Best scanning manager program?

[HvdV]

Yes, but that was only the case when each device used a different
custom interface, usually something based on the bidirectional
parallel port. All scanners these days communicate through the
standard interface usually USB, sometimes FireWire. So that low level
communications layer is abstracted and applications don't have to
worry about that. That takes care of the most difficult part, the
"how".

What's left is the "what" i.e. the command set unique to each device,
actually it's only a subset, the device specific codes. And that's
elementary to implement.

It would surprise me if it is really as easy as you say. As I wrote earlier,
hardware manufacturers, and I'm thinking in particular of the optics field,
tend to be protective about their products, if only by employing incompetent
programmers on their side.

Yes, but that's not the case here. Besides, as I say the interface is
well know (i.e. USB). What's left is the manufacturer specific codes.

Even if Minolta refused to supply this information because of
Vuescan's abysmal record, any even moderately competent programmer can
figure this out. There are even tools to do that e.g. USBSnoopy.

Reverse engineering can be a time eater...

No, just for me. It would be of no use to most people because its
focus is to scan "raw" i.e. without any bells and whistles. All that I
do later in Photoshop. All I want a scanner program to do is *scan*.
The whole point (together with my "merge" TwinScan program) is to
extend the dynamic range of dense Kodachromes which my LS-50 doesn't
penetrate. So I scan each image twice, once normal exposure and once
overexposed (+4 AG) to get the shadows. Then I align the images,
color/exposure adjust and merge using a *hard* border (no feathering).

Why not take up contact with the KDE or Gnome people so we can have a KScan,
supporting say the LS-50 and the KM-5400-II? BTW, what do you use for
correlation?

How bad is it? I tried SilverFast very briefly but its concept is
completely opposite of what I want to do. SilverFast puts a lot of
effort on user-friendliness and "auto everything". I, on the other
hand and as mentioned above, want to scan as raw as possible. Even
after turning everything off in SilverFast it still did a lot of
processing and I never managed to get a raw scan out if. But I didn't
try very hard, I must add...

It's not really that bad. At first I was running it on an alpha version of
OSX 10.4, and often it did hang on the manual focus tool. With the official
10.4 it works fine. Still, the 'hang' was poorly handled, even after killing
Silverfast I could not restart it easily, might be shared memory segments
lying around, but I didn't investigate.
There are attempts to auto everything, but AFAIK you can control everything
yourself. There is also an interactive scratch remove function for the tough
cases (did crash once).

Yes, but the whole point of the PowerPC Alliance (or whatever they
were called) was to brake Intel's monopoly. So to sabotage it right
from the start because of Apple's greedy self-interest is just mind
boggling. It's self-defeating? Like that story of the scorpion and the
frog...

That might have been IBM's agenda, but what was Apple's interest? At that
time RISC chips were superior to Intel chips, so it was a good idea to
replace the old Moterola 68xxx by one of Alpha, MIPS, Sparc, Power, etc.

Now, if the Power PC won and Apple *then* broke away, I could
understand that, but to break away before they achieved anything is a
case of "sawing the branch you're sitting on"...

Well, IBM definitely achieved something, look at the #1 and #2
in the Top500 http://www.top500.org. Intel/SGI too with #3, but with a processor
intended to break its own monopoly! And failing at it too!

Apple achieved to build nice and profitable hardware *and* lived through a
radical change in the OS, quite an achievement. To get not too far OT, for
image processing the G5 does quite well. Speculating, the break away from IBM
might be necessary to keep up with the Intel IA-32 chips, and especially with
the AMD chips. That the Power architecure is superior for scientific
applications including image processing is not so important if you only want
to run a webbrowser.

just my 2 cts, Hans

 

[Roger S.]

Don, I don't disagree with much in your last post responding to me. I
have no desire to search the archives at this point to read your past
threads looking for examples of bias as I'll just wait for new ones to
appear.

One thing that is causing us grief is sloppy syntax.
I can see why you say I misinterpreted the statement above which used
the word "inferior." The original was ambiguous. It wasn't clear that
you were basically saying that some users has self-admittedly
"inferior", or less demanding requirements. I read it as Don saying
that VS users invariably have absolutely inferior requirements. The
original was ambiguous as I read it.

About the "overreacting" and "emotion" statements, I think you just
aren't concerned with how your readers perceive what you write and that
you might have difficulty understanding how they would feel.

Don wrote:
It is *not* belittling!!! It's an explanation because they provide no
fact!

This "explanation" that our conclusions are invalidated by our emotions
doesn't go over well with me as the reader, but clearly you either
don't care or are incapable of caring how others feel. If belittling
others isn't your intent, it's still the result of your words, so you
might want to be mroe careful.

You claim to have no emotional investment in this issue and to simply
be a reporter of fact. Let's say that the facts as you see it add up
the conclusion that Vuescan is a buggy, unreliable program with basic
errors that inexplicably reappear and an abusive author. Now given
this conclusion, there is more than one way to react. From this
conclusion you could react with indifference- that you don't use VS so
why would you care. You could react with sacrasm, that VS users are
suckers and deserve what they get. But you react with outrage and have
a need to tell everyone the truth about this sham, to expose VS for the
fraud it is, and to bring the truth to those who need to hear it. I
think all of these responses are plausible reactions to the same
conlusion, so wouldn't you say that it's your values, emotions and
personality that led you to this particular reaction? Why do you feel
a need to enlighten the masses and counter the VS apologists?
Going further, I'd also argue that emotions, personality and values led
you to draw this particular conclusion from this set of facts. Now
this isn't to say that your conclusion is unsupported or invalid! More
on that in a minute.

