P
Philip Tobias
UrbanVoyeur said:
Yes. The phrase "get a life" comes to mind, doesn't it?
....pt
Yes. The phrase "get a life" comes to mind, doesn't it?
....pt
B
Bart van der Wolf
Ralf R. Radermacher said:
Trying to scan at an exposure factor of 100 (or more), assumes the
need to penetrate a density of over D=5.00. Since color accuracy of
most slide films is seriously compromised at densities over 3.00, it
seems a decent hardcoded limit to avoid damaging the sensor or
blooming. It also demonstrates that the programmer considered
exceptional user choices, and prevented them.
Bart
D
Daniel
Haven't you got anything better to do??
D
Don
That's just plain silly... Trying to defend the indefensible i.e.,
another sloppy VueScan bug.
If it were a conscious decision by the (so-called) programmer to block
this exposure, then an alert box would come up. It does *not*!
Instead, VueScan *secretly* (!) performs a *random* exposure which has
no relation to the setting! Again: a *random* exposure! That's a BUG!
As to how useful such high exposures are: Ever heard of contrast
masking or high dynamic range (rhetorical question)? Not only are such
exposures useful, they are *necessary*! And I didn't even mention
"Kodachrome"! ;o)
Don.
U
UrbanVoyeur
Don said:
What are you talking about? Plenty of applications from Microsoft, Abode
and others "roll over" or "roll back" their spinners when you reach the
limit of the range. Its a very common way of limiting the input. Its not
a bug, it just is.
And what is the so-called random exposure? I have only had consistent,
accurate results from vuescan.
I've scanned a number of Kodachromes (200-300 at least), old and recent
with Vuescan and never had a problem. Color were reasonably accurate,
grain was fine, contrast was good and focus was sharp.
What exactly was your problem when you tried to scan Kodachromes?
D
David J. Littleboy
UrbanVoyeur said:
He tried to scan his Kodachrome with a scanner known not to be capable of
scanning Kodachrome. Kennedy spent a lot of time explaining why it wasn't
possible with that scanner, but Don never got anywhere near understanding
what was going on. (LED scanners measure points in the spectrum, whereas our
eyes average over a range of frequencies (when the slide is illuminated with
white light). So what the eye sees can be very different from what an LED
scanner sees with Kodachrome, whose dyes have weird spectral responses.)
David J. Littleboy
Tokyo, Japan
U
UrbanVoyeur
David said:
OIC. Not much you can do about that.
D
Don
That's just patently wrong!
Every scanner can scan Kodachromes but it may require more work if the
manufacturer does not provide a workable, pre-chewed solution.
When correct individual Analog Gain is applied to compensate for both
LED response *and* Kodachrome response curve there's no problem.
You seem to have selective memory.
The context was: Why doesn't Nikon retrofit Kodachrome mode (i.e.
enable it in NikonScan) for older scanner models.
What Kennedy said is that due to the many variables (LED, film
response, etc) they *may* not be a *single* and *literal* correction
value. I maintained that it was a marketing decision. Do note that
these two statements are not in conflict!
Kennedy specifically did *not* say that an adaptive, empirically-based
solution was impossible. Indeed, that's exactly what I did in the end.
Next time, may I suggest you check the archives first...
Don.
D
Don
Yes, but they don't go on and perform the task with an unknown value!
They roll back the display (and probably Beep to alert you to it) and
do *not* perform the task until you acknowledge the change.
VueScan performs the scan *first* and only then is the exposure
display secretly (!!) changed. Unless you go back and check it you
wouldn't even know! So what exposure was used? Nobody knows.
It's simply extremely sloppy programming, so typical of VueScan.
BTW, just because some other sloppy programs may fail to do proper
boundary checking is totally beside the point. The excuse *but mom, he
did it too...* didn't work when you were 5 and it doesn't work now...
In a word, it's yet another VueScan bug.
When the actual exposure used is unknown and doesn't correspond to the
exposure specified beforehand. Especially, when the display is
*secretly* modified after the scan!
Great! Enjoy!
The same problems everyone has when scanning them on a Nikon scanner.
Blue cast (actually absence of red), dark scans, etc. Indeed, it's a
particularly dark Kodachrome which revealed that particular VueScan
bug.
Don.
B
Ben Fullerton
: >I have only had consistent,
: >accurate results from vuescan.
: >What exactly was your problem when you tried to scan Kodachromes?
: The same problems everyone has when scanning them on a Nikon scanner.
