Coolscan 8000 v. 9000?

[Alan Browne]

Okay, the hunt is still on but the 8000 is CAD$1200 cheaper than the
9000. (I guess Nikon Canada have a few in inventory to get rid of...)

So what should I expect to be different in actual performance? Any
skeltons in the closet with the 8000?

Does it use LED's or fluorescent?

(PS: Picked up the Hassy 500 C/M yesterday...)!!

Cheers,
Alan.

 

[degrub]

expect slower scan, since you have to use the single line CCD mode to
avoid mechanical resonance (banding)
Somewhat better KC scans.

Otherwise look at the Nikon website for the other differences.

 

[David J. Littleboy]

Okay, the hunt is still on but the 8000 is CAD$1200 cheaper than the 9000.
(I guess Nikon Canada have a few in inventory to get rid of...)

So what should I expect to be different in actual performance? Any
skeltons in the closet with the 8000?

The skeleton in the 8000 closet is that the DOF is near zero and you have to
have the film very flat to get a sharp corner to corner scan. But I'd guess
that the 9000 has the same skeleton. A 6x6 scan on the 8000 should be under
8 minutes or so.

Hmm. If those are new units from Nikon, that's a hard choice. (I'd avoid a
used or refurbished 8000, but I'm a wimp.)

It's a similar problem to the Minolta 5400/5400II: they are similar enough
that no one has both.

Does it use LED's or fluorescent?

As I understand it, both use 4 colored LEDs and no filters on the CCD. The
Nikon scanners expose all four channels for each pixel _without moving the
film_, so pixel channel registration is perfect.

(PS: Picked up the Hassy 500 C/M yesterday...)!!

Congrats. I really enjoyed using mine over the years I had it. The damn
thing fits one's hands so perfectly. In Adam's biography he says that it
became his camera of choice from the day an early prototype arrived on his
doorstep.

David J. Littleboy
Tokyo, Japan

 

[Alan Browne]

As I understand it, both use 4 colored LEDs and no filters on the CCD. The
Nikon scanners expose all four channels for each pixel _without moving the
film_, so pixel channel registration is perfect.

Just DL'd the manuals. The other skelton for the 8000 is that it is 14
bit/color depth, not 16. Real issue? Maybe not.

(3 LEDs, BTW)

Scan times for 6x9 begin around 3 minutes and go to 15.5 for ICE/ROC/GEM
scans. ... 7 minutes for ICE only. (8 bit depth).

Cheers,
Alan

 

[eastside]

I've owned both. The LS-9000 yields visibly better scans; has a much
improved ICE; is much faster; has a "fixed" version of the negative mode
scan. Both use LEDs.

According to Nikon the LS-9000 is alone among the current lineup of scanners
in that it has Digital ICE Professional, which implements better defect
correction. The defect correction is so good on the LS-9000 that one should
just leave it on.

Dane

 

[David J. Littleboy]

I've owned both.

Yahoo! Real info!

The LS-9000 yields visibly better scans; has a much improved ICE; is much
faster; has a "fixed" version of the negative mode scan. Both use LEDs.

According to Nikon the LS-9000 is alone among the current lineup of
scanners in that it has Digital ICE Professional, which implements better
defect correction. The defect correction is so good on the LS-9000 that
one should just leave it on.

Thanks! Sounds as though it's more worthwhile than I thought. I just wish it
weren't US$3000 in Tokyo. Sigh. (I think my 8000 needs another cleaning.
Sigh, again.)

David J. Littleboy
Tokyo, Japan

 

[Kennedy McEwen]

Okay, the hunt is still on but the 8000 is CAD$1200 cheaper than the
9000. (I guess Nikon Canada have a few in inventory to get rid of...)

So what should I expect to be different in actual performance?

The 9000 uses 16-bits per channel while the 8000 uses 14-bits.
This means the 9000 will scan deep blacks better, with more overhead.

The 9000 has Pro-ICE, while the 8000 has ICE-4.
This means that the 9000 can clean up any dirt on Kodachrome a bit
better than other ICE equipped scanners but, like all ICE equipped
scanners, it still has difficulty because of the emulsion structure.

Any skeltons in the closet with the 8000?

The main issue is a type of banding, apparently caused by a mechanical
resonance. The solution is to enable the "fine scan" mode, which
deselects 2 of the 3 CCD lines. This eliminates any banding, giving
perfect results, but at the expense of each scan taking 3 times longer
than the advertised time. If you scan with ICE, autoexposure and
autofocus enabled, this doesn't make much difference to the overall scan
time, since it doesn't affect the time for these options.

Does it use LED's

LED's what?

or fluorescent?

Like all Nikons, it uses LEDs for illumination. The 9000 & 8000
scanners are equipped with a tri-linear CCD, but this is not colour
filtered. Instead of the conventional approach, the tri-linear CCD is
used to reduce the scan time by scanning three lines at once. As
mentioned above, the problem that beset the 8000 was that the larger
steps required by the scanner head to achieve this high speed scan set
up a mechanical resonance with many units, causing a discontinuity in
the scan position which appeared as a form of banding in the image. The
solution to the problem on the 8000 is to operate in the slower scan
mode using only one of the CCD lines.

 

[Don]

It's a similar problem to the Minolta 5400/5400II: they are similar enough
that no one has both.

Actually, they are *significantly* different! The former uses a
conventional light source while the latter uses a (white) LED. This is
then reflected in many other ways such as ICE, anti-grain method, etc.
5400II has much more in common with a Nikon scanner than with a 5400.

One possible reason nobody has both could be much more mundane and
nothing to do with specs. People who "jumped the gun" and invested in
5400 may not feel like throwing more money Minolta's way *shortly
thereafter* to get Mark II even though (on paper, at least) it would
appear to be a significant improvement.

As I understand it, both use 4 colored LEDs and no filters on the CCD. The
Nikon scanners expose all four channels for each pixel _without moving the
film_, so pixel channel registration is perfect.

Actually, it's *not*! A simple test is to scan B&W film in color -
which is how I came across it. Below you'll find a couple of thoughts
Kennedy had about the subject after I asked.

In brief, things working against perfect channel alignment are:
- residual scanner head motion
- finite misalignment of LED sources on the optic axis.

Don.

 

[eastside]

I've owned both.

Yahoo! Real info!

Speaking of which you can get more info in the archives of
http://groups.yahoo.com/group/coolscan8000-9000/

Thanks! Sounds as though it's more worthwhile than I thought. I just wish
it weren't US$3000 in Tokyo. Sigh. (I think my 8000 needs another
cleaning. Sigh, again.)

David J. Littleboy

One other thing: the banding problem that the LS-8000 had is fixed on the
LS-9000. So if you've been scanning in SF mode on the LS-8000 the effective
gain in speed is huge.

Dane

 

[Alan Browne]

Okay, the hunt is still on but the 8000 is CAD$1200 cheaper than the
9000. (I guess Nikon Canada have a few in inventory to get rid of...)

So what should I expect to be different in actual performance? Any
skeltons in the closet with the 8000?

Does it use LED's or fluorescent?

(PS: Picked up the Hassy 500 C/M yesterday...)!!

Cheers,
Alan.

Thanks all for great replies.

Conclusion: it sounds too good because it is... back to the 9000...

Cheers,
Alan