In thinking about my own biases and preconceptions which color the
"facts" about Vuescan here's what I've concluded. For me, Vuescan is a
piece of software which fulfulls my image capture requirement better
than the stock scanner driver, which I find unusable. I wouldn't
recommend VS to a novice as you really need to know how your scanner
works or you may never know when VS isn't working right. It's taken me
many months of reading these newsgroups to get a handle on how VS works
(and how it's supposed to work when it fails!). VS can be a plausible
alternative to other scanner software if you are willing to spend the
time to figure out the interface and how it works and can yield good
results. I come to this forum and photo.net to share what I learned
the hard way after a lot of frustration with VS and a scanner in
general. If people who know something about scanning ask on this forum
if they should give it a try, my response is a qualified yes, as I get
results with it that are much better than the alternative, which was
dealing with TWAIN or not scanning at all and handing over my negs to
the local lab.

Don's answer is almost always an unqualified no, that Vuescan is not up
to any job (Don- I'm sure you can find evidence of this without my
help). Now, I'm not the only one getting decent results from Vuescan.
Look at the scanner bakeoff or other photo forums. If VS were to
change radically one way or the other, my experience with it would make
me change my opinon. My experience over the last 12 months since I
bought the scanner has been that VS bugs have gone done and performance
has improved significantly (calibration, IR cleaning, IT8 support,
interface usability, memory usage, preview quality). If you would like
details I can provide them as each one of these things was a source of
frustration at one time. If it had gone the other way and gotten
worse, I'd be warning others not to make my mistake. I think I'm being
fair based on my first-hand experience.

Don will look at this and say that it's not factual and that
"improvement" is subjective whereas bug reports (binary- bug or no bug,
quantifiable- how many bugs?) are objective. I think that this
subjective and supportable evidence is much more coherent than cobbled
together bug posts from the last few year. While the conclusion I draw
from it is different from Don's, it is also supportable by fact and
hundreds of hours of experience. If someone wants to assess whether or
not VS is a workable program, I leave it up to them to ask the right
questions for their use.

I guess I'm a rabid Vuescan fan after all : )

 

[Roger S.]

Okay, I said I was going to wait for Don's next post, but I was
tempted, and had to look at the beginning of the thread to remember why
I posted here in the first place. It was this statement.

Don wrote about Vuescan:
"However, if you don't care for quality and just want a quick a dirty
web scan it just may do the trick"

This is not the only conclusion one can draw from the "facts" that VS
has and has had bugs. So Vuescan is only okay to use if you want a
"Quick [and] dirty web scan" and don't care about quality? That may be
Don's opinion, but my experience, the experience of other users on this
forum, and the results of the Scanner Bake-off contradict it. Frankly,
I think this hyperbolic statement is unsupported by the evidence Don
has presented to date. Don, how can you "objectively" prove such a
sweeping statement- what criteria would you use?

 

[Father Kodak]

I just joined this group, and just read 40+ messages on this topic.
All I can say is,

1. there appear to be two viable non-vendor choices. Silverfast and
VueScan.
2.Neither one sounds "ideal."
3. This discussion seems to have morphed into a software engineering
food fight. I work in a Silicon Valley enterprise software company
(though in marketing, not engineering), so I can appreciate this
discussion in a way. However, I suspect most readers of this group
don't care any longer.

Can I restart this discussion, for my particular situation. My
questions, for the situation below, are:

1. Between Silverfast and VueScan, which gives me the highest degree
of automation when I need it (for a box of slides or strip of film)?

2. Which gives me the most flexible options for a custom settings, for
a more precise scan, probably one at a time?

3. In real life, under real usage, what is the minimum PC config I
would want to run your recommended software? processor speed, RAM,
disk storage.

4. Assuming that I expect my workflow (I haven't gone digital quite
yet) to be "centered" on Photoshop, not Photoshop Elements, which
version of Silverfast would I want? And why?

5. As a side question, what are the specific issues regarding
Kodachrome and recent model Nikon scanners? Can either of these two
packages compensate for these issue?

6. Can either of these packages utilize the built-in image correction
features such as ICE, etc.

Now, thank you for reading this far. Here is my situation:

I have:

5000+ black and white negs, mostly Plus-X and
Tri-X, home developed in various developers. Maybe a small amount of
black and white film based on C-41.

About 1000 color negatives, various Kodak films, C-22 and C-41
processes. Mostly in 4 or 6 image strips.

At least 8000 Kodachrome 25 and 64 slides. K12 and K14
processes. 2 x 2 mounts. Mostly Kodak, some third party lab in
either cardboard or plastic. Some in uncut rolls.

About 1000 Ektachrome high speed slides, pushed to 1600 and
3200. Mostly cardboard mounts. Some in rolls.

I would like to scan most of these images, and do my sorting/culling
after scanning, particular the negatives. Probably in Photoshop, on a
20" monitor. Beats a loupe.

Being a looong time Nikon user, I will probably opt for the 5000 plus
the slide autoloader, so I can batch-scan overnight. With this volume
to be scanned, I can justify the cost of the scanner plus slide
autoloader. Besides, I will stay with film for at least one more
year, possibly longer, until the Nikon digital camera of my "dreams"
is on the market and at a reasonable price. I expect this to occur
not before 2007, maybe 2008.

So I expect to end up scanning 15 K, maybe up to 20K images My
equipment cost per scan could be well under ten cents, or ten plus
scans for one dollar US.

Having my druthers, I would like a "RAW" format output done at 2000
lines or greater. I expect to be doing digital photography with RAW
format.

(I'm upgrading my PC in anticipation of large files and many of them.
I will probably use an 80/160 GB SCSI tape drive for backup.)

Thank you in advance.

 

[Noons]

1. Between Silverfast and VueScan, which gives me the highest degree
of automation when I need it (for a box of slides or strip of film)?

I reckon they're just about the same...

2. Which gives me the most flexible options for a custom settings, for
a more precise scan, probably one at a time?

Vuescan has more settings, but it takes a while to learn
how to use them all effectively. And it can scan negatives
into a kind of raw format which lets you do later colour
balance. I don't think Silverfast can do the same.

3. In real life, under real usage, what is the minimum PC config I
would want to run your recommended software? processor speed, RAM,
disk storage.

Don't think it makes a diff which of Vuescan or Silverfast you use.
Get a fast CPU, at least 512Mb RAM. And the biggest disk(s) you can
afford. USB2 is a must. More later.

