
>>Now you wish they had selected a
>>neutral balance for Kodachrome but, despite its popularity, it is a
>>relatively minor player in the scheme of things, especially in the past
>>decade, and is completely different from the emulsions typified by the
>>E-6 process.
>
>Actually, I think the reasons are probably more mundane. Nikon is a
>Japanese company and Kodak is essentially non-existent in Japan. (One
>of the bones of contention in trade talks involving Japan's
>protectionism.) And since Japan (in audio-video and mobile
>communications, to name two) is so far ahead it wasn't until the rest
>of the world caught on and started using Nikon film scanners that the
>Kodachrome problem became an issue.
>
Now you're assuming that Nikon are incapable of assessing their major
market sector! Sales of scanners in the US (even excluding the rest of
the world) swamp those in Japan. Even in the US however, chromogenic
emulsions vastly outnumber Kodachrome - and have done so for years.
Partly this is because the KC development process is so variable that
Kodak themselves had difficulty maintaining it in more than a couple of
plants, let alone document it for general release, and partly it is for
commercial reasons of maintaining a closed in-house standard.
>This creates the situation where a consumer product (CoolPix) is
>superior to a semi-pro piece of equipment (LS-30) - all technical
>differences notwithstanding - and that just seems wrong.
Rubbish!! What you are suggesting is that all Japanese film scanners of
a similar vintage would be incapable of working with KC because of
US/Japanese trade wars! Absolute bunkum! Canon, for example, had film
scanners on the market since before the LS-30 was launched and have no
problem scanning KC. Minolta and Olympus similarly - both also Japanese
manufacturers. Many other professional scanners are manufactured in
Japan. The technical approach taken by Nikon on their scanner line
determines its spectral capabilities, it has nothing to do with the
manufacturing country. Even a recent LED scanner, such as the LS-5000,
will not handle the colour balance of all KC emulsions as uniformly as a
digital camera. This has NOTHING to do with the country of manufacture
(and guess where your Coolpix camera was made) but EVERYTHING to do with
the technical differences.
>
>Your point was that I can't expect Nikon to support an obsolete
>product and I'm simply saying this product never worked to start with!
>
It worked, and continues to work, perfectly well within its
capabilities! Indeed, its predecessors, the LS-20 (which I had) and the
LS-1000, and its upmarket partner the LS-2000 also worked perfectly
well, within their capabilities. None handled KC by default.
>>
>>Let's just examine some LED examples
>...
>
>>Now, do you still think that all LEDs are the same?
>...
>>No doubt it will come as a major surprise to you that the profile of the
>>LS-50 is completely different from that of the LS-4000 despite these
>>scanners being nominally identical in all other respects to the user and
>>merely a single generation apart!
>
>That is not relevant on two counts:
>
>1. Even if the differences exist, they are minor. The basic technology
>is still the same. It's still the same paradigm.
>
>2. The settings are all under software control. This means it's really
>up to Nikon.
You clearly still don't understand the issue of spectral response - but
there is little point in labouring this further. Suffice to say that
your assumption that the only difference is only software is naive, at
best - the software settings are only available for scanners which can
cope with the difference.
> If they so chose, they could (in theory) produce a
>balance for every batch of Kodachrome out there. Of course, that would
>be silly, but it is possible.
Well, at least you understand that now! So why do you still insist that
the difference between your scanner and more recent KC capable devices
is merely software?
> However, given that, there is absolutely
>no excuse for leaving Kodachrome out completely on the LS-30.
>
So the fact that they may need dozens of KC settings, with no way for
the end user to tell which related to any of their emulsions, is not an
issue for you? I can tell you it would certainly be an issue for many
of their users! Without analysing the emulsion, not even Kodak can tell
you where any particular KC film was processed - it was shipped all over
the world depending on which plant had most capacity at the time.
>There may be another reason: My disassembly of NikonScan shows that
>there are basically 3 modules to handle the setup. These are based on
>*connection* type! FireWire and USB modules have a Kodachrome setting,
>SCSI one doesn't. So, my guess (and it's only a guess at this point as
>the disassembly is a long term project) is that Nikon was simply lazy
>and didn't feel like messing with a module (SCSI) which had most bugs
>ironed out. In NikonScan4 they, apparently, abandoned it completely.
>
As opposed to the more logical guess that the older scanners which used
the older SCSI interface SCSI, simply were incapable of handling KC with
a single colour balance.
>(Why should a correction profile produced for one LED scanner bear any
>relevance to that of another?)
>
>The relevance is in the basic underlying technology. You used an
>automotive metaphor early on, so let me use one too. It's like the
>difference between driving different car models. The controls may be
>slightly different, but the paradigm is still the same.
>
That is irrelevant: the basic principle of scanning and image is exactly
the same as that developed by Harry Nyquist in 1921 for the electronic
transmission of images across telegraph wires - but Bell Labs didn't
complain that his original equipment didn't produce the same quality of
image as a 1980s FAX machine, let alone the excellent capability of your
LS-30!
