scott_brim

A note for Les Sarile: I've taken a close look at your scanned photos and have reached the conclusion that you were fortunate enough to buy one of the few CoolScan 5000 units that didn't suffer from flaring issues.

 

The use of an optically-inferior material for the protective window that covers the CCD is the most plausible source of the flaring problem. If we took your Coolscan 5000 apart and compared the CCD in your unit with the one in mine, I wonder what differences we might find? My unit was purchased in late January, 2004.

 

If it weren't for the flaring problem, I do believe that using DEE in combination with Analog Gain would be a more useful means of retrieving shadow detail from contrasty Kodachrome slides. Noise might still be present in the shadows, but selective use of a noise-reduction Photoshop plug-in could deal with that problem.

I second what J. Harrington says about the Nikon 5000 ED scanner. For myself, I can't afford an Imacon 848 or 949, so I have to make do with the Nikon. There are issues with this scanner, flare and noise being the most important issues as they affect Kodachrome. On the other hand, it does a very good job on E6 films and on color negatives.

 

I have roughly 500 Kodachromes that have marketable historical and/or artistic value among the thousands of slides in my collection. In my first pass-through of scanning these 500 slides --- starting in January 2004 --- I used the time to learn the scanner, to learn Photoshop, and to judge the quirks of each individual image in light of the experience I was gaining. I am now in the process of rescanning every single one of those 500 images based on two years of previous experience working with the Nikon and with Photoshop.

 

First, I remove each slide from its original cardboard mount and put it into a plastic frame I made myself that allows full scanner coverage of the entire frame. Using manual focus is the next step, doing a survey across the frame to get the best compromise focus point. I then take at least four 16-bit scans of the same slide using various settings, sometimes five or six scans.

 

For the first four basic scans, I'll scan with ICE and without, and with Scan Image Enhancer and without. I compare the four scans and make a judgement as to whether image degradation caused by ICE is sufficient to warrant manual cleaning in Photoshop, and as to whether the scan-enhanced image is more acceptable than the non-enhanced version.

 

If none of the first four scans turn out to be an acceptable base image for further work in Photoshop, then I choose the best of the four and start tweaking exposure of the next series of scans so as to bring out detail while minimizing flare. However, I keep all of the previous scans in my archive -- just in case.

 

In some cases, with difficult contrasty slides, I expose for highlight in one scan and for shadow in the other, and then merge the two together in Photoshop to create an acceptable first-generation TIF image, sometimes with judicious use of Neat Image to deal with noise.

 

ROC, GEM, and DEE provide assistance on some problem chromes, especially faded ones, but more often than not this is at the expense of enhanced noise in the dense areas. If you use ICE and DEE together combined with Analog Gain, the image will be significantly degraded, at least for enlargements greater than 8 x 12.

 

With the very worst of my problem Kodachrome slides, those being mostly Kodachrome 64's as opposed to Kodachrome 25 and Kodachrome II, I will just send them out for a drum scan -- if the image is valuable enough to justify this -- and save myself the time and trouble of getting that first acceptable base image.

 

If the Minolta 5400's had been reliable pieces of equipment out of the box, those might have been a better choice for what I'm doing. In spite of my reservations about it, I still believe the Nikon to be the best choice overall for what I'm doing, knowing that I can't come close to affording an Imacon. I've made some 24" x 36" enlargements from Kodachrome II using this scanner that are eye-poppers. But you have to know the tools to do this.

... Scott, I'd really like to offer CCD's with the glass removed, but that is way out of my league .....

 

It's way out of my league, too, but I do wonder what it would cost Nikon to have a specialized version of this CCD manufactured without the window --- or with a better quality of optical surface if the window has to stay there --- and what the difference in price would be to the consumer. $20? $40?

Uhhh... Erik ... Might you be thinking of selling an after-market, upgraded version of this component? If you are, what might it cost? (It never hurts to ask. Really.)