4. Assuming that I expect my workflow (I haven't gone digital quite
yet) to be "centered" on Photoshop, not Photoshop Elements, which
version of Silverfast would I want? And why?

Can't help. Don't use Photoshop myself, I'm a GIMP man.

5. As a side question, what are the specific issues regarding
Kodachrome and recent model Nikon scanners? Can either of these two
packages compensate for these issue?

Can't help: haven't tried Kodachrome yet.

6. Can either of these packages utilize the built-in image correction
features such as ICE, etc.

They can use the hardware infra-red facility but don't use the
proprietary processing algorithm.

I would like to scan most of these images, and do my sorting/culling
after scanning, particular the negatives. Probably in Photoshop, on a
20" monitor. Beats a loupe.

Reason why I went for scanners as well!
If I may offer a suggestion: go for a lcd panel display
rather than a conventional monitor. Much easier on the eyes.

autoloader. Besides, I will stay with film for at least one more
year, possibly longer, until the Nikon digital camera of my "dreams"
is on the market and at a reasonable price. I expect this to occur
not before 2007, maybe 2008.

With luck. The way they're wasting time with gimmicks,
it won't happen... I'm getting ready to go MF meanwhile.

So I expect to end up scanning 15 K, maybe up to 20K images My

Man! You're in for a lot of pain and long hours...

(I'm upgrading my PC in anticipation of large files and many of them.
I will probably use an 80/160 GB SCSI tape drive for backup.)

May I suggest a DVD/CD writer instead? And the best recording
media you can find/afford? Tapes are slow (unless you're prepared
to spend big), can be notoriously unreliable and usually
require specialised software to operate effectively.

With either Windows or Linux, DVD/CD writables are a breeze
nowadays. Pick a good long-lasting media supplier (delkin, verbatim)
and you got a much more cost-effective solution. Besides,
tapes take up a lot of space.

HTH

 

[nailer]

excuse me for being lazy, but your story is long, and my time limited.

1. RAW format applies to digital cameras, scaners work differently.
Save in 16 bit tiff or psd, providing the scanning progie supposts 16
bit scans.
2. Considering you have BW silver based films a scaner with diffused
illumination would be a better choice.
3. Considering all color films - get something with infrared chanel as
well.
4. try vuescan and any other progz by yourself, do not expect a
meaningful reply, unless there is another person with similar needs.
You must decide (compromise).
5. get the fastest PC with as much RAM you can afford. Double core -
why not? hard disks are cheao these days, RAM too.

and most of all - how many days to retirement? It's gonna to take a
while to scan THAT amount of frames.
from my experience with Nikon LS30 on P4 - 3 hours for one roll.


On Thu, 15 Sep 2005 01:04:25 -0700, Father Kodak

*I just joined this group, and just read 40+ messages on this topic.
*All I can say is,
*
*1. there appear to be two viable non-vendor choices. Silverfast and
*VueScan.
*2.Neither one sounds "ideal."
*3. This discussion seems to have morphed into a software engineering
*food fight. I work in a Silicon Valley enterprise software company
*(though in marketing, not engineering), so I can appreciate this
*discussion in a way. However, I suspect most readers of this group
*don't care any longer.
*
*Can I restart this discussion, for my particular situation. My
*questions, for the situation below, are:
*
*1. Between Silverfast and VueScan, which gives me the highest degree
*of automation when I need it (for a box of slides or strip of film)?
*
*2. Which gives me the most flexible options for a custom settings, for
*a more precise scan, probably one at a time?
*
*3. In real life, under real usage, what is the minimum PC config I
*would want to run your recommended software? processor speed, RAM,
*disk storage.
*
*4. Assuming that I expect my workflow (I haven't gone digital quite
*yet) to be "centered" on Photoshop, not Photoshop Elements, which
*version of Silverfast would I want? And why?
*
*5. As a side question, what are the specific issues regarding
*Kodachrome and recent model Nikon scanners? Can either of these two
*packages compensate for these issue?
*
*6. Can either of these packages utilize the built-in image correction
*features such as ICE, etc.
*
*Now, thank you for reading this far. Here is my situation:
*
*I have:
*
* 5000+ black and white negs, mostly Plus-X and
*Tri-X, home developed in various developers. Maybe a small amount of
*black and white film based on C-41.
*
* About 1000 color negatives, various Kodak films, C-22 and C-41
*processes. Mostly in 4 or 6 image strips.
*
* At least 8000 Kodachrome 25 and 64 slides. K12 and K14
*processes. 2 x 2 mounts. Mostly Kodak, some third party lab in
*either cardboard or plastic. Some in uncut rolls.
*
* About 1000 Ektachrome high speed slides, pushed to 1600 and
*3200. Mostly cardboard mounts. Some in rolls.
*
*I would like to scan most of these images, and do my sorting/culling
*after scanning, particular the negatives. Probably in Photoshop, on a
*20" monitor. Beats a loupe.
*
*Being a looong time Nikon user, I will probably opt for the 5000 plus
*the slide autoloader, so I can batch-scan overnight. With this volume
*to be scanned, I can justify the cost of the scanner plus slide
*autoloader. Besides, I will stay with film for at least one more
*year, possibly longer, until the Nikon digital camera of my "dreams"
*is on the market and at a reasonable price. I expect this to occur
*not before 2007, maybe 2008.
*
*So I expect to end up scanning 15 K, maybe up to 20K images My
*equipment cost per scan could be well under ten cents, or ten plus
*scans for one dollar US.
*
*Having my druthers, I would like a "RAW" format output done at 2000
*lines or greater. I expect to be doing digital photography with RAW
*format.
*
*(I'm upgrading my PC in anticipation of large files and many of them.
*I will probably use an 80/160 GB SCSI tape drive for backup.)
*
*Thank you in advance.

 

[Bruce Graham]

Man! You're in for a lot of pain and long hours...

and thats just learning how, then more pain and many more hours doing it,
then starting over as you get better.