: Blue cast (actually absence of red), dark scans, etc. Indeed, it's a
: particularly dark Kodachrome which revealed that particular VueScan
: bug.
I have yet to try Vuescan on either of my scanners (Nikon CoolScan IV and
Microtek ScanMaker E6) and I am looking for suggestions as to which of the
many versions in the development history of Vuescan is felt to be the
easiest to use and the most completely de-bugged.
I am NOT a pro, and my main objectives are to digitize a few thousand
Kodachromes and fewer thousands of negatives - some with adequate quality
for family archives and probably a few with the best quality I can get for
sections (eg. one person from a group shot) cropped from slides and 35mm
negatives.
Many will be used only for computer or TV viewing but some will be printed
in sizes of up to 8 x 12 in.
While I am here, any comments on what version of NikonScan might give the
best CoolScan IV-ED results for use with W2000Pro or W98SE? (I use both)
I would like to spread out the extended 'learning curve' involved in,
eventually, approaching the level of expertise of many of the contributors
to this newsgroup - and get some usable, if not top quality, results along
the way.
As for those who think that Vuescan is crappy, why not just dump it and
ignore it - as I have done with one of the 'bundled graphics editors'
which was included with one of my earlier digital cameras?
I expect that there are many who use and like them. That is ok with me.
Live and let live works for many of the things that I do not agree with.
Ben Fullerton
: >accurate results from vuescan.
: >What exactly was your problem when you tried to scan Kodachromes?
: The same problems everyone has when scanning them on a Nikon scanner.
: Blue cast (actually absence of red), dark scans, etc. Indeed, it's a
: particularly dark Kodachrome which revealed that particular VueScan
: bug.
I have yet to try Vuescan on either of my scanners (Nikon CoolScan IV and
Microtek ScanMaker E6) and I am looking for suggestions as to which of the
many versions in the development history of Vuescan is felt to be the
easiest to use and the most completely de-bugged.
I am NOT a pro, and my main objectives are to digitize a few thousand
Kodachromes and fewer thousands of negatives - some with adequate quality
for family archives and probably a few with the best quality I can get for
sections (eg. one person from a group shot) cropped from slides and 35mm
negatives.
Many will be used only for computer or TV viewing but some will be printed
in sizes of up to 8 x 12 in.
While I am here, any comments on what version of NikonScan might give the
best CoolScan IV-ED results for use with W2000Pro or W98SE? (I use both)
I would like to spread out the extended 'learning curve' involved in,
eventually, approaching the level of expertise of many of the contributors
to this newsgroup - and get some usable, if not top quality, results along
the way.
As for those who think that Vuescan is crappy, why not just dump it and
ignore it - as I have done with one of the 'bundled graphics editors'
which was included with one of my earlier digital cameras?
I expect that there are many who use and like them. That is ok with me.
Live and let live works for many of the things that I do not agree with.
Ben Fullerton
U
UrbanVoyeur
Don said:
You say potato...
I'm not so sure about that.
I used a Nikon LS-4000 with my Kodachrome slides - 25, 64 and 200. I had
no problems with color cast or exposure on Vuescan. If anything the reds
were a bit over the top, but that is a characteristic of Kodachrome. It
would have helped to have a calibration slide, as I have for ektachrome
and fujichrome, but by the time I found a source, I was done.
Initially I was using a Minolta Multi, but switched to the Nikon. Same
good results from Vuesan on either one, but the 4000 produce higher res,
& batch fed scans.
I would note that 75% of my K-14 slides were the correct exposure and
the rest within +/- 1/2-3/4 of the correct exposure. They may have a lot
to do with the results. Kodachrome is know go "wierd" when undexposed.
It is not at all like E-6. Some IlfoChrome/CibaChrome printers won't
touch dark Kodachrome for that reason - you can't get a decent print.
It also stands to reason that a dark slide will produce a dark scan.
Kodachrome is a high contrast film and goes to black very quickly.
Bullet proof black, in fact. The blue cast may have just been amplified
sensor noise or LED spill over in the dark areas.
U
UrbanVoyeur
Don said:
So, unless I missed something, what you're saying, essentially, is that
you had trouble scanning Kodachromes, and that it is a problem with the
Nikon scanner?
Where exactly is Vuescan at fault here?
BTW, it *is* possible to get excellent scans with Kodachrome on an LS-4000.
R
Robert Feinman
I find I get the best results from locking the settings in Vuescan
and making all adjustments in Photoshop afterwards. You need to lock
the settings so that the full range of densities is captured.