For the work I'm doing, flare is the single most important problem that keeps the Nikon 5000 from being a truly professional-grade tool. The other issue is noise in the shadows of legacy Kodachrome 64 and Kodachrome X.

 

The scanner does much better with properly exposed Kodachrome IIs and Kodachrome 25s -- even ICE works fairly well on these -- but flare can still be an issue depending on how much contrast is in the scene. And it can be an issue with the latest E-6 films too, if there is too much contrast.

 

I find it very strange that this problem has not been addressed by Nikon as an improvement to their basic design, because the mechanicals of the scanner appear to be very well engineered. If they fixed this problem and improved the noise performance somewhat, I would go out and get the new model straight away.

 

I will say this much too after reading the other thread: If someone offers an after-market fix to the current scanner that is shown to reduce or eliminate the flare issue, I'll by that fix and see what happens.

Sounds identical to mine. It sounds like a piece of heavy-duty machinery, and fortunately, it's built that way too. I've used my Coolscan 5000 weekly for two years now without a mechanical problem of any kind.

One other thing, in addition to what I said earlier: The context in which you perform your workflow makes a difference in your philosophy as to how you do things.

 

My context is that I have a series of roughly 300 photographs from 20 to 35 years old --- mostly 35mm Kodachrome II and Kodachrome 25 with some in Kodachrome X or 64 --- that have become very marketable as prints in the last five years.

 

I'm painstakingly restoring these photographs for use in a digital workflow with the intention of selling prints over the Internet and through just one retail outlet.

 

I think I can sell enough prints that there is a choice to be made between volume photograpic print reproduction on the one hand, and print-on-demand offset or laser on the other.

 

With CostCo charging only $1.50 for an 8x12 photographic print, and with color reproduction that is clearly superior to print-on-demand offset or laser; then for my situation, it makes sense to use CostCo's Noritsu printer as my volume production machine.

 

Making a Noritsu-compatible color proof is the last step in my image restoration process. The approach is to have a complete digital trail from scan to final print for each image I'm marketing, performed in sequence as follows:

 

--- Nikon 16bit TIF scan file (AdobeRGB). NikonScan works fine for me.

 

--- PSCS 16-bit TIF cropped, cleaned, color/contrast adjusted and sharpened base file (AdobeRGB); (May be more than one depending in the rendering purpose and how experimental I get. Useful experiments are not discarded.)

 

--- Various 16-bit TIF files in the Noritsu's color space for each format and size that I'm printing, adjusted as necessary for Noritsu-specific gamut, contrast, and color correction issues;

 

--- A corresponding 8-bit JPEG file for each of the 16-bit Noritsu-adjusted TIF files, plus a hardcopy proof print from the Noritsu for each image.

 

When I'm about to print a run of fifty (or whatever), I submit a test print in the morning, make sure it is OK in comparison with the proof print done weeks or months earlier. If it is, then I submit the order for the full production run.

 

In the two months that I've been using the local CostCo's Noritsu printer, I have yet to experience any detectable variation from my original proof print. I attribute this consistency to the local CostCo staff's professionalism in maintaining and servicing their machine and in understanding what it is that I'm attempting to accomplish.

 

I'm in a funding-constrained enviroment here, so I had to make choices as to what were the most important tools in making this project work:

 

--> The Nikon 5000 scanner. Works well on most of my images except for the Kodachrome X and 64. These take a lot of work but are less than 20% of my collection, fortunately. The Imacon 949 would be much better, but I can't afford it.

 

--> Two previously existing Gateway 700 XLs each with 1 gig of memory and 2.2 P4 processor and Gateway VX920 CRT 19" monitor. I need more memory and keep on acquiring more and more disk and offline storage capacity. My approach to color management forces the requirement for more digital storage space.

 

--> Adobe Photoshop CS. This project would be impossible without Photoshop. I am redoing some of my earlier efforts now that I'm more experienced with digital workflows.

