SHARPNESS problem - Nikon Coolscan - need hints or comments

[Ian Woodrow]

Hi

I'm relatively new to slide scanning and now have a Nikon Coolscan III
(LS-30) to mess around with. Now, one thing I noticed was that using the
Nikon Scan software the colour of the initial scan is pretty bad with a lot
of blue tint... it would need a LOT of footering and manipulating to improve
the scan. However, I tried VueScan and it seems to make a good job
colourwise of the initial scan but in both cases the SHARPNESS is not what I
would expect. Yes, there is always a BIT of manual sharpening to be done
with Photoshop or whatever, but the initial scan from my LS-30 doesn't look
too good at all. A friend of mine told me the sort of unsharp mask values
he normally uses with scans but using similar I have no joy. Even
increasing by quite a bit i am not getting great results. I notice this
most easily where the subject has writing (as writing is detailed so more
noticeably blurred).

Perhaps someone else out there in NewsLand has had similar problems or knows
what I am talking about ? I have been scanning at 3200dpi initially. If
someone would like to have a look at a couple of sample scans I can scan to
whatever spec you want .... let me know what size/resolution etc you would
normally scan at and I can scan a couple at the same and email the results
for your comments.


Cheers for any hints or help


Woody

 

[Philip Homburg]

I'm relatively new to slide scanning and now have a Nikon Coolscan III
(LS-30) to mess around with.

Perhaps someone else out there in NewsLand has had similar problems or knows
what I am talking about ? I have been scanning at 3200dpi initially. If
someone would like to have a look at a couple of sample scans I can scan to
whatever spec you want .... let me know what size/resolution etc you would
normally scan at and I can scan a couple at the same and email the results
for your comments.

First of all, the LS-30 is supposed to have a resolution of 2700 ppi.

Second, make test scans without ICE, and with the focus set to the area of
interest.

If after that, a scan of a sharp slide is not sharp, you may have a dirty
mirror.

 

[Don]

I'm relatively new to slide scanning and now have a Nikon Coolscan III
(LS-30) to mess around with. Now, one thing I noticed was that using the
Nikon Scan software the colour of the initial scan is pretty bad with a lot
of blue tint... it would need a LOT of footering and manipulating to improve
the scan.

There are two causes for this. One is Nikon Color Management. Turn it
OFF! It's useless.

Secondly, you're probably scanning Kodachromes (all Nikon scanners
have problems with them). In general you need to boost red Analog Gain
(a lot) and green to some extent in order to get an image with
sufficient dynamic range in all channels to be able to fix it up
later.

However, I tried VueScan and it seems to make a good job
colourwise of the initial scan

Be aware that VueScan is very buggy. However, depending on your
requirements, it may do the job. Also, if the original is not
"perfect" (as in dark Kodachromes) VueScan will have problems with it.
Whether you notice them depends on how much quality you're after.

but in both cases the SHARPNESS is not what I
would expect.

Again, only a guess, but are you scanning mounted slides? If so, it's
probably warped film which is causing parts of the image to be out of
focus. For all their other many advantages Nikons do have a narrow
depth of field.

Perhaps someone else out there in NewsLand has had similar problems or knows
what I am talking about ? I have been scanning at 3200dpi initially.

This may be the cause of your sharpness problems. LS-30's native
resolution is 2700 and if you are using 3200 then the image is
enlarged in software. That causes blurring. Especially, when you use
such odd upsampling.

Don.

 

[David Blanchard]

There are two causes for this. One is Nikon Color Management. Turn it
OFF! It's useless.

This statement has been made so many times it has become a "fact" ;-)

So, the other day, I tested Nikon CM on a slide (not Kodachrome; Fuji
Velvia actually) of very colorful wildflowers. I scanned with CM
turned on and with it turned off. And my results show that the colors
with it turned on are better and more true to both the slide than with
it turned off.

I'm using NikonScan 3.1.2 on a Macintosh. I'm curious if the claims
of poor CM are from those using other operating systems.

-db-

 

[Don]

This statement has been made so many times it has become a "fact" ;-)


So, the other day, I tested Nikon CM on a slide (not Kodachrome; Fuji
Velvia actually) of very colorful wildflowers. I scanned with CM
turned on and with it turned off. And my results show that the colors
with it turned on are better and more true to both the slide than with
it turned off.

I'm using NikonScan 3.1.2 on a Macintosh. I'm curious if the claims
of poor CM are from those using other operating systems.

Good point!!

However, years ago when I was still wet behind the ears, I wrote to
Nikon's so-called "support" (exactly about Kodachromes on LS-30) and
they themselves advised that NCM be turned off - permanently!

Now, Nikon "support" have since proven repeatedly how incompetent they
are but in this case they may have just got it right. As the saying
goes: "Even a broken clock is correct twice a day"... ;o)

Seriously though, you may very well be correct that the problem is OS
specific because I'm, indeed, on a Windows machine. I'm also curious
what others will say.

Don.

 

[John]

This statement has been made so many times it has become a "fact" ;-)

So, the other day, I tested Nikon CM on a slide (not Kodachrome; Fuji
Velvia actually) of very colorful wildflowers. I scanned with CM
turned on and with it turned off. And my results show that the colors
with it turned on are better and more true to both the slide than with
it turned off.

I'm using NikonScan 3.1.2 on a Macintosh. I'm curious if the claims
of poor CM are from those using other operating systems.

-db-

I use NikonScan 4 on Windows 2000. I agree that at first sight, the colours
seem closer with Nikon Colour Management turned on, and you do have to work
harder if you start from a raw scan without a custom profile. However, Nikon
Colour Management is very heavy handed with the shadow areas (i.e. serious
clipping) - unacceptably so. It may be that Nikon supply duff profiles as
standard, but since you can't add your own custom profiles, even if this
were so it makes the whole system useless.

 

[Kennedy McEwen]

Hi

I'm relatively new to slide scanning and now have a Nikon Coolscan III
(LS-30) to mess around with.

<Snip>
Soft results and blue tint are sure signs of dirty optics but, as others
have said, the scanner only does 2700ppi, so you will get and
interpolated image at 3200ppi - not enough to give the poor results you
describe though.

Take a look at this website on how to disassemble the unit to access the
parts that need cleaning (the mirror and the outside surface of the
lens). The LS-2000 is almost the same as the LS-30 inside.
http://www.vad1.com/photo/dirty-scanner/ls2000-cleaning/

Do NOT disassemble the lens or move it or the entire barrel from its
mounted position!! You will spend ages trying to get it to focus at all
again and will probably never have the same range of focus once you do.

Also be careful because the mirror is front silvered and the surface is
easily damaged. If you cannot get it clean just with a blower then try
lens tissue and a couple of drops of iso-propyl alcohol.

So how are things in GU these days? Is the library still sliding down
the hill? ;-)

 

[Kennedy McEwen]

This statement has been made so many times it has become a "fact" ;-)

So, the other day, I tested Nikon CM on a slide (not Kodachrome; Fuji
Velvia actually) of very colorful wildflowers. I scanned with CM
turned on and with it turned off. And my results show that the colors
with it turned on are better and more true to both the slide than with
it turned off.

I'm using NikonScan 3.1.2 on a Macintosh. I'm curious if the claims
of poor CM are from those using other operating systems.

Nope, its nothing to do with the OS. I scan almost everything with NCM
on - and that includes all photographic images. The only time I turn it
off is for test purposes when I want a specific gamma curve or some
other *known* default condition.