Similarly the engine management software the latest Lexus will be
absolutely no use on a Ford Model-T! Even unleaded fuel is incompatible
with car engines of only a few generations back. The basic operation
and controls of most cars haven't changed much in the years I have been
driving, but both the detail of the technology and their resulting
capability has changed dramatically. Just the same as scanners,
including LED scanners. Just because the principle is the same means
very little, particularly in terms of backwards compatibility - which is
the key issue here.
>>But there is no optimum recommended for your scanner - that is it, pure
>>and simple, down to the end user. Too many variables to cope with! As
>>I pointed out to you way back in the thread, even the images that Rafe
>>posted from a much more recent scanner show variability between KC
>>emulsions - and that is for a scanner where Nikon have gone to some
>>trouble to produce something like an acceptable medium, something like a
>>profile which varies a limited amount across a range of dye spectra.
>
>But that's exactly what I'm entitled to
No it isn't!
>and it proves that it is
>possible.
No it doesn't!
It proves that a perfect balance across that range of identically named
emulsions is not even possible with the latest scanners, let alone the
long obsolete LS-30!
> With all the warts and caveats it, nevertheless, gives a
>decent starting point. That's all I'm asking. The lion share of the
>(known and mundane) work is done and I can take care of fine tuning.
>Compare that to the LS-30 image where I always have to start from the
>bottom of the hill.
>
Why should you be any different to anyone else that used the same
scanner? Others, including myself, had the same problems with that and
similar generation systems! And, to be frank, KC colour balance was one
of the least significant issues! We all developed our own techniques
for dealing with the limitations of the technology - just as we do
today, although the limitations are now different.
Why do you assume, given the documented variability of KC spectral
characteristics, that an "official LS-30 Kodachrome" balance would be
any closer to a useful profile than the ones that you, and I and
everyone else who has ever used older Nikon LED scanners with KC,
developed manually based on the developed emulsion stock that we own?
You yourself have used the term "hit and miss" to describe the success
of your own adjustments - stop and think about why that is the case for
a minute! There is only one possible reason why your own adjustment is
not "standard" for all of your stock, the same reason as an "official
LS-30 Kodachrome" setting would be just as "hit and miss" - and I have
been trying to explain that to you for the past dozen or so messages in
this thread!
>
>But that's not what I'm asking. Retrofitting Kodachrome on the LS-30
>is not a "free upgrade" but to *fix* a flaw!
>
Whilst it may well be a limitation, it certainly is not a flaw.
>The problem with your argument is twofold:
>1. The early model (LS-30 in this case) has a flaw (Kodachrome).
That is actually a flaw in your entire thesis. Where, in any Nikon
literature for the scanner, does it state that the device will provide
perfect KC scans under default settings? Unless that is specifically
stated then it is NOT a flaw. YOU have assumed that since it operates
fine with one type of emulsion then it will operate fine with all - that
is a flaw in your logic, not in the scanner!
>2. There is a generic setting to correct this flaw (at least in part,
>but at the very least produce acceptable results).
>
You ASSUME that there is such a generic setting - an assumption without
ANY supporting evidence.
>That's what I hold Nikon responsible for.
>
Supplying you with what they offered instead of supplying what you
thought you wanted. If that was a trading requirement then NOBODY would
be in business anywhere! It isn't Nikon's fault you bought something
that did not do the job you wanted - only one person is to blame for not
understanding, specifying and researching the capability of their chosen
solution prior to purchase - and you'll see him facing you in the
mirror!
Don
>I haven't suggested it is NOT a major blue cast, but it certainly isn't
>"orders of magnitude greater/worse" - it is exactly 27% difference in
>hue!
Well, then we're down to definitions. What to you is an "order of
magnitude"?
To me, when one steps up from 1s (as is 1%) to 10s (as in 27%) that's
an order of magnitude.
>Increasing the exposure on all three channels equally, by two EV, will
>result in values around R=32, G=40, B=80, and again S=0.6 (the
>approximation is only one of rounding since the additional exposure
>permits more precision of the data than simply multiplicative factors on
>the original quantisation, however this only affects the precision of
>the saturation, not its magnitude).
Actually, I agree with you there because I've discovered this myself a
long time ago. That's also the very reason I'm not satisfied with a
"blind" increase in analog gain but wanted Nikon to provide
recommended values.
Increasing master (RGB) analog gain does indeed change the color
balance, which actually surprised me so I find your explanation a
pleasant serendipitous discovery I will pursue further.
>>I think you misunderstood the purpose for posting the CoolPix image.
>>
>>All that notwithstanding, however, the purpose of the CoolPix image
>>was to show what the slide actually looks like. It was *not* to
>>compare relative *technical* merits of the two devices. It other
>>words, the CoolPix image was just *control*. I can't show you the
>>slide itself so the CoolPix image is the next best thing. That's all!
>>
>That may well be what you originally intended it for, however you then
>went on to write about how the Coolpix could cope readily with
>Kodachrome characteristics whilst the Coolscan could not.