Scanning is only worth the effort for your *chosen* images, so don't loop
the loupe just yet.

 

[David Blanchard]

If I may offer a suggestion: go for a lcd panel display
rather than a conventional monitor. Much easier on the eyes.

A discussion in another forum indicated that low-end and middle range
LCDs still don't have the color fidelity that a quality CRT has. If
you choose LCD you should consider high end units (i.e., 20" LCD in
the $2000-2500 price range!!) Me? I'm sticking with my CRT for now.

May I suggest a DVD/CD writer instead? And the best recording
media you can find/afford? Tapes are slow (unless you're prepared
to spend big), can be notoriously unreliable and usually
require specialised software to operate effectively.

Tapes may be slow but typically hold more data than other formats.
I've had very few reliability issues with tapes and I've been doing
tape backups of one type or another for 20 years. Still have the
tapes and can still read them. For "specialized software" on either
Mac or PC, use Retrospect from Dantz.


Also, I noticed in the original query that only Vuescan and Silverfast
were mentioned for use on a Nikon scanner. Why not NikonScan? I've
used all three (Silverfast in demo mode only, however) and continue to
use both Vuescan and Nikonscan. Sometimes one works better than the
other; sometimes the other works better.

-db-

 

[Noons]

Bruce Graham apparently said,on my timestamp of 15/09/2005 8:43 PM:

Scanning is only worth the effort for your *chosen* images, so don't loop
the loupe just yet.

I'm finding that as well and I'm nowhere near the number of images
this guy is going for! What I'm doing now is scanning the older
negatives and slides and archiving them in raw. That way even if the
film strips crumble into dust, at least I got something to go back to.
And only the choice ones of the modern stuff, like you say.
Life is too short...

 

[Don]

It would surprise me if it is really as easy as you say. As I wrote earlier,
hardware manufacturers, and I'm thinking in particular of the optics field,
tend to be protective about their products, if only by employing incompetent
programmers on their side.

No, it really is simple. The number of commands is quite limited and,
as I say, there are even programs out there to monitor the USB bus for
you automatically.

Reverse engineering can be a time eater...

Yes, in theory, but in this case it's relatively simple. You can
usually deduce the number of parameters and their function from
commands you supply to the native software you're monitoring and from
the stack.

Now, if you did not have that native software, then it would be hard.
But given the native software it's relatively easy.

Why not take up contact with the KDE or Gnome people so we can have a KScan,
supporting say the LS-50 and the KM-5400-II?

Because it really isn't anything special and I have no time to take on
official development duties. When I finish scanning my slides all
programs will be mothballed and I will probably sell the scanner
because I've converted to digital.

BTW, what do you use for correlation?

Basically square root of mean of square differences but I tweaked it a
lot. I looked at several approaches and stuff I got from this site
figures prominently:

http://astronomy.swin.edu.au/~pbourke/analysis/correlate/

That might have been IBM's agenda, but what was Apple's interest? At that
time RISC chips were superior to Intel chips, so it was a good idea to
replace the old Moterola 68xxx by one of Alpha, MIPS, Sparc, Power, etc.

Actually, IBM was desperate at that point and were hoping Power PC
could brake Intel's monopoly and, by extension wrestle control back
from Microsoft. I suspect Apple was afraid of that so they sabotaged
the whole project before it even started! Like I say, "it's Apple's
nature", like that scorpion/frog story...

BTW, the Power PC was *not* RISC in spite of all IBM (and Apple)
propaganda! At the time it had over 120 (!!!) instructions and was
implemented on 3 chips (maybe 4, I don't remember anymore - my Power
PC manuals are in storage)! That's not RISC by any stretch of
imagination in spite of RISC-like pipeline and some other relatively
insignificant RISC similarities.

Don.

 

[Don]

I just joined this group, and just read 40+ messages on this topic.
All I can say is,

1. there appear to be two viable non-vendor choices. Silverfast and
VueScan.

A question back: Why do you insist on a non-vendor program?

This is a very important point. If you're just starting, no program
can replace experience i.e. you can't buy experience. And only after
you've grasped it all is when you'll be able to evaluate the programs
according to your specific needs (which will change as you learn!).

2.Neither one sounds "ideal."

It all depends on your requirements. Unfortunately that's a catch-22
and requirements will change with time. Expect to re-start scanning
several times *if* you're examining the results carefully.

1. Between Silverfast and VueScan, which gives me the highest degree
of automation when I need it (for a box of slides or strip of film)?

I'd say there's no substantial difference in this respect.

2. Which gives me the most flexible options for a custom settings, for
a more precise scan, probably one at a time?

Vuescan nominally appears to have more settings but that's misleading
as many are broken. SilverFast is much more stable but aimed more at
the "auto everything" crowd.

3. In real life, under real usage, what is the minimum PC config I
would want to run your recommended software? processor speed, RAM,
disk storage.

For pure *scanning* the requirements are very low. Any computer with
USB 2 will do. Scanners are relatively slow devices.

Now, when it comes to image editing afterwards, the story changes
completely. Get as much RAM as you can and as fast a processor as you
can! Plus a second (even a third!) drive! For example, one for Windows
scratch, one of PS scratch and one for data.

Disk storage size depends on quality of your scans and whether you
will archive the originals or only save the end result, for example,
as JPG. As orientation, a full 16-bit scan at 4000 resolution will
occupy about 133 MB, less if you do some cropping.

4. Assuming that I expect my workflow (I haven't gone digital quite
yet) to be "centered" on Photoshop, not Photoshop Elements, which
version of Silverfast would I want? And why?

Not really related.

Assuming you aim for highest quality all you want from a scanner
program is to produce maximum (that's *native*, *not* interpolated
resolution) and maximum bit-depth (which at this point is 16-bit).

5. As a side question, what are the specific issues regarding
Kodachrome and recent model Nikon scanners? Can either of these two
packages compensate for these issue?

NO!!! And NO again!!!! Oh yes, also, NOOOOOOOO!!!!!! ;o)

Nikon - Kodachrome is a deadly combination! ;o) This is from someone
who's been struggling with it for 3 years now!