It also helps if you can profile your scanner or get a profile from
the manufacturer.
Then in Photoshop you assign the profile when opening and convert to
AdobeRGB for editing.
This way the scanner gets to do its job without any "automatic"
guesses by the software.
The latest version of Vuescan works fine. Download and try it
for yourself.
N
Neil Gould
For those scanners, I think you'd be fine to get the current version ofRecently said:
Viewscan. As for bugs, I think that one participant in this group has one
up their posterior, so I wouldn't worry about it. Viewscan has worked as
advertised with my scanners, and I like its interface better than some
very popular and expensive scanning software.
Neil
D
Don
Yes and no. There are some limitations of the scanner due to the light
source (which apply to *all* Nikon scanners, BTW) but these can be
easily overcome by applying correct amounts of Analog Gain (AG).
The problem was Nikon failed to retrofit the "Kodachrome" (KC) mode to
the LS-30 (the scanner in question). This is purely a software
retrofit i.e. an option in NikonScan (NS). It was probably due to
marketing reasons and/or laziness. Each connection type in NS has a
different module. And Nikon just didn't want to go back and fix the
"obsolete" SCSI module.
Instead, the so-called Nikon "support" blamed everything from phases
of the moon to the day of the week... I went along with all of their
fancy "explanations" disproving them all in turn - even though it was
blatantly evident from the start where the problem was.
I even sent in a slide on their request knowing darn well it was a
monumental waste of time! The whole thing took a couple of months. In
the end, Nikon admitted what I (and I'm sure they) knew all along.
That it was their failure to retrofit the KC mode.
My point exactly! ;o)
Some VueScan "supporters" do - as a matter of course - often try to
change the subject when it becomes uncomfortable... ;o)
Indeed! Proving that what David said was patently wrong. I came up
with approximate AG settings myself for the LS-30 but I wanted
"proper" settings from Nikon - only to get a royal run-around and then
a grudging confirmation of what I was saying all along.
As I wrote, after I got the LS-50 which does have the KC mode and
compared the scans with the KC mode on and off, it generally confirmed
my initial LS-30 settings.
Don.
D
Don
Well, we are digressing a bit. The Kodachrome reference in my reply to
Bart related to the exposure (length). However, David "understood" is
as a reference to the cast problem.
Having said that, though, KC (actually Nikon's uncooperative attitude)
and massively incompetent so-called "support" (there were several
other instances) is what caused these scratch marks on my walls... ;o)
So, this is always an interesting digression.
I tried getting a KC reference slide both in North America and in
Europe but these were just impossible to find. On reflection, I wonder
how useful they would be in the end because there isn't a single KC
but many different versions.
Well, you have to quantify "good". VueScan - due to its demonstrated
unreliability and bugs - is bound to corrupt the data one way or
another. I mean, all emotions aside, if we are objective that really
is a given.
Having said that, though, it all comes down to one's requirements. One
may not see or even be aware or simply don't care about any of these
many VueScan shortcomings. It all depends on one's needs and testing
procedures (how thorough or detailed they are, or even if they exist).
All that may not be a major factor and I have even advised some people
to try VueScan - given they are aware of the shortcomings.
So, "works for me" is not really an objective factual evaluation.
YES!!! That's the key problem. Actually, technically, they aren't
underexposed. I mean, they are perfect when used as intended i.e. to
be projected. That's why people loved them!
The problem is the migration into the digital domain. All film
scanners have trouble with KCs to some extent but Nikons are
particularly "touchy".
No, it's the characteristic Kodachrome curve:
http://www.kodak.com/global/en/professional/support/techPubs/e55/f002_0486ac.gif
What's really frustrating is that Nikon could have been much more
helpful. In the end, however, even the KC mode doesn't go far enough
because it's really only a compromise solution. It also breaks down
rapidly as you increase the exposure because their algorithm appears
aimed at nominal exposure ("absolute 0") and is not adaptive.
NikonScan makes this more difficult by hiding absolute exposure
because AG adjustments are *on top* of this "secret" baseline
exposure. AutoExposure can be turned off but it's very difficult to do
(must turn off both the scanner and NikonScan) and then there is only
a limited, cumulative range of only +4 AG.
In this respect - at first blush - VueScan appeared to have the upper
hand: no relative baseline but always absolute exposure display, no
upper limit (well, later I learned otherwise...). Which is why I tried
it in the first place.
But then the problems started... First of all, VueScan uses this inane
"multiplier" instead of the standard ev setting. Then there was the
exposure bug (which started this sub-thread). Finally, (back then) VS
didn't even have individual RGB AG setting!!