In the context of commercial application of technology!
You yourself changed the context to the commercial domain and those
comments were my response to that.
One of the problems here is that you keep changing the subject. Every
time I ask a simple question you descend into technical minutiae
ending up answering a *different* question I never asked. So, once
again, some consolidation is in order to get back to the subject...
>However your statement "this consumer camera beats this dedicated
>(semi-pro) film scanner!" indicates that you expect it to.
Again, those were side comments made in a different context and not
related to technical merits. Simple mention of "consumer" and
"professional" is a clear indication of their non-technical nature.
>>Now, I understand you may wish to bring up technical differences
>>between the two devices, but that's not at discussion here. What is,
>>is Nikon's failure to provide recommended values to produce a balanced
>>LS-30 Kodachrome scan. I'm not asking for an identical image to the
>>CoolPix one, but a reasonable image given LS-30 limitations. And I
>>stress that repeatedly in the thread.
>>
>But you assume that Nikon have, or are prepared to invest time and money
>in providing that for an obsolete scanner, in fact one of the least
>capable scanners for which they do not provide KC support.
I think you vastly exaggerate the effort it takes to create a
generalized Kodachrome *setting* (I'm not, and never have been,
talking about *profiles*). With the equipment available to Nikon staff
any low level technician could come up with one.
Also, this should have been done years ago when the scanner was not
obsolete!
>You also assume that only ONE such profile exists.
Indeed I do, given the parameters of "general settings" and the fact
that *all* subsequent Nikon scanners only provide *one* (general)
Kodachrome profile/setting.
How many Kodachrome *settings* do you have in NikonScan for your
scanner? One! Well, there you go! *That's* what I'm taking about.
This is a case of you answering a question *not* asked. I never said
that there is one "perfect" *profile*, but I did say there exists one
"general" *setting*.
I'm taking about a "happy medium", a "general setting", an "average",
a "one size fits all", the "least of all evils"... That is not only
possible but, I repeat, it exists in all subsequent scanners.
It's simply axiomatic that a generalized profile does exist once
parameters are wide enough. Inexact? Yes! Imperfect? Yes! Better than
what Nikon provides now for LS-30 (i.e. nothing)? A definitive, yes!
>>You already agreed that such a general setting is possible (with all
>>the caveats attached).
>
>No, I have NOT!!
OK, what do you call the *single* Kodachrome setting available for
scanners which Nikon decided to support?
>> Just like Nikon provides Kodachrome settings
>>for subsequent scanners (with all appropriate caveats also attached).
>>
>Scanners which have been designed to be capable of doing so, not
>scanners that you WISH were, but which obviously were not!
But that contradicts what you just wrote above. You can't have it both
ways...
According to you, Kodachrome is a moving target and the differences
between different stock is not linear, therefore no one setting can be
possible or available to cover all.
And then you claim that subsequent scanners "have been designed to be
capable of doing so". How? Do those scanners have magic LEDs? Those
scanners also have limitations and yet a "generalized" Kodachrome
setting has be created for them.
It's just common sense that once all the caveats are clearly stated,
there is an *optimum* setting! A perfectionist may not be happy with
it, but it's not made for perfectionists. That's why it's called
"general".
Even if the LS-30 LEDs only had 1 bit of color depth, there is an
optimum setting. It will be a compromise, but there always is. It's
just so axiomatic and I don't understand why you fail to see that?
>>Actually, I think the reasons are probably more mundane. Nikon is a
>>Japanese company and Kodak is essentially non-existent in Japan. (One
>>of the bones of contention in trade talks involving Japan's
>>protectionism.) And since Japan (in audio-video and mobile
>>communications, to name two) is so far ahead it wasn't until the rest
>>of the world caught on and started using Nikon film scanners that the
>>Kodachrome problem became an issue.
>>
>Now you're assuming that Nikon are incapable of assessing their major
>market sector!
It's all black or white with you, isn't it? I make a general side
comment to calm things down and you immediately take it to the extreme
making all sorts of assumptions and accusations.
Are you familiar with the "first to market" concept? Or "first mover"?
There are many marketing reasons why unfinished products are released
but I don't think we should digress into yet another side issue...
>> However, given that, there is absolutely
>>no excuse for leaving Kodachrome out completely on the LS-30.
>>
>So the fact that they may need dozens of KC settings, with no way for
>the end user to tell which related to any of their emulsions, is not an
>issue for you? I can tell you it would certainly be an issue for many
>of their users! Without analysing the emulsion, not even Kodak can tell
>you where any particular KC film was processed - it was shipped all over
>the world depending on which plant had most capacity at the time.
So how come there is only *one* Kodachrome setting on all subsequent
Nikon scanners?
>>There may be another reason: My disassembly of NikonScan shows that
>>there are basically 3 modules to handle the setup. These are based on
>>*connection* type! FireWire and USB modules have a Kodachrome setting,
>>SCSI one doesn't. So, my guess (and it's only a guess at this point as
>>the disassembly is a long term project) is that Nikon was simply lazy
>>and didn't feel like messing with a module (SCSI) which had most bugs
>>ironed out. In NikonScan4 they, apparently, abandoned it completely.
