LS-50: Initial impressions

[Don]

I got the LS-50 a few days ago but due to various things like bad
luck, sloppy programming (both Nikon's and Microsoft's), etc... I just
managed to get it going. Long story but 2 notebooks, 4 disks, 2 OSes
and 5 days later the LS-50 is finally up and running with my old LS-30
exibiting a bad case of sibling rivalry.

Multi-pass multi-scanning just doesn't work, as I suspected, because
4000 dpi is just too precise and it's impossible to align two images,
let alone multiple ones. Unfortunately, looking at dark areas, again
as I suspected, shows how essential multiscanning really is.

However, I'm very impressed with ICE4 (on Kodachromes). It really does
remove dust and scratches without "fuzzification" like ICE(1) did on
my LS-30 (non-Kodachromes). Furthermore, ICE4 also seems to smooth out
the image without softening (the edges remain sharp) which is very
suprising. It almost looks like the image was multiscanned about 4
times, if not more. Quite amazing. But those are first impressions and
I'm sure I'll find a few warts down the road.

Speaking of Kodachromes, LS-50 seems to do the same things to correct
the notorious blue cast (actually absence of red) I used to do
manually on my LS-30 with Analog Gain (after months of frustration)
only LS-50 seems somewhat conservative when applying this Kodachrome
correction.

I'm not too keen on ROC/GEM/DEE on principle, but I ran one scan each
to see what they do. Not impressed, but this was a cursory glance.

4000 dpi is really nice but with file sizes to match. Initially this
caused some slowdown during scanning but once I split disks usage over
3 disks (one for the OS TEMP directory, one for the NikonScan TEMP
directory, and one to save the images) things picked up considerably.

The dust cover on the MA-21 slide adapter is a nice touch but I find
the slide goes in too deep and it's hard to fit it in snugly like it
was on the MA-20 slide adapter in my old LS-30. This shows up with
slides not perfectly straight. I still have to try the old FH-2 film
strip holder. Also, the scanner is BIG!

Not so sure about NikonScan 4, though. I ran the 4.0.1 upgrade on the
CD but this never shows up in the program itself - it still reads
4.0!? Sloppy... I also managed to crash NS4 - which never happened
with NS3. Actually, it even brought Win2K down! Quite an
"accomplishment"! I see Nikon still employs C "programmers"... ;o)

All in all, good impressions, but it's early days. I only ran a couple
of tests and I'm sure various kinks will start showing up as I
continue using it. After all, it's a Nikon... ;o)

Don.

 

[David J. Littleboy]

I got the LS-50 a few days ago but due to various things like bad
luck, sloppy programming (both Nikon's and Microsoft's), etc... I just
managed to get it going. Long story but 2 notebooks, 4 disks, 2 OSes
and 5 days later the LS-50 is finally up and running with my old LS-30
exibiting a bad case of sibling rivalry.

Multi-pass multi-scanning just doesn't work, as I suspected, because
4000 dpi is just too precise and it's impossible to align two images,
let alone multiple ones. Unfortunately, looking at dark areas, again
as I suspected, shows how essential multiscanning really is.

You're doing something wrong. The whole point of the Nikon design is that it
does multi-sampling from all four of the RGBI channels for a given pixel
without moving the film. NikonScan should have a 1x to 16x setting
somewhere. That's all you need.

David J. Littleboy
Tokyo, Japan

 

[Don]

You're doing something wrong. The whole point of the Nikon design is that it
does multi-sampling from all four of the RGBI channels for a given pixel
without moving the film. NikonScan should have a 1x to 16x setting
somewhere. That's all you need.

LS-50 doesn't do *single-pass* multi-scanning. (LS-5000 does.)

What I'm referring to is *multi-pass* multi-scanning. The idea is to
repeat the scan several times and then manually combine these scans in
Photoshop.

It can still be done but requires sub-pixel image shifting which is
very finicky and time consuming.

Don.

 

[Philip Homburg]

You're doing something wrong. The whole point of the Nikon design is that it
does multi-sampling from all four of the RGBI channels for a given pixel
without moving the film. NikonScan should have a 1x to 16x setting
somewhere. That's all you need.

Of course, the firmware in the CoolScan V doesn't support multi-sampling.
Nikon's budget models always lack some useful features.

 

[Ralf R. Radermacher]

Of course, the firmware in the CoolScan V doesn't support multi-sampling.
Nikon's budget models always lack some useful features.