Indeed, it was my challenge to the author which resulted in individual
RGB AG being added to VueScan (after months of futile begging by meek
VueScan groupies)! And the thanks I get from (some) VueScan "users" is
vicious hate messages! ;o)
Don.
D
Don
Asking for a particular version is academic, because only the latest
version is available on the site. So, unless, someone volunteers to
send you their favorite version, there's no other way of getting it.
But before writing anything else (and this is the key) try it out
yourself! And then come back with specific questions. There are many
dedicated VS victims... erm... users ;o) more than willing to help.
I take it, you guessed by now what I think of it... ;o)
Re de-bugged: No such thing! (Never mind "completely"!)
Re easiest to use: See above... ;o)
Seriously though, VS has a very "unique" user interface. One of the
most common problems (and indeed bugs) is the unpredictable
interaction of seemingly unrelated options on different tabs and
sometimes not even displayed. The common "cure" for this is to delete
the INI file and start from scratch.
Again, all that is moot: Try it out yourself!
Whether pro or amateur it really depends on how much quality you're
after.
With such a low threshold, more subtle VueScan bugs will *not* affect
you and VueScan may indeed be suitable (assuming you like the user
interface and the results).
The only thing you have to worry about are more blatant bugs like
"can't scan at all" or "cropping doesn't work", etc.
Once you find a version without those bugs, you should be OK. After
that do resist the urge to upgrade!! And if you feel you must, make
sure you keep the old version backed up!!!
Use the latest version which supports your scanner, which (I believe)
is 3 on W98 and 4 on W2K.
That's going to be tricky. Be prepared for a lot of frustration along
the way. The only thing that will get you there is time.
Even though it may look boring, but start by reviewing the archives.
Many of your future pitfalls are "been there, done that" and you'll
find plenty of useful messages there. However, humans usually learn
best from their own mistakes.
So, stick around and ask questions. There are plenty of knowledgeable
people here who are more than willing to help!
However, if you do want to get to that level, the more you learn the
more likely it is that you'll find VueScan "problems" bothersome. But,
again, since we learn best from our own experiences wrestling with
VueScan's may end up being a good learning exercise.
We do, but some VueScan users are not happy with that and insist on
trying to start a flame war. Ignoring them, apparently, just makes
them angrier... ;o)
Don.
U
UrbanVoyeur
Don said:
I think there was a misunderstanding on my part here. I thought that you
were taking the position that the problems you experienced with
Kodachrome were related to Vuescan, when you clearly state that it was a
Nikon failing. A apologize for that.
Just curious: Why the LS-50 and not the LS-5000, which captures more
bits and has the batch option?
U
UrbanVoyeur
Don said:
Yes, and it is very difficult to get calibration slides for older
emulsions. I wish I could remember where I found the Kodachrome
calibration slides. I do recall it was about $200-300 per slide from one
place and more from another - as opposed to Wolf Faust's $50-60 E-6
calibration slides.
Good = consistent, predictable, repeatable, low noise, absence of
artifacts, sharply focused, reasonably accurate color (given the lack of
a calibration slide), well saturated, highlights and shadows that hold
detail.
I could get good results from either the Minolta or the Nikon with
Vuescan, but the nikon had the edge in contrast range, res, and automation.
VueScan - due to its demonstrated
I don't find that, and i don't accept it as a given. If I did, I
wouldn't use the product. I hold no allegiance to any image processing
software. If *any* product doesn't work, I stop using it and find one
that does. I spend too much time and effort shooting the images to
waste time with software product that doesn't do what I need.
The unprocessed scans (no software adjustments) from Vuescan and
Silverfast on the Nikon are identical. When I start to use the
correction and adjustment features of each the output varies, but that
is to be expected - they are different programs using different
algorithms. But I would not say either "corrupts" the image. And since
the the raw/unprocessed scans are identical, I'd say that qualifies as
objective reality to me.
.I suppose I could crank up Nikon Scan and get a third opinion, but the
program was so buggy on XP SP2 last time I tried it, that I really don't
want to install it again.
I guess I fall into the "can't find a problem" camp. Therefore its a
non-issue for me.
Well, not being a professional reviewer and having experience with only
a limited number of film scanners, "works for me" is all I can attest
to. 5 scanners, 3 photographers, 1000's of slides. No problems. Works as
well or better than Silverfast.
Unless we do a statistically valid sample, none of this is "objective".
J
John
UrbanVoyeur said:
Well said!