>>
>As opposed to the more logical guess that the older scanners which used
>the older SCSI interface SCSI, simply were incapable of handling KC with
>a single colour balance.
Well, that's a stretch, isn't it. Since when is color balance a factor
of interface type?
Besides, aren't there models of Nikon scanners with multiple
connection methods? That is to say, the *same* scanner offered with
different connections.
That also shoots down your basic premise that there "may" be something
"magical" about LEDs in scanners which have the Kodachrome option in
NikonScan.
>It proves that a perfect balance across that range of identically named
>emulsions is not even possible with the latest scanners, let alone the
>long obsolete LS-30!
Once again, how do you reconcile that with only *one* Kodachrome
setting on all those Nikon scanners?
>Why do you assume, given the documented variability of KC spectral
>characteristics, that an "official LS-30 Kodachrome" balance would be
>any closer to a useful profile than the ones that you, and I and
>everyone else who has ever used older Nikon LED scanners with KC,
>developed manually based on the developed emulsion stock that we own?
For the simple reason that I don't have the equipment Nikon does. All
my results, even though *vastly* superior to what Nikon's so-called
"support" produced, are in the end only my own *guesses* with all the
imperfections in the link of causalities, from my own eyes, to
inaccuracies in the calibration of my equipment, etc, etc.
Take by comparison a Nikon scanner which does have a Kodachrome
setting. It's an objective setting arrived at with the scientific
method, rather than my own (or anybody else's) guesswork.
>You yourself have used the term "hit and miss" to describe the success
>of your own adjustments - stop and think about why that is the case for
>a minute!
I thought that was self-evident!
It's because they are *subjective guesses*! I have no measuring
equipment to arrive at objective values like Nikon does. It is this
they used to come up with *one* general Kodachrome setting on
subsequent scanners.
>>The problem with your argument is twofold:
>>1. The early model (LS-30 in this case) has a flaw (Kodachrome).
>
>That is actually a flaw in your entire thesis. Where, in any Nikon
>literature for the scanner, does it state that the device will provide
>perfect KC scans under default settings?
Or more accurately:
Where, in any Nikon literature for the scanner, does it state that the
device will NOT provide KC scans out of the box?
>>2. There is a generic setting to correct this flaw (at least in part,
>>but at the very least produce acceptable results).
>>
>You ASSUME that there is such a generic setting - an assumption without
>ANY supporting evidence.
Except for the fact that all subsequent Nikon scanners only have just
*one* such general Kodachrome setting.
Don.
Kennedy McEwen
<phoney...@yahoo.com> writes
>Every
>time I ask a simple question you descend into technical minutiae
>ending up answering a *different* question I never asked.
Don, there is little point in discussing this further - I have explained
to you the reasons why your expectation is impractical with the device
you have, and I suspect that Nikon have done much the same too.
You consider such explanations to be "technical minutiae", yet you still
ramble on about "subsequent scanners", "nonlinearity", "bit-depth",
"optimum settings", "magic LEDs" etc. which all indicate that you have
not understood any of the issues involved. I suggest that you read over
the thread again in slow time, in particular those areas concerning
spectral bands of emulsion dyes and LED emission spectra and also search
out further explanations of metamerism and its causes.
When you eventually do understand the issues, and it is not difficult so
I am confident that you will, it should then be obvious why a
manufacturer endorsed setting for your scanner, without hardware
modification, cannot be any less variable than the settings you have
already derived yourself - but which you describe as "hit and miss". You
*might* then understand why such a manufacturer would not be prepared to
endorse such a "hit and miss" setting.
Sadly, however, I suspect that you will continue to stamp your feet and
blame Nikon for not giving you what you demand when in fact they have
upgraded their scanner range *including the hardware* to provide exactly
what you demand - albeit before you even realised what you were going to
demand!
Don
><phoney...@yahoo.com> writes
>>Every
>>time I ask a simple question you descend into technical minutiae
>>ending up answering a *different* question I never asked.
>
>Don, there is little point in discussing this further - I have explained
>to you the reasons why your expectation is impractical with the device
>you have, and I suspect that Nikon have done much the same too.
No, they haven't. Quite the contrary! Once I progressed to the
so-called "3rd-level support" they insisted that LS-30 is perfectly
capable of scanning Kodachrome. After a three month run-around this
"3rd-level support" produced a "solution" vastly inferiour to my
amateur bunglings.
>You consider such explanations to be "technical minutiae", yet you still
>ramble on about "subsequent scanners", "nonlinearity", "bit-depth",
>"optimum settings", "magic LEDs" etc. which all indicate that you have
>not understood any of the issues involved. I suggest that you read over
>the thread again in slow time, in particular those areas concerning
>spectral bands of emulsion dyes and LED emission spectra and also search
>out further explanations of metamerism and its causes.