And the most annoying thing is that it's mostly firmware.

At least they have different interfaces, nowadays. But the whole
difference between an LS-30 and a LS-2000 is a few bits of firmware and
1000 USD.

Ralf

 

[Mike Engles]

I got the LS-50 a few days ago but due to various things like bad
luck, sloppy programming (both Nikon's and Microsoft's), etc... I just
managed to get it going. Long story but 2 notebooks, 4 disks, 2 OSes
and 5 days later the LS-50 is finally up and running with my old LS-30
exibiting a bad case of sibling rivalry.

Multi-pass multi-scanning just doesn't work, as I suspected, because
4000 dpi is just too precise and it's impossible to align two images,
let alone multiple ones. Unfortunately, looking at dark areas, again
as I suspected, shows how essential multiscanning really is.

However, I'm very impressed with ICE4 (on Kodachromes). It really does
remove dust and scratches without "fuzzification" like ICE(1) did on
my LS-30 (non-Kodachromes). Furthermore, ICE4 also seems to smooth out
the image without softening (the edges remain sharp) which is very
suprising. It almost looks like the image was multiscanned about 4
times, if not more. Quite amazing. But those are first impressions and
I'm sure I'll find a few warts down the road.

Speaking of Kodachromes, LS-50 seems to do the same things to correct
the notorious blue cast (actually absence of red) I used to do
manually on my LS-30 with Analog Gain (after months of frustration)
only LS-50 seems somewhat conservative when applying this Kodachrome
correction.

I'm not too keen on ROC/GEM/DEE on principle, but I ran one scan each
to see what they do. Not impressed, but this was a cursory glance.

4000 dpi is really nice but with file sizes to match. Initially this
caused some slowdown during scanning but once I split disks usage over
3 disks (one for the OS TEMP directory, one for the NikonScan TEMP
directory, and one to save the images) things picked up considerably.

The dust cover on the MA-21 slide adapter is a nice touch but I find
the slide goes in too deep and it's hard to fit it in snugly like it
was on the MA-20 slide adapter in my old LS-30. This shows up with
slides not perfectly straight. I still have to try the old FH-2 film
strip holder. Also, the scanner is BIG!

Not so sure about NikonScan 4, though. I ran the 4.0.1 upgrade on the
CD but this never shows up in the program itself - it still reads
4.0!? Sloppy... I also managed to crash NS4 - which never happened
with NS3. Actually, it even brought Win2K down! Quite an
"accomplishment"! I see Nikon still employs C "programmers"... ;o)

All in all, good impressions, but it's early days. I only ran a couple
of tests and I'm sure various kinks will start showing up as I
continue using it. After all, it's a Nikon... ;o)

Don.

Hello

Multiscanning does work, if you do not close the twain module.
The LS50 is really very consistent.

Mike Engles

 

[ThomasH]

Yeah, I still have my notes about "installation" of LS4000 with
Nikon LS4000ED. I even called the 13 days(?if my memory serves me)
I needed to get it done "time BS" (before scanning.) I got several
"level 2" calls to their support and I made an over 10 letters size
pages long list of bugs and problems with Nikon Scan 3.0, Nikon 4
has also an impressively long list of issues.

We are blessed by Vuescan!

You're doing something wrong. The whole point of the Nikon design is that it
does multi-sampling from all four of the RGBI channels for a given pixel
without moving the film. NikonScan should have a 1x to 16x setting
somewhere. That's all you need.

I think this is one of the things which they disabled in LS50 on purpose!

A similar strange issue is with SA21 versus SA30: The SA30 front inset
is mechanical identical (except for a small brush on the entry) but
the firmware disables reading of longer film strips with SA21.

Just as well they could be selling the rear roll only for a fraction
of the price, but so you are paying much more for the duplicate of the
entire complex thing.

Thomas

PS: SA30 is my primary scanning tool. I simply could not imagine
scanning without batch scan functionality. Since scanners cost
less now, if your budget allows it: ....take LS5000 rather!!
It is most probably the last scanner you will get in your life,
spread the cost over several years and see how few it costs
per month!

 

[Don]

Hello

Multiscanning does work, if you do not close the twain module.
The LS50 is really very consistent.

Mike Engles

Hi Mike,

I don't use the TWAIN module but run NikonScan standalone.

