Some Qs about Nikon Coolscan 8000 ED

[brbo]

Hi!

Currently I have a Microtek 120tf (4000dpi medium format scanner) and a Minolta DiMage 5400 v1 (5400dpi small format scanner). Now, I'm very happy with my scanners. Both are probably at least as good as Nikon counterparts (Minolta is probably even noticeably better).

If I was a BW shooter I wouldn't even think of replacing those two scanners, but I use colour film 95% of the time and I must admit that sometimes I miss the lack of ICE in my Microtek 120tf. So, I've been thinking that I could probably get a Nikon Coolscan 8000 ED if I sold my two scanners.

I've been reading about CS 8000 and I would like some input from people who've used it (or even have some first hand knowledge of how this Nikon compares to my scanners).

So far, I gather that:

a) Nikon would produce somewhat harsher/grainier files (LED light source vs. cold cathode (additionally, my Minolta even has a built-in grain dissolver)); what is your opinion on that?

b) CS 8000 ED needs "fine mode" enabled otherwise you get banding; how slow does that make it? I always scan at full resolution and with 16x multi-pass (but don't use multi-exposure), my both scanners are fairly slow at that settings, so much slower with Nikon in "fine mode" would probably not be welcomed.

c) Holders. I must use glass holders with ANR glass for medium format. Nikon has two glass holders. The cheaper one doesn't seem to have ANR glass. Am I wrong? Is the glass easily replaceable with ANR glass? Does the more expensive, rotating holder, have ANR glass for both glass plates?

d) I scan panoramic format on 135 film. I scan it in my Microtek in a glass holder. I don't need any film masks or whatever, I just put my film between two glass plates and scan the whole area with sprocket holes included. I've heard that you need film masks on Nikon otherwise you get "light leaking" from the unmasked sides into the film area. If that is so, would that mean that scanning film with sprocket holes included is not possible with Nikon? Like this...






e) Is there any way to scan into the unexposed edge of 120 film on Nikon 8000? My Microtek 120tf can get just a mm or two into the edge, I wish it could scan the whole width of the film.

f) Any other thoughts, opinions and comments are welcome.


Thanks!

 

[mfogiel]

I have the 9000 version, but I think it is the same as 8000.
Re:
a) - no. on the contrary
b) - it makes it about twice slower, but I don't know your scanner
c) - I use the non rotating glass and have never seen newton rings, but I scan B&W, for colour you can use special masks, that mitigate the issue
d) - you can do the same, you can also use the 35mm glasss holder after cutting away the dividers
e) - no
f) - read this thread and look at Fernando's tests
http://www.rangefinderforum.com/forums/showthread.php?t=138307

 

[sevo]

The non-rotating and rotating glass holders both are AN glass top, polished bottom.

While scanning with sprocket holes is possible, it is not advisable. It will reduce the dynamic range, unless you overexpose the holes, where they will spill into the image.

 

[Fernando2]

Hello,

first of all: I did a full review of the Nikon CS8000.
In Italian, but with tables and graphics and image samples; plus I can translate parts for you.

http://www.effeunoequattro.net/htdocs/freecontent/FC_ProvaNikon8000/index.php

Then:

a) Nikon would produce somewhat harsher/grainier files (LED light source vs. cold cathode (additionally, my Minolta even has a built-in grain dissolver)); what is your opinion on that?

The Nikon produces scans a bit grainier than the Microtek 120TF (which I had before the Nikon), but less grainy than the Minolta 5400-I (which I still have) even when you enable the GD on the Minolta.
The Nikon has almost no aliasing (so almost no grain aliasing).
I really would not categorize the 8000/9000 as "grainy". Unless you compare it to flatbeds (which lack the resolving power). 😉

b) CS 8000 ED needs "fine mode" enabled otherwise you get banding; how slow does that make it? I always scan at full resolution and with 16x multi-pass (but don't use multi-exposure), my both scanners are fairly slow at that settings, so much slower with Nikon in "fine mode" would probably not be welcomed.

The timings for the Nikon are in my review; I recommend you read the table and compare to the timing table of the Minolta 5400 in my other review:

http://www.effeunoequattro.net/htdocs/freecontent/FC_minolta_elite/cinque.htm

Very briefly: even with Fine Mode enabled, the Nikon 8000 still outperforms the Minolta 5400, if you enable the GD on the Minolta. Same if you enable ICE and multisampling on both.
But to have more details, I recommend looking at my tables.

BTW I would not recommend going 16x on either scanner: I've seen that the heat buildup tends to curve the film, so you end up with a different focus than measured at start.
4x is a good compromise from my dynamic range measurements.

Can't comment on the glass holder (don't have it), nor on panoramic scans (never did it), but:

I've heard that you need film masks on Nikon otherwise you get "light leaking" from the unmasked sides into the film area. If that is so, would that mean that scanning film with sprocket holes included is not possible with Nikon?