However, it all goes back to your requirements and how much quality is
enough quality. There are two basic problems:

A generic one, regarding dynamic range. Kodachromes have a very wide
dynamic range with which even 16-bit scanners struggle (if you want
the highest quality) so you will have to at the very least multi-scan,
or better still use High Definition Range or "Twin Scan" methods.

The specific one is that Nikon's Kodachrome mode does *not* go far
enough and the scans will be still have a blue cast. The darker the
original the more pronounced the blue cast in the scan will be.

Some programs will try to correct this automatically, but it's your
call if that's satisfactory. Both Vuescan and SilverFast are free to
try so give them a spin and compare to NikonScan.

6. Can either of these packages utilize the built-in image correction
features such as ICE, etc.

Silverfast can (I believe, so I'll let users confirm), Vuescan can't!
Indeed, any vendor specific features are absent in Vuescan. SilverFast
implements some but not all (as far as I know). For example, I don't
think SilverFast implements DEE and SIE which are available in
NikonScan. Now, whether you want DEE and SIE is a different question.

Now, thank you for reading this far. Here is my situation:

You're welcome. )

I have:

.... cut ...

The media is not really important except for two special cases:
1. Kodachromes (see above)
2. Silver-based B&W negatives. ICE doesn't work with them.

So I expect to end up scanning 15 K, maybe up to 20K images My
equipment cost per scan could be well under ten cents, or ten plus
scans for one dollar US.

However, your hair-pulling and wall-climbing cost will be very high if
you aim for high quality! ;o)

Having my druthers, I would like a "RAW" format output done at 2000
lines or greater. I expect to be doing digital photography with RAW
format.

If you want to "scan raw" then the scanner program becomes relatively
unimportant. The main difference between scanner programs is what they
do *after* the scan. But if you want to do all that editing yourself,
all you want the scanner program to do is scan.

So the whole equation changes and you want a program that will let you
"turn things off" easily and just scan. As well as mess with the image
the least! Both Vuescan and SilveFast are equally "bad" in this
respect. The former due to bugs/unreliability, the latter due to
stress on "auto everything".

BTW, if you do "scan raw" then you want to scan at *native* resolution
of your scanner.

Thank you in advance.

You're most welcome. Hope it helps.

Don.

 

[Don]

Don, I don't disagree with much in your last post responding to me. I
have no desire to search the archives at this point to read your past
threads looking for examples of bias as I'll just wait for new ones to
appear.

Then prepare for a very long wait! ;o)

The problem has always been that people (*some* Vuescan "fans") often
overreact but are hard pressed when asked for specifics.

One thing that is causing us grief is sloppy syntax.
I can see why you say I misinterpreted the statement above which used
the word "inferior." The original was ambiguous. It wasn't clear that
you were basically saying that some users has self-admittedly
"inferior", or less demanding requirements.

Exactly! I repeat that all the time!

Above paragraph is precisely an example of overreaction without paying
attention to context or actually reading what I write.

I read it as Don saying
that VS users invariably have absolutely inferior requirements. The
original was ambiguous as I read it.

Actually, it wasn't if you read it carefully. I use very precise
language.

About the "overreacting" and "emotion" statements, I think you just
aren't concerned with how your readers perceive what you write and that
you might have difficulty understanding how they would feel.

Not true at all. I'm always considerate, polite and never abusive.

And that's even after some really obnoxious outbursts out there some
aimed at me personally (which I never respond to)!!

So, given the context of such (over-) reactions and my responses your
conclusion is therefore clearly wrong and not based on fact.

Don wrote:
It is *not* belittling!!! It's an explanation because they provide no
fact!

This "explanation" that our conclusions are invalidated by our emotions
doesn't go over well with me as the reader, but clearly you either
don't care or are incapable of caring how others feel. If belittling
others isn't your intent, it's still the result of your words, so you
might want to be mroe careful.

I can't be held responsible when people *misinterpret* what I write
because they don't bother to read!

How can you blame *me* for that? Isn't that the fault of those who
overreact?

You claim to have no emotional investment in this issue and to simply
be a reporter of fact. Let's say that the facts as you see it add up
the conclusion that Vuescan is a buggy, unreliable program with basic
errors that inexplicably reappear and an abusive author. Now given
this conclusion, there is more than one way to react. From this
conclusion you could react with indifference- that you don't use VS so
why would you care.

Because the purpose of this forum is to assist each other. When there
is a question regarding Vuescan and these *essential facts* are
omitted, simply stating them should not be a problem, should it?

And yet when *some* Vuescan "fans" see that, they just go ballistic! I
then calmly supply the evidence, which they can't refute so without
providing any evidence themselves, they just get ever more agitated.

You could react with sacrasm, that VS users are
suckers and deserve what they get.

That's not my style. I may post a short humorous post (see end of
sentence!) but some Vuescan "fans" have demonstrated a total lack of
sense of humor starting with their "great leader" himself. ;o)

But you react with outrage

There you go again!

No, I do NOT! Please stop making such baseless accusations without any
proof!!! Repeating it is not going to make it true!

The only people who overreact with outrage and abuse are *some* of the
said Vuescan "fans"!

Don's answer is almost always an unqualified no

That's just patently false! How can you expect a decent response when
you continue with such, well... slander!

!!!===> How do you explain my supplying the Vuescan site and telling
people to try it themselves, then!? <===!!!

.... the rest based on these false premises omitted for brevity ...

I guess I'm a rabid Vuescan fan after all : )

Fan, definitely, but not a "fan" (i.e. rabid) ;o)

Seriously, you may have been put off by some of my messages initially
because you did *not* view them in full context - i.e. the last 3
years - or really *read* what I wrote but just joined the angry crowd
of peasants with pitchforks... ;o)

Unfortunately, you never took the trouble to reexamine your initial
false impression and continued to base your opinion on those incorrect
first impressions in spite of ample evidence to the contrary.

As you read my messages more carefully (e.g. in this thread!) and
especially looking carefully to catch my "bias" I expect you will
change your opinions considerably.