However, what is most important is not to remove the slide between the
scans. But even then there are minor misalignments. I was getting very
good results with LS-30, but the problem with LS-50 is twofold:

First, of course, the higher resolution. So even the smallest of
misalignments are much more pronounced. Secondly, I find that on the
LS-50 the slide goes in too deep and it's hard to fit it properly as I
was able to do on the LS-30.

One caveat, so far I only used cardboard mounted slides (Kodakchromes)
and that may have aggravated the problem. We'll see what happens when
I try plastic mounts or the film strip holder.

Don.

 

[Don]

And the most annoying thing is that it's mostly firmware.

At least they have different interfaces, nowadays. But the whole
difference between an LS-30 and a LS-2000 is a few bits of firmware and
1000 USD.

In most cases the capability is still in the firmware but simply
disabled (after identifying the hardware) because it's much easier to
maintain a single code repository rather than fork it.

But, then again, considering how buggy Nikon software is, who knows
what they're doing... ;o)

Don.

 

[Don]

We are blessed by Vuescan!

I would venture a guess that we have a widely different definition of
"blessed"... ;o)

Warts and all, but I'd pick NikonScan any day.

I think this is one of the things which they disabled in LS50 on purpose!

Indeed! Despicable marketroids...

A similar strange issue is with SA21 versus SA30: The SA30 front inset
is mechanical identical (except for a small brush on the entry) but
the firmware disables reading of longer film strips with SA21.

That's why I was concerned regarding the FH-2 strip holder (from my
LS-30) and the new MA-21 slide adapter. The aforementioned abominable
marketroids decided to leave the FH-3 (FH-2's successor) out of the
package and make it an "optional extra"... :-/

Since scanners cost
less now, if your budget allows it: ....take LS5000 rather!!

I would, but I just coudln't afford it at this time. Actually, I'd be
quite happy with the LS-4000 as I don't really need the twin CCD
arrays. But it's too late now...

Don.

 

[David J. Littleboy]

LS-50 doesn't do *single-pass* multi-scanning. (LS-5000 does.)

Oops. Sorry for the noise.

David J. Littleboy
Tokyo, Japan

 

[Raphael Bustin]

But, then again, considering how buggy Nikon software is, who knows
what they're doing... ;o)

I don't quite agree. I had some installation problems
when I first got my LS-8000 a few years ago, but it's
been rock solid since then. Just a few gripes:

1. haphazard identification of frame
boundaries occasionally on 645 film

2. if using auto exposure and "normal"
workflow, compression of shadow tones
when scanning negtives.


#1 is an annoyance. #2 is more serious,
but there are excellent workarounds that
don't hurt much.

As scanner drivers go, NS is about as
good as any I've seen.


rafe b.
http://www.terrapinphoto.com

 

[Philip Homburg]

On Wed, 16 Jun 2004 21:28:03 GMT, (e-mail address removed) (Don) wrote:
As scanner drivers go, NS is about as
good as any I've seen.

I don't know how bad other scanner drivers are, but on an absolute scale,
NikonScan 3 is quite bad. However, it is possible to make good scans
with NikonScan and it doesn't crash.

 

[Philip Homburg]

:
PS: SA30 is my primary scanning tool. I simply could not imagine
scanning without batch scan functionality. Since scanners cost
less now, if your budget allows it: ....take LS5000 rather!!
It is most probably the last scanner you will get in your life,
spread the cost over several years and see how few it costs
per month!

There is always the LS-9000... I think that for A3 sized prints you want a
6000 dpi scanner. It should be possible to build a 9000 dpi scanner based
on the sensor in the LS-8000. For high contrast subjects such as trees
against the sky, 4000 dpi is bit low.

 

[Don]

Oops. Sorry for the noise.

No problem, David. I still appreciate you trying to help.

Don.

 

[Don]

I don't quite agree. I had some installation problems
when I first got my LS-8000 a few years ago, but it's
been rock solid since then. Just a few gripes:

I was half-facetious, up there, but...

My installation "problems" were really more of an ambiguous nuisance
with both NikonScan (NS) and Microsoft to blame.

NS4 refused to install on my Win98 - it's really picky about 98*SE*. I
have installed all the Service Packs and - according to MS - that
should be functionally equivalent to SE, but Nikon begs to differ...

In all this time (since 1998) I have only had one other program insist
that what I have is not SE (Adobe Acrobat). All other software
requiring SE was quite happy to run on my 98+Service Packs and that
includes some low level drivers like the PCMCIA USB 2.0 card.