Indeed the Nikon 8000 (not the 9000) suffers from light leaking.
Or, better: it suffers from internal parassite reflections which show up as "light leaks", or flare (they improved the lens coating in the 9000).

Like this...

You would end up with some halos around the sproket holes.
Not too severe (probably not into the actual picture), but visible.

e) Is there any way to scan into the unexposed edge of 120 film on Nikon 8000? My Microtek 120tf can get just a mm or two into the edge, I wish it could scan the whole width of the film.

Not really. You can catch half a mm of unexposed edge on both sides, but that's it.

f) Any other thoughts, opinions and comments are welcome.

The Nikon 8000 is a great scanner, with some shortcomings.
It does not reach the stellar resolving power of the 5400, which was absolutely an unmatched, unique beast in the consumer world (only drum scanners, Creos and Imacons do better).
It has some flare issues (not really problematic in real world scans).
It has serious depth-of-field issues; this is the most problematic part. Curly films need glass holders or focus-stacking.
I routinely focus-stack my Coolscan scans (my Minolta scans, too) using HeliconFocus.

Fernando

 

[brbo]

Thank you all!

Seems there is little sense in replacing Microtek and Minolta with Nikon. I would only gain ICE on medium format (and ability to use Vuescan with Nikon 8000 which I can't with Microtek as it doesn't work well enough with it), but lose on all other aspects.

Fernando, did you make a glass holder for your Minolta 5400? I bought a spare holder for my Minolta with the intention of modifying it to take a piece of ANR glass. It would be great if you could share your design or findings about the mod (if you've done it, of course).

Edit: Arhg, I don't know why I assumed that you had a glass holder for your Minolta. I re-read your post and find no mention of such a thing. Sorry.

 

[Fernando2]

Still, I was actually hoping to build a glass holder for the Minolta 5400! 😀
I even purchased glass inserts from ScanScience. But then I gave up: my purpose was to build a glass holder for wetmounting, and I did not find a way to prevent fluid leaking into the scanner.
Plus I don't think there's enough space to use glass on both sides. Thin glass (for microscope observations) breaks too easily (tried that with the Microtek).

About replacing both your scanners with a Coolscan: the only real reason why I would do that, is for longevity.
They no longer sell CCFL lamps for those 2 models.
Coolscans use LEDs, they last much longer and don't lose efficiency during their life.

Fernando

 

[brbo]

I've got a spare lamp for Minolta. I guess I'll be using those two scanners until they die. I hope that is still a long time away...

I'll only use one glass insert for my glass holder. I hope it will still be an improvement over the stock holders.

 

[Fernando2]

I've got a spare lamp for Minolta. I guess I'll be using those two scanners until they die. I hope that is still a long time away...

I'll only use one glass insert for my glass holder. I hope it will still be an improvement over the stock holders.

Excellent!
You did a very clever thing, purchasing a spare lamp.
May I ask you the maker and model of the lamp?
Maybe it's possible to source a similar model, somewhere (even if I doubt it 🙁 ).

 

[brbo]

Excellent!
You did a very clever thing, purchasing a spare lamp.
May I ask you the maker and model of the lamp?
Maybe it's possible to source a similar model, somewhere (even if I doubt it 🙁 ).

Fernando, I'll try to give you any detail I can about the lamp when I get home. There was a guy selling those lamps a year or two ago on eBay for a while (with auctions, 1 lamp per week). I never checked if the lamp does indeed work with Minolta 5400.

 

[brbo]

Sadly, no markings on the lamp...

 

[rbrooks]

I routinely focus-stack my Coolscan scans (my Minolta scans, too) using HeliconFocus.

I tried this many years ago but I had trouble. In the meantime I'm sure the software and hardware have improved. Can you share your current technique with us?

 

[Fernando2]

Sadly, no markings on the lamp...

I suspected it. 🙁
Manifacturers used to ask "custom designed" lamps to Osram, Philips etc., without known specifications, and then add beefy margins (say $125 instead of $25).

Thanks anyway!

Fernando

 

[Fernando2]

I tried this many years ago but I had trouble. In the meantime I'm sure the software and hardware have improved. Can you share your current technique with us?

Sure!
I use Vuescan for this, but maybe it works with NikonScan as well (can't remember if it allows manual focus)

The process is a bit boring; it would be vastly better if only Ed Hamrick (author of Vuescan) would agree to implement some automatic procedures I asked him for ages. 😛

1) I measure the focus value on various spots of the frame. Most critical zones of the image and/or borders, center. Say 3 to 8 measures.