>
>When you eventually do understand the issues, and it is not difficult so
>I am confident that you will, it should then be obvious why a
>manufacturer endorsed setting for your scanner, without hardware
>modification, cannot be any less variable than the settings you have
>already derived yourself - but which you describe as "hit and miss". You
>*might* then understand why such a manufacturer would not be prepared to
>endorse such a "hit and miss" setting.
The reason my results were hit and miss have nothing to do with the
hardware but with the lack of tools. I would get the same results on
an LS-9000 if it had *no* Kodachrome setting just like LS-30 has no
Kodachrome setting.
>Sadly, however, I suspect that you will continue to stamp your feet and
>blame Nikon for not giving you what you demand when in fact they have
>upgraded their scanner range *including the hardware* to provide exactly
>what you demand - albeit before you even realised what you were going to
>demand!
Well, that's one way of avoiding the simple question of why there is a
*single* Kodachrome setting on scanners Nikon arbitrarily chose to
support.
Nevertheless, and in spite of your often condescending and
confrontational tone, I again thank you for all your input. I really
mean this.
Don.
Kennedy McEwen
<phoney...@yahoo.com> writes
>
>The reason my results were hit and miss have nothing to do with the
>hardware but with the lack of tools.
Why would adequate tools prevent any setting that you derive from
changing from one KC film to another? Tools will only let you get a
more accurate setting for one particular batch. You have described your
results as "hit and miss", so having "hit" once, you should then achieve
the same level of consistency between slides that ANY setting, no matter
how accurately derived, will achieve. You appear to be suggesting that
your "amateur bungling" has never been close to a hit in the first
place!
>
>>Sadly, however, I suspect that you will continue to stamp your feet and
>>blame Nikon for not giving you what you demand when in fact they have
>>upgraded their scanner range *including the hardware* to provide exactly
>>what you demand - albeit before you even realised what you were going to
>>demand!
>
>Well, that's one way of avoiding the simple question of why there is a
>*single* Kodachrome setting on scanners Nikon arbitrarily chose to
>support.
>
I have NOT avoided the question. It is quite simple and I have told you
many times - including in that last post: different spectral
characteristics are required for the illumination source to minimise
metamerism. Unfortunately, you demean that as "magic LEDs" - thus
demonstrating the truth in that first law of science fiction.
Don
<r...@nospam.demon.co.uk> wrote:
>>The reason my results were hit and miss have nothing to do with the
>>hardware but with the lack of tools.
>
>Why would adequate tools prevent any setting that you derive from
>changing from one KC film to another? Tools will only let you get a
>more accurate setting for one particular batch.
Unless, of course, they provide an average, general setting to address
all variants equally well (or equally badly, if you wish). And before
you jump all over that *please consider* what does the *single*
Kodachrome setting do on scanners where it is available?
How does that *single* setting deal with all different KC variants?
Even if those scanners "may" (as you put it) have a superior light
source, the KC variations are still there. That can only lead to the
conclusion that a single KC setting is possible, no matter how
imperfect.
And that's the key! I'm *not* asking for a *perfect* setting. No
general setting can ever be perfect and that's not at discussion.
>>>Sadly, however, I suspect that you will continue to stamp your feet and
>>>blame Nikon for not giving you what you demand when in fact they have
>>>upgraded their scanner range *including the hardware* to provide exactly
>>>what you demand - albeit before you even realised what you were going to
>>>demand!
>>
>>Well, that's one way of avoiding the simple question of why there is a
>>*single* Kodachrome setting on scanners Nikon arbitrarily chose to
>>support.
>>
>I have NOT avoided the question. It is quite simple and I have told you
>many times - including in that last post: different spectral
>characteristics are required for the illumination source to minimise
>metamerism.
So, are you saying that such a light source is *not* susceptible to KC
variations?
This is the contradiction I find difficult to accept because it's just
so logically inconsistent.
The general settings for different light sources may be qualitatively
different when compared to *each other*, but within their own domain
(i.e. for a particular light source) they are the best *overall*
(general) setting for that light source. That's all I'm saying.
I also think you misconstrued my demanding my rights as petulance.
It's a cultural difference between Europe and North America, and I
don't think I will offend anyone (apologies if I inadvertently do) but
European consumers are far less assertive than North American ones.
I know whereof I speak as I have lived in many European countries for
several years each (Germany, Britain, Holland and Belgium, to name a
few).
There is a definitive difference between customer service and consumer
rights on the two continents. Europeans are better protected at
government levels through various bodies (which is perhaps why they
are more accepting of authority), while North Americans have better
personal rights and are more assertive as individuals (perhaps to
counteract the comparatively lesser amount of bureaucratic
protection). You seem to misinterpret this assertiveness.
And since you have a tendency to twist things into extreme positions,
let me stress that this is all a question of degree. It's not black or
white.
Don
Kennedy McEwen
<phoney...@yahoo.com> writes
>On Tue, 6 Jan 2004 02:20:24 +0000, Kennedy McEwen
><r...@nospam.demon.co.uk> wrote:
>
>>>The reason my results were hit and miss have nothing to do with the
>>>hardware but with the lack of tools.