So, I had to install Win2K on a spare drive and that opened up another
can of worms as I got mired in dual boot contortions... 5 days
later...

1. haphazard identification of frame
boundaries occasionally on 645 film

I personally try to use as few of the "auto" functions as possible and
prefer to do things like cropping and exposure manually.

2. if using auto exposure and "normal"
workflow, compression of shadow tones
when scanning negtives.

That's why I may do my first scan with "auto" just to get a rough
ballpark. After that I fine tune manually. I even turn clipping off.

As scanner drivers go, NS is about as
good as any I've seen.

I'm generally pleased too although I don't use most of the additional
functionality such as curves and friends. I prefer to do all that in
postprocessing with Photoshop. I just basically want the scanner
driver to, well... drive the scanner...

NS3 never crashed once, and the NS4 crash (bringing W2K down) was when
during Preview I inadvertently clicked on the histogram button.
Haven't repeated the procedure since because I'm too busy trying
things and don't fancy going through the defrag should W2K crash
again.

Instead, right now I go by "if it hurts, don't do it" although I plan
to return to this eventually to get to the bottom of it.

Don.

 

[Mike Engles]

Hi Mike,

I don't use the TWAIN module but run NikonScan standalone.

However, what is most important is not to remove the slide between the
scans. But even then there are minor misalignments. I was getting very
good results with LS-30, but the problem with LS-50 is twofold:

First, of course, the higher resolution. So even the smallest of
misalignments are much more pronounced. Secondly, I find that on the
LS-50 the slide goes in too deep and it's hard to fit it properly as I
was able to do on the LS-30.

One caveat, so far I only used cardboard mounted slides (Kodakchromes)
and that may have aggravated the problem. We'll see what happens when
I try plastic mounts or the film strip holder.

Don.

Hello

In my experiments, I did 8 scans and found the error was 1 pixel in the
long dimension at 4000dpi.
That is pretty good. What I did find was that multiscanning did not do a
great deal.

Mike Engles

 

[Don]

Hello

In my experiments, I did 8 scans and found the error was 1 pixel in the
long dimension at 4000dpi.
That is pretty good. What I did find was that multiscanning did not do a
great deal.

Hi,

On the LS-30 I don't think I ever got a full pixel shift, it would
always be sub-pixel - a half or a third of a pixel. Also,
multi-scanning did improve things quite a bit even after only 4 scans,
but that's because LS-30 is so limited.

I'm still testing the LS-50 re pixel shift, but re quality/noise in
dark areas even a single pass scan is so much better, especially with
ICE4 on, as I mentioned elsewhere. In my initial tests, ICE4 (in
addition to removing dust and scratches) also appears to reduce grain
without loss of sharpness!

Don.

 

[Kennedy McEwen]

In my initial tests, ICE4 (in
addition to removing dust and scratches) also appears to reduce grain
without loss of sharpness!

It can do, if it interprets the grain as defects and attempts to conceal
it as such. Remember that the resolved contrast of the image recorded
on the film has deteriorated virtually to zero by the time it reaches
granular dimensions (ie. MTF is zero well before the spatial frequencies
of a granular scale are reached) so the effect on the final image may
not appear noticeably softer. However, be careful with Kodachrome since
this effect of ICE can vary to a much more severe level, literally from
one image to another, simply due to the amount of retained silver in the
final emulsion. You can find KC films where some frames scan almost
perfectly without loss of sharpness and others look positively soft
until ICE is turned off. Just a word of warning before your assumptions
become enforced.

 

[Mike Engles]

Hi,

On the LS-30 I don't think I ever got a full pixel shift, it would
always be sub-pixel - a half or a third of a pixel. Also,
multi-scanning did improve things quite a bit even after only 4 scans,
but that's because LS-30 is so limited.

I'm still testing the LS-50 re pixel shift, but re quality/noise in
dark areas even a single pass scan is so much better, especially with
ICE4 on, as I mentioned elsewhere. In my initial tests, ICE4 (in
addition to removing dust and scratches) also appears to reduce grain
without loss of sharpness!

Don.

Hello

Nikonscan is now updated to 4.02. They have cleared up the problem with
the colour and contrast difference between preview and 14 bit output.
Dark kodachromes in conjunction with ROC/GEM are very good.
I really do think that Nikonscan is now very good. They have also had a
look at the problems with multiple images on a strip, but I have not
looked at that yet.

Mike Engles