Just look at this very thread? After *really* examining what I wrote
you already changed your impressions a lot (first sentence in this
message)! Still some way to go, but I'm patient! ;o)

Don.

 

[Don]

Okay, I said I was going to wait for Don's next post, but I was
tempted, and had to look at the beginning of the thread to remember why
I posted here in the first place. It was this statement.

Don wrote about Vuescan:
"However, if you don't care for quality and just want a quick a dirty
web scan it just may do the trick"

This is not the only conclusion one can draw from the "facts" that VS
has and has had bugs. So Vuescan is only okay to use if you want a
"Quick [and] dirty web scan" and don't care about quality?
Correct.

That may be
Don's opinion,

No, objective fact:

When you scan for Web you use very low resolution and output in JPG.
Only a *tiny* fraction of original (or "potential") data is still
present in such an image.

Given all that, it's *impossible* to tell a Vuescan JPG web scan from
the "best scanner program in the Universe" JPG web scan.

Hence, the above statement holds.

but my experience, the experience of other users on this
forum, and the results of the Scanner Bake-off contradict it. Frankly,
I think this hyperbolic statement is unsupported by the evidence Don
has presented to date. Don, how can you "objectively" prove such a
sweeping statement- what criteria would you use?

See above.

Don.

 

[Don]

And, you've written "tons of them" that are equally functional across a
wide range of devices... say 400 or so?

I'm sorry Neil but that shows you really don't understand this at all.

Please don't overreact as I'm only trying to explain it:

The number of devices is totally irrelevant. What you're dealing with
is an *image*. Whether it comes from one device or from 1 billion
devices *doesn't matter*!

Given this image (*regardless* of source!) the task is to crop it!
That's all! Where this image comes from is totally beside the point!

Excuse me? I asked you a question about your knowledge, and YOU responded
with a sophistic reply. To make it clear, your response is based on a
presumption about Hamrick's algorithm, not on knowledge about it. So,
while your "experience" may lead one to think your conclusion is
plausible, lacking specific knowledge and fact, it is not, and therefore
it is misleading, and therefore sophistry.

The above explanation should put things into perspective. (I hope?)

What "all" are you referring to?

Cropping, "universal problem"... all your challenges in the original
message.

What "all" are you referring to?

Chase up the archives for the following thread where you'll find "all"
who have reported this cropping problem (over *several* Vuescan
versions!):

--- cut ---

So it looks like a serious bug with the cropping system, as you suggest. Don
will say "told you so" -well, he did! Stick to your working version.
Upgrade at your peril!

--- cut ---

And, I'd see it as not surprising, especially for a reversed-engineered
product that directly addresses the hardware.

Huh? What does that have to do with anything?

Cropping is *not* related to any one scanner!!! It's to do with the
image *after* it has been received from the scanner!!!

Why am I not surprised?

You are not surprised I don't have a Minolta?

Double huh? What "point" are you trying to make?

First, I'd be surprised, as I have found bugs in Silverfast Ai 6 for my
scanner, and even some UI problems that would affect all scanners.

Keep your eye on the ball and don't change the subject.

Did or does SilveFast *ever* have "Vuscan stripes" when used with
Minolta. *That's* the question!

How do you know, as you don't have a Minolta? How do you know that it
isn't a case that Silverfast/Minolta users just have more realistic
expectations than you do?

Because in *TWO* years not a *single* SilverFast user has complained!

Then, what should one make of your above statement, "What do you base all
your challenges on, besides sophistry?" Not only is this statement "name
calling", it's an outright lie.

Oxford Concise:
Sophist: Captious or fallacious reasoner, quibbler.

Responding to everything with an argumentative "how do you know"
without providing any fact would be a canonical example of that.

The rest, I'm afraid, is far too incongruent to respond in any
meaningful way.

Don.

 

[DBM]

I do not know why people continue to argue with Don. Don apparently exists
for controversy and debate even if it does not add to the knowledge of the
group. I have met a few people like Don in my lifetime. You will never
come out ahead in a discussion with Don as he knows it all and only he is
correct. As he says one needs to read the archives. I have been reading
this group for years now and in several instances he encourages an argument
so he can continue to *prove* his knowledge and debating skills. He has yet
to post any examples of his skills for this group. He makes claims that is
not supported with examples posted for all to read and see.



dbm

 

[HvdV]

Hi Roger,

(and how it's supposed to work when it fails!). VS can be a plausible
alternative to other scanner software if you are willing to spend the
time to figure out the interface and how it works and can yield good
results. I come to this forum and photo.net to share what I learned
the hard way after a lot of frustration with VS and a scanner in
general.

You have to admit that this is a bit strange: you pay for a program with a
GUI and need to 'figure out the interface'. Unless you're dealing with
something very complex, a GUI should be self explanatory, or at least there
should be a well written manual with a 'howto' section to prevent any
frustration. In these respects the software which comes with the KM5400 is
better, too bad its 16bit support is practically absent.

I guess I'm a rabid Vuescan fan after all : )

Can't say I have reached that state yet with Silverfast...

-- Hans

 

[HvdV]

Because it really isn't anything special and I have no time to take on
official development duties. When I finish scanning my slides all
programs will be mothballed and I will probably sell the scanner
because I've converted to digital.

You might consider putting the stuff on sourceforge and see whether someone
takes over.

Basically square root of mean of square differences but I tweaked it a
lot. I looked at several approaches and stuff I got from this site
figures prominently:

http://astronomy.swin.edu.au/~pbourke/analysis/correlate/

Did you get sufficient accuracy with that?

Actually, IBM was desperate at that point and were hoping Power PC
could brake Intel's monopoly and, by extension wrestle control back
from Microsoft. I suspect Apple was afraid of that so they sabotaged
the whole project before it even started! Like I say, "it's Apple's
nature", like that scorpion/frog story...

Fortunately, in this case the frog was poison proof and is making heaps of
profit from its effort.