>>
>>Why would adequate tools prevent any setting that you derive from
>>changing from one KC film to another? Tools will only let you get a
>>more accurate setting for one particular batch.
>
>Unless, of course, they provide an average, general setting to address
>all variants equally well (or equally badly, if you wish).
No - it will still result in EXACTLY the same variation. Only the
centre point of the distribution will be different, which might be good
or bad - the variation will be EXACTLY the same. You have achieved
variations that you term "hit and miss" with your own empirically
derived setting. Any other single setting will achieve EXACTLY the same
variation.
> And before
>you jump all over that *please consider* what does the *single*
>Kodachrome setting do on scanners where it is available?
>
It provides a balance from the spectrally integrated product of dye
density and LED emission which is reasonably consistent from batch to
batch. That cannot be achieved without matching the LED emission to the
emulsion's spectral characteristics.
>How does that *single* setting deal with all different KC variants?
We've been through this several times - you either can't or won't
understand, its difficult to tell which since it has been spelled out in
very simple terms. One last time though - and I mean it this time!
For simplicity of numbers, lets take hypothetical nice round number
wavelengths. Consider one layer of the emulsion which, lets say,
contains a dye which has a spectral range from, say 500nm to 700nm, with
a dye density which increases linearly from the extreme wavelengths to
peak in the centre at 600nm. In the scanner, the density of this
particular layer is sensed by measuring the transmission of one LED
which has a narrow band of wavelengths centred at 600nm, the peak
density of the dye itself. Everything works fine, as the dye density
increases for different parts of the image, the scanner senses less LED
light transmitted through the dye and produces an output signal.
Now, lets, say that this dye varies throughout the batch of emulsions at
its extreme wavelength by up to +/-20nm. In short, the overall
bandwidth of the dye is variable by up to +/-20%. This which would
result in a change in the visible density of the dye of some magnitude,
because the visible density is the sum of the products of the dye
transmission and the eye response at every wavelength in the spectrum.
So, to correct this variable dye bandwidth, the development of the film
is tweaked, so that more/less colour coupling occurs in the process and
the visible density of that dye layer remains constant under a specified
wideband source of light. In short, to the eye, there is NO difference
to the apparent density between batches of the emulsion - and no
difference to any sensor which measures the density over a large
bandwidth, such as a digital camera with gelatin filters (also dyes) on
the appropriate pixels.
However, to the scanner measuring transmission of ONLY the LED
wavelengths, the variation in bandwidth of the dye is not detected - so
instead of consistent colour balance, the scanner sees the tweaks in the
total amount of dye that have been applied to correct for its loss in
bandwidth. Consequently, a variation of colour reproduction of that dye
between batches is produced by the scanner - which is invisible, or
almost so, to the eye or any wideband sensor like a trilinear CCD
scanner or a digicam. This is a manifestation of metamerism - the
signal integrated over all wavelengths has a different variation
depending on the light source, the response of the sensor and the
spectral characteristic of the dye. It is why some clothes look pretty
much the same colour under daylight as they do under the fluorescent
lights of the shop you bought them in whilst others look completely
different - spectral characteristics of the dye, light source and your
eye.
What you need is some means of measuring the amount of the dye present
in a part of its spectrum which is susceptible to *both* the variations
in bandwidth and the density, so that both can balance out in
approximately the same way that the process tweaking was originally
intended to. In our example, two ways of achieving this is simply to
move the LED wavelength away from the dye peak wavelength or increase
its bandwidth to cover a larger part of the dye spectrum. That way the
scanner senses a better balance of both the dye spectral variation and
the amount of dye used to compensate for it - ideally reaching the same
balance that the eye does.
Now, real dyes don't have spectral densities which increase linearly
towards the peak. The have a range of incremental densities, plateaus,
peaks, troughs and decrements between their upper and lower wavelength
range. Some of these areas change in density from batch to batch,
others remain consistent. Now, you might get the best signal to noise
measuring the transmission of the LED at the spectral peak of the dye
but, as seen in the example above, that makes the instrument much more
susceptible to the variations in the amount of dye used to compensate
for the integrated spectral density. If you move the LED to a
wavelength which does actually change its density between batches then
you achieve a much more consistent result, but also require a more
sensitive detector and ADC. Similarly, if you use a wider waveband LED
then you achieve the same thing. It's unlikely that you can compensate
for the variation completely just by changing the LEDs, but you can
certainly reduce the effect significantly.
Even in the single KC setting for those modern Nikon scanners which
support it, there is still a residual variation from batches. However
this is MUCH less than for scanners which were optimised for the
spectral characteristics of the E6 dyes.
>
>So, are you saying that such a light source is *not* susceptible to KC
>variations?
>
>This is the contradiction I find difficult to accept because it's just
>so logically inconsistent.
>
It is only inconsistent if you neglect the spectral issues, as you seem
determined to!