2) I take the min and max focus value. Say it's 0.400 and 0.600

3) The useable DOF of the CS8000 is about 0.035-0.045. So I calculate the focus steps I need to stay within the DOF. In our example: 0.440, 0.480, 0.520, 0.560

4) I do the scans with manual focus; focus values as above

5) Focus stacking into HeliconFocus (need the 64 bit version for 120: it uses a huge memory chunk). I use method "B" with Radius = 10-30 and Smoothing = 4-8

Fernando

 

[rbrooks]

The process is a bit boring; it would be vastly better if only Ed Hamrick (author of Vuescan) would agree to implement some automatic procedures I asked him for ages. 😛

1) I measure the focus value on various spots of the frame. Most critical zones of the image and/or borders, center. Say 3 to 8 measures.

2) I take the min and max focus value. Say it's 0.400 and 0.600

3) The useable DOF of the CS8000 is about 0.035-0.045. So I calculate the focus steps I need to stay within the DOF. In our example: 0.440, 0.480, 0.520, 0.560

4) I do the scans with manual focus; focus values as above

5) Focus stacking into HeliconFocus (need the 64 bit version for 120: it uses a huge memory chunk). I use method "B" with Radius = 10-30 and Smoothing = 4-8

Have you tried with less focus measures, i.e. only 2 instead of the 3 to 8 you mention? I ask because my experience (years ago before Heliconfocus) has been that if I set one focus point just a bit off centre that gets everything but the edges. The edges would then need another scan at a focus point near the corner (which results in a slightly different focal length so one cannot easily stitch). The centre and edge scan might be enough? I made some scans years ago before the newer Helicon and Photoshops. I'm betting that I'd get a decent composite scan with just the centre + edge. Do let us know if you've tried that solution and if so then why you've rejected it for the 3 to 8 focal points.

Guys, if software has come along like it seems it has, this is a game changer. The main problem with all these film scanners has always been keeping the negative flat. If there's an easy way to multi-scan for focus point and then have software such as Helicon stitch them back together that's a big deal.

I should have some time over the coming holidays to do some tests.

 

[sevo]

I suspected it. 🙁
Manifacturers used to ask "custom designed" lamps to Osram, Philips etc., without known specifications, and then add beefy margins (say $125 instead of $25).

The main issue is that CCFL lamps were considered semi-permanent (having a MTBF longer than the five years expected use life of professionally owned computer equipment) and their length was one of the more easily variable parameters in a scanner design. As suppliers could cut relatively smally batches (of, say, 100) to length, the makers did not feel that they had to standardize for the benefit of their service department either.

The electrical parameters don't vary critically much with size - where possible, it will often be a solution to ream out a bit more of the lamp case to fit the next bigger readily available lamp (shorter lamps, while being installable without modification, tend to deliver noticeably bad margin illumination).

 

[philipus]

About fine mode - my 9000 most definitely does not take 2x as long with fine mode enabled (for 35mm). I barely notice any difference actually.

 

[Fernando2]

Have you tried with less focus measures, i.e. only 2 instead of the 3 to 8 you mention?

Yes, it may work if film is very flat to begin with; but it does not work in many cases (film bulges in the strangest ways! 😀 )
Anyway, it does not take much time to do 5-6 measures.

What takes a long time, is doing all the needed scans! 🙁

That's why I prayed Ed Hamrick to implement an automatic multi-scan function into Vuescan, when one inputs min focus/max focus/focus step and Vuescan does all the work.
But probably I was the only one asking for it.
Maybe if other Vuescan users join the party and mail him about the same thing... :angel:

Guys, if software has come along like it seems it has, this is a game changer

It is, I assure you.
It works very well, I use it as my default workflow since at least 2 years.
The very important part: you have to load HF with ALL the steps it needs.
Don't skimp on focus step. It needs a *complete* stack of scans with progressive focus, without any "disjointed" scan.
Then it does a tremendous job.

Fernando

 

[rbrooks]

That's why I prayed Ed Hamrick to implement an automatic multi-scan function into Vuescan, when one inputs min focus/max focus/focus step and Vuescan does all the work.
But probably I was the only one asking for it.
Maybe if other Vuescan users join the party and mail him about the same thing...

Ok, I've cleared some time where I can experiment with scans at different focus settings. I want to see how many scans are needed to combine together a composite for Helicon focus that (a) doesn't have artefacts (b) is better than a straight scan.

If this works well, let's write a joint email to Vuescan to ask him to automate this.

 

[rbrooks]

Workflow:

1) Use manual focus to get a feel for the range of focus of your negative(s). My negatives are relatively flat but still show OOF edges.

2) Save scans for each of the key focus points using a naming convention such as adding the manual focus offset used. I've found that with my negatives 2 or 3 focus points worked better than 8. 8 confuses the stitching software and didn't add too much value when I looked at the maps in Photoshop.

3) Bring the TIFFs into Photoshop and do the Automate Photomerge and Auto-Blend Layers "trick".

Photoshop seems easier to use than Heliconfocus.

The pictures attached shows each of the 3 layers I choose for the final stitch.

 

[rbrooks]

Here is the final stitched image.