BTW, the Power PC was *not* RISC in spite of all IBM (and Apple)
propaganda! At the time it had over 120 (!!!) instructions and was
implemented on 3 chips (maybe 4, I don't remember anymore - my Power
PC manuals are in storage)! That's not RISC by any stretch of
imagination in spite of RISC-like pipeline and some other relatively
insignificant RISC similarities.

Well, neither the scorpion nor the frog are renowned for their beauty. Other
designs like the R3000 might have been cleaner, but were are they now? (in
your dishwasher, maybe even in your scanner )
Anycase, this is more a topic for comp.arch

-- Hans

 

[Neil Gould]

Cropping, "universal problem"... all your challenges in the original
message.

Again, I asked you two questions about your KNOWLEDGE, both of which can
be answered by facts (if you had such facts, which it is now clear that
you don't): "How do you KNOW that the algorithm is simple?" where the ONLY
factual reply would be "Because I've seen the algorithm, and it's simple";
and, "How do you *know* it's a unversal problem?" where the ONLY factual
reply would be "Because the alogorithm and subsequent problem can be shown
to be device-independent." Neither was your reply, and what you did come
back with is completely subjective and unprovable.

Here's two data points for you:
1) VueScan's cropping works exactly as expected with my scanners, and I've
never noticed a problem of any kind with this aspect of the program.

2) In looking at the list of changes to VueScan on Hamrick's site, I
didn't see one reference to fixing a "cropping problem". If it were a
"universal problem", it would probably be mentioned.

Item one proves your notion of universality to be patently false. Item two
suggests that there wasn't a universal problem to begin with, though it
could be inconclusive if not for item one.

Chase up the archives for the following thread where you'll find "all"
who have reported this cropping problem (over *several* Vuescan
versions!):

So, the "all" you are referring to is "all of those that reported having
the problem in comp.periphs.scanners"? Why would someone that has no such
problem report anything? Beyond that, you do realize that the users in
this newsgroup that experience problems don't qualify as "universal",
don't you?

Keep your eye on the ball and don't change the subject.

"The ball" is your statement that "SilverFast has *no* problems with
Minolta..." If nothing else, the UI problems that I've noticed (and are
recognized by LaserSoft, btw) would affect Minoltas as well as any other
scanner, ergo, your claim is patently false.

Did or does SilveFast *ever* have "Vuscan stripes" when used with
Minolta. *That's* the question!

I thought you didn't want to change the subject? YOU JUST DID!

Because in *TWO* years not a *single* SilverFast user has complained!

That only supports the idea that SF/Minolta users have more realistic
expectations than you. FWIW, I don't complain about either SF or VueScan;
they are both useful, and they both have "problems".

Neil

 

[Neil Gould]

Hi Roger,


You have to admit that this is a bit strange: you pay for a program
with a GUI and need to 'figure out the interface'.

Comprehending a UI is a pretty subjective matter. What is easily grasped
by one person may be all but opaque to another.

Compared to some other scanning programs, I find SF's UI to be redundant,
circular, and at times frustrating. OTOH, VueScan's UI seems a bit too
rudimentary in its use of dialog boxes. For example, there doesn't appear
to be adequate consideration for font sizes as displayed by different
monitor card resolutions, which can make setting its parameters difficult.

Neil

 

[Roger]

On Thu, 15 Sep 2005 01:04:25 -0700, Father Kodak

OK, I'll add my thoughts to the mix.
(Sorry, I didn't mean to get carried away)

I just joined this group, and just read 40+ messages on this topic.
All I can say is,

1. there appear to be two viable non-vendor choices. Silverfast and
VueScan.

Don't discard NikonScan either

2.Neither one sounds "ideal."

I've not found one yet, I'd consider "ideal". Each has their points.
I contantly switch between NikonScan and VueScan. I like them both.

3. This discussion seems to have morphed into a software engineering
food fight. I work in a Silicon Valley enterprise software company
(though in marketing, not engineering), so I can appreciate this
discussion in a way. However, I suspect most readers of this group
don't care any longer.

Can I restart this discussion, for my particular situation. My
questions, for the situation below, are:

1. Between Silverfast and VueScan, which gives me the highest degree
of automation when I need it (for a box of slides or strip of film)?

"I think" VueScan will give you a higher degree of control, but with
that comes a steeper learning curve, and more work at least initially.

On thing to remember about slides. NO program or pieces of equipment
is completely reliable with auto feeders. It may run trouble free for
hours only to jam when you turn your back, or it may jam incessantly.
This is more the fault of the slides than the equipment. Paper slide
holders are prone to curling as well as edge curling and spreading.
Some plastic slide mounts refuse to feed in specific directions. I
had several thousand that would only feed backwards which is no big
problem, you just have to remember to do it.

2. Which gives me the most flexible options for a custom settings, for
a more precise scan, probably one at a time?

Again, I would say, VueScan, but much of the work will most likely go
just fine using NikonScan.

3. In real life, under real usage, what is the minimum PC config I
would want to run your recommended software? processor speed, RAM,
disk storage.

Image processing is one of the most CPU intensive operations you can
find. As an opinion, get the biggest, baddest machine you can afford
and then at least tripple the size of the drive you thought you
needed.

4. Assuming that I expect my workflow (I haven't gone digital quite
yet) to be "centered" on Photoshop, not Photoshop Elements, which
version of Silverfast would I want? And why?

I'd get the latest version of what ever, but check the forums to see
if any one is having a problem with a specific scanner with that
version.

5. As a side question, what are the specific issues regarding
Kodachrome and recent model Nikon scanners? Can either of these two
packages compensate for these issue?

Some claim so, I've never seen any "with the exception" of the IR
cleaning regardless of scanner can be inconsistent between batches of
Kodachrome. It's a dye transfer process which is quite different from
Ektachrome, Fuji, and others (E7).

6. Can either of these packages utilize the built-in image correction
features such as ICE, etc.

Kinda, sorta. They have their own way of using the IR source, but
*seem* to work fine for me with the exception of some Kodachrome.