Don
<r...@nospam.demon.co.uk> wrote:
>> And before
>>you jump all over that *please consider* what does the *single*
>>Kodachrome setting do on scanners where it is available?
>>
>It provides a balance from the spectrally integrated product of dye
>density and LED emission which is reasonably consistent from batch to
>batch.
Exactly! And that balance is, precisely, what I have been talking
about all along!
I was well aware before this thread even started that, as such (a
balance), it will not match all cases perfectly. That was a given. But
it's a much better starting point than no setting at all.
>Even in the single KC setting for those modern Nikon scanners which
>support it, there is still a residual variation from batches. However
>this is MUCH less than for scanners which were optimised for the
>spectral characteristics of the E6 dyes.
Right! And, as I also said, that means a balance for scanners not
optimized would be less accurate. But I, both, accepted and expected
that too.
Changing the subject slightly, a few questions. Regarding the formula
for saturation you mentioned before:
S= (Max(R,G,B) - Min(R,G,B)) / Max(R,G,B)
First of all, I seem to have read somewhere that conversion between
RGB and HSI (might have been LAB, though!) is lossless. Is that
correct? By this I mean, for example, loading an image into Photoshop
and then switching between RGB and HSI (or as Photoshop calls it "HSB"
- that is the same as HSI, I take it?).
Actually, I should clear up the term "saturation". In this context, to
me, it means how "clear" the color is (yes, I realize how silly that
"definition" is). Is that correct? If so, where does the intensity fit
into this? (Photoshop defines saturation as "the strength or purity of
the color" and it "represents the amount of gray in proportion to the
hue" while B (brightness, which I take as Photoshop's synonym for
Intensity) as "relative lightness or darkness of the color".)
Next, you write:
Thus with R=8, G=10, B=20, S=(20 - 8)/20 = 0.6
Increasing the exposure on all three channels equally, by two EV,
will result in values around R=32, G=40, B=80, and again S=0.6
How did you arrive at these values? They are *quadruples* of your
example value in order to achieve *doubling* of the EV value. I'm
guessing that, like sound, it's probably a logarithmic scale. However,
just be sure, what is the relationship between the two?
And finally, what are the other two formulas for expressing intensity
and hue as a function of RGB?
In other words (in addition to understanding all this better) what I'm
getting at is this: How do I use analog gain (which according to the
scanner docs is equivalent to EV) to "brighten up" a scan without
changing the color balance? As you wrote, and I myself observed first
hand, simply raising master RGB analog gain alters the overall color
balance. (The reason I want to use analog gain is to do this at the
earliest stage of the process.)
Thanks!
Don.
Kennedy McEwen
<phoney...@yahoo.com> writes
>On Thu, 8 Jan 2004 00:26:58 +0000, Kennedy McEwen
><r...@nospam.demon.co.uk> wrote:
>
>>> And before
>>>you jump all over that *please consider* what does the *single*
>>>Kodachrome setting do on scanners where it is available?
>>>
>>It provides a balance from the spectrally integrated product of dye
>>density and LED emission which is reasonably consistent from batch to
>>batch.
>
>Exactly! And that balance is, precisely, what I have been talking
>about all along!
>
>I was well aware before this thread even started that, as such (a
>balance), it will not match all cases perfectly. That was a given. But
>it's a much better starting point than no setting at all.
>
of your "hit"s! Nikon cannot provide anything better than the same
degree of "hit and miss", no matter how they derive the setting for the
LS-30 and publishing such a setting would open them to far greater
support costs than simply not publishing one at all!
>
>Changing the subject slightly, a few questions. Regarding the formula
>for saturation you mentioned before:
>
> S= (Max(R,G,B) - Min(R,G,B)) / Max(R,G,B)
>
>First of all, I seem to have read somewhere that conversion between
>RGB and HSI (might have been LAB, though!) is lossless. Is that
>correct?
No, although unlike lossy compression such as JPEG, the losses are
constrained to the initial conversion in each direction. This is fairly
obvious to see when you realise where the losses come in. In RGB, using
8bits per channel, there are (2^8)^3 individual colours, which is 2^24,
or 16,777,216 individual colours. In HSI, the H (hue) & S (saturation)
settings are irrelevant if the I (intensity) is zero. Similarly, for
values of S which are zero, the H value is irrelevant. Similarly, for
low values of I, the discrete steps in H & S do not map onto unique
values of RGB, and the same for low values of S. Consequently, fewer
colours can be represented in 8bpc HSI than in RGB, so the conversion
from RGB to HSI must involve some loss of information which cannot be
recovered. The same is also true, in the reverse direction. In short,
conversion between RGB and HSI cannot be lossless unless the bit depth
in one of those colour spaces is significantly higher than in the other.
>Actually, I should clear up the term "saturation". In this context, to
>me, it means how "clear" the color is (yes, I realize how silly that
>"definition" is). Is that correct?
No, saturation defines how "pure" the colour is, or how "uncontaminated"
by other colours it is, or how much the colour deviates from a neutral
grey. This is easier to understand in terms of the primary colours in
RGB or the secondary colours in CMY, but applies equally to any mixture.