Now, thank you for reading this far. Here is my situation:

I have:

5000+ black and white negs, mostly Plus-X and
Tri-X, home developed in various developers. Maybe a small amount of
black and white film based on C-41.

ICE does not work with B & W negatives so you do not have the luxury
of IR scratch and dust removal. I've not seen a small scanner yet
where I liked the results with B & W, but there are many I haven't
tried as well.

About 1000 color negatives, various Kodak films, C-22 and C-41
processes. Mostly in 4 or 6 image strips.

At least 8000 Kodachrome 25 and 64 slides. K12 and K14
processes. 2 x 2 mounts. Mostly Kodak, some third party lab in
either cardboard or plastic. Some in uncut rolls.

How long have your had these masochistic tendencies?
You are talking many hundreds of hours of work here.

http://www.rogerhalstead.com/scanning.htm may help, but I do need to
update it.

About 1000 Ektachrome high speed slides, pushed to 1600 and
3200. Mostly cardboard mounts. Some in rolls.

Rolls are easier in 5 image strips (standard holder size) or if you
have a roll adapter. If you have negatives or positives rolled up
tight you will never do so again after trying to scan the warped
little buggers.

I would like to scan most of these images, and do my sorting/culling
after scanning, particular the negatives. Probably in Photoshop, on a
20" monitor. Beats a loupe.

There may be some you not only don't want to bother scanning, there
may be some you don't want to get in your equipment. As the basic
scanning with IR cleaning on the LS-5000 ED runs about 30 seconds and
you can easily take that well past a minute culling does make sense.

I did the "old family photos" so every thing was scanned and I
probably have another year or two working part time to finish. I'm
past twenty some thousand and have about 70 some DVDs full with
another set as backup.

If you use Photoshop in conjunction with VueScan it can open the
images automatically in Photoshop for editing. Scanning a strip of
five negatives or slides takes a lot of resources. When you add
Photoshop to that and the images to it, you need a *LOT* of
horsepower. I found going from 512 megs to one Gig of RAM was like
night and day. This computer is a 64 bit, 3.4 Gig Athlon with 2 Gigs
of DDR RAM at 400 MHz. The network is approaching three *terabytes*
and will hit four shortly.

I do make liberal use of external USB drives. A couple of them are
250 Gig, but the last three are all 300 Gig. They are faster than
most networks, but I'm using a hardwired gigabit network via Cat5e.
Most of the machines have two external drives and two or three
internal. One has a 400 Gig SATA RAID and this one will be getting
one soon.

One note on the USB drives. If you purchase the drive and enclosure
seperately you can get the very same drive and enclosure for $30 to
$50 less. This involves about 4 screws and two cables. Either way
you still have to format the drive. I leave them as all one
partition.

Being a looong time Nikon user, I will probably opt for the 5000 plus
the slide autoloader, so I can batch-scan overnight. With this volume

Probably not. The autoloader is a nice one, but you are at the mercy
of your slides. You learn to roll the edges of paper slide holders by
smoothing them with the back of a thumbnail or the handle of a kitchen
knife.

to be scanned, I can justify the cost of the scanner plus slide
autoloader. Besides, I will stay with film for at least one more
year, possibly longer, until the Nikon digital camera of my "dreams"
is on the market and at a reasonable price. I expect this to occur
not before 2007, maybe 2008.

Going with the voice of experience I'd say It's going to take you
longer than that to scan in all those images.

As to the dream camera, I have a couple of nice digital cameras, but
nothing fancy. A Olympus E20N and a Nikon D70 with a good set of
lenses. I also have an F4S and my spare is an 8008S. I gave up long
ago waiting for my "dream camera".

So I expect to end up scanning 15 K, maybe up to 20K images My
equipment cost per scan could be well under ten cents, or ten plus
scans for one dollar US.

That is one big job.

Having my druthers, I would like a "RAW" format output done at 2000
lines or greater. I expect to be doing digital photography with RAW
format.

Tis not quite the same thing, but you can scan with the basic settings
set to get the most out of image's dynamic range and then post
process. I scan at 4000 dpi, but you are looking at some very *large*
files. That's 60 some megs per image at 8 bit color depth and 128
megs at 16 bit color depth for a full frame 35 mm negative of slide.

(I'm upgrading my PC in anticipation of large files and many of them.
I will probably use an 80/160 GB SCSI tape drive for backup.

My smallest drive is a 250 Gig although I do have a 400 Gig RAID made
from two 200 Gig SATA drives.

My own preference is to stay as far away from tape drives as I can
get. My profession was computers, I have my degree in the field, and
I worked my way up to project manager so I have more than a passing
acquaintance with them. Good tape drives are expensive and do not
allow for random searches as you can do on a CD and DVD.

All 4 computers here have dual layer DVD drives. Single layer DVDs
(4.6 Gig) are almost free if you shop around a bit. My last 25 were
free with rebate and I think I paid 20 or 30 cents for the hundred
prior to those.

A word of caution on storage of digital images. No medium is
permanent. Generally, magnetic media such as hard drives are
considered temporary although I'd think they have lifetimes measured
in years. If you use windows, use NTFS and not FAT 32.

"We thought" images were going to be long lived on CD and DVD, but
some strange ailments are turning up in some isolated cases. Remember
the back of the CD is sensitive to damage while the face can be
blocked with scratches and dirt. Still the information is burned into
the layer on the back. DVDs are a sandwich and are sensitive to
flexing. The writing is done between the two layers of plastic. Do
not pop them out of jewel cases by lifting on the edges. Press down in
the center and dump it in your hand, but holding it by the edges.
Store CDs and DVDs on edge, in a cool dry place out of direct
sunlight.

Remember too, that no mater how good your filing system, your back up
system, and your equipment, most trashed files come from user mistakes
and not the equipment.

Good Luck,

(20,000 plus done, another few thousand to go and about 200# of prints
yet to scan)

Roger Halstead (K8RI & ARRL life member)
(N833R, S# CD-2 Worlds oldest Debonair)
www.rogerhalstead.com