100% saturated red, for example, contains no green and no blue. The
amount of red can change from 0 to 255 (assuming 8bpc) but, provided
there are no other colours present, then it is still 100% saturated. A
conversion to HSI would thus give a saturation value of 255. The
intensity, I, simply represents the amount of the colour present. Adding
equal amounts of green and blue to the original amount of red reduces
its saturation, because the colour is now a mixture of grey and red, so
the difference between the colour and neutral grey is reduced. Adding
unequal amounts of green and blue changes both saturation and hue, since
the colour now represents an amount of neutral grey together with a
particular colour which is a mix of red and *either* green or blue.
You can consider RGB colour space as defined by a cube, with axes of R,
G & B. Any colour within that cube is represented by its value of R, G,
& B. HSI colour space is defined as a cylinder. The central axis of
the cylinder represents intensity, the distance of the colour from that
axis represents saturation and the angle between the shortest line from
the axis to the colour and a defined reference (nominally pure red)
represents hue. Thus the conversion between RGB and HSI is essentially
one of cartesian to cylindrical co-ordinates. Whilst both describe
space equally well, pacing constraints on the maximum value and the
precision of any of the axes in either description, such as using 8bpc,
inevitably results in both describing different colour spaces with
different degrees of accuracy - a cube is not a cylinder and vice versa.
Consequently, some of the cube falls outside of the space which can be
described by a cylinder, if the largest cylinder is defined to reach the
edges of the cube, or some of the cylinder falls outside the space which
can be described by a cube, if the largest cylinder is defined to reach
the corners of the cube. Clearly, if some space remains inaccessible in
one or other, or even both, of the descriptions then it follows that
some of the positions within the space which are described uniquely by
one description will not be defined with sufficient precision to be
unique in the other description.
>If so, where does the intensity fit
>into this? (Photoshop defines saturation as "the strength or purity of
>the color" and it "represents the amount of gray in proportion to the
>hue" while B (brightness, which I take as Photoshop's synonym for
>Intensity) as "relative lightness or darkness of the color".)
>
I think you should find this description compatible with the first
paragraph above.
>Next, you write:
>
> Thus with R=8, G=10, B=20, S=(20 - 8)/20 = 0.6
>
> Increasing the exposure on all three channels equally, by two EV,
> will result in values around R=32, G=40, B=80, and again S=0.6
>
>How did you arrive at these values? They are *quadruples* of your
>example value in order to achieve *doubling* of the EV value. I'm
>guessing that, like sound, it's probably a logarithmic scale. However,
>just be sure, what is the relationship between the two?
>
EV is "Exposure Value" and derives from a photographic term used to
define the correct exposure for film for a given light level. The light
level corresponding to EV 0 is defined as an exposure of 1 second, with
a lens of f/1 using ISO100/21 film. With twice as much light, the EV
increases by one, and the correct exposure can be obtained with a film
speed of ISO50/18, a shutter speed of 1/2sec or a lens aperture of f/1.4
Consequently a change in light level of x4 is 2EV, x8 3EV, x18, 4EV etc.
The first page that comes up on a Google of "Exposure Value" is
http://www.chem.helsinki.fi/~toomas/photo/ev.html
which describes the relationship adequately, but there are many.
>And finally, what are the other two formulas for expressing intensity
>and hue as a function of RGB?
>
I = Max(RGB) + Min(RGB)
Since H defines an angle, the conversion formula is more complex than
the others.
If I=0 then H=0
If I>0 then H=acos(2xR-G-B)/(2*sqrt((R-G)^2+(R-B)*(G-B))/(2*pi)
if B>G then H=1-H
The results must be scaled appropriately for use.
>
>In other words (in addition to understanding all this better) what I'm
>getting at is this: How do I use analog gain (which according to the
>scanner docs is equivalent to EV) to "brighten up" a scan without
>changing the color balance? As you wrote, and I myself observed first
>hand, simply raising master RGB analog gain alters the overall color
>balance.
That's not what I wrote - raising master RGB analogue gain will not
change the colour balance or the saturation significantly. All it can
do is increase the precision that the hue of any pixel is determined to
by increasing the utilised range of the ADC. That can result in a
*minor* change to both hue and saturation but only as a consequence of
data precision and rounding errors. As you can see from the equation
above, hue is normalised so that any scaled increase in RGB results in
no change to the hue (or saturation, as previously explained), only the
intensity. In the previous example, if you put the values of R, G & B
into the equation for H then you get a result of 0.64 irrespective of
the analogue gain.
All of this assumes, of course, that the dark current of the CCD and any
DC offset introduced between the CCD output and the ADC input has been
equally and accurately calibrated beforehand. If residual DC offsets
exist in any of the colour channels then adjustments of the master RGB
channels will NOT result in proportional increases in RGB and "all bets
are off" as to the hue, saturation or even intensity of the final result
- but there is little you can do about it anyway.