 

--> QuickGamma and QuickMonitorProfile get me by well enough without requiring purchase of a calibration device such as One-eye or Spyder. The use of a capable CRT is probably why this works for me. If I am forced by industry trends to move to an LCD, then doing so probably means having to buy a hardware calibration device.

 

--> CostCo's staff and their Noritsu printer. They give me consistent reproduction quality at a very affordable price. I have near absolute control of the product's quality, and can do just enough proof printing to verify the previous workflow steps, but no more than is absolutely necessary.

 

The two most important decisions that have effected project costs have been to forgo a color calibration device for the monitor and to postpone the purchase of an inkjet printer such as the R1800. When I can afford it, I will eventually get a One-eye and an R1800 printer. Photoshop is worth every penny, and in fact represents amazing value for the money it costs. Events and experience have proven the Nikon 5000 to be the right choice for scanner, even if it has problems with legacy Kodachrome X and 64.

My experience at my local CostCo in using their Noritsu QSS-3412 printer and in employing profiles for this machine that I downloaded from Drycreek Photo has been very positive.

 

This result was been achieved with a modest investment in time and experimentation to see how well it would work. After a few minor adjustments to my monitor in making it a little less bright, I now get exactly what I want, and it works so well in terms of affordability and color quality that I have postponed a decision to buy an Epson R1800 or R2400.

 

I run Photoshop CS and use QuickGamma to adjust my VX920 CRT monitor -- four years old and still going strong, knock on wood --- checking and readjusting the monitor with QuickGamma at the start of every editing session.

 

I use Adobe 1998 for both the native scan on my Nikon 5000 and for the initial work in PS in creating a device-independent 16-bit TIF base image.

 

When I'm ready to print a particular image, I convert the Adobe 1998 base image to CostCo's custom Noritsu profile using relative colormetric and black point compensation. When proofing, I do not use "white paper" for proofing since doing so indicates less contrast than the Noritsu actually produces.

 

I make proofing adjustments to the 16-bit TIF image after it has been converted to the Noritsu profile and then save it as a separate base image for use when I need to print it again.

 

As a last step, I convert the image to an 8-bit JPEG, check it out for correct translation, and save that version as well. This is the file that actually goes to CostCo.

 

I burn the pictures I'm going to print that day to a CD and either hand it directly to the technician if he/she is not too busy, or else use the kiosk setup to get the files transferred into their system.

 

All the technicians at this particular CostCo understand color management theory and they know what it means, and why, when I ask for autocorrection to be turned off.

 

While I was standing at the counter a few weeks ago, a local professional photographer came by to pick up her workday's printing, and we had a discussion as to why she is using CostCo. She is one among several local professionals who now do the majority of their printing using this Noritsu machine.

 

She told me that the CostCo technicians at this store know digital photography and color management, they know their equipment and keep it running in top shape, and that they are positioning their store to handle volume printing requirements from the local professional community.

 

In particular, she saves money and time because she doesn't have to buy archival inks for an inkjet printer, she doesn't have to manage a chemical darkroom, and Photoshop gives her infinite flexibility to control her process from one end to the other. She also makes her own custom profiles for this machine every month for her own use.

 

For the very modest investment in time and experimentation that I've put into it, the results have been very successful.

 

OK, is the corollation between my monitor and a final print 100% spot-on in terms of saying they always look exactly identical to each other?

 

No, not quite, but the differences between the two are always consistant. So I close the gap by mental estimation, using my experience with past prints of similar color content and contrast to decide which specific contrast and color corrections should be made.

 

Maybe it is not as precise a color management scheme as a dedicated professional photographer might use, but I'm getting professional quality 8 x 12 prints for $1.50 apiece that could hardly be any closer to my expectations.

" .... The Coolscan 5000 needs improvement in only one major area in my opinion: and it isn't a big improvement that is necessary; noise vs. signal i n the dark areas ...."

 

In addition to a reduction in noise, the problem with flare in high contrast images needs to be addressed in the next model of the Coolscan series, whenever that happens. This means using higher quality CCDs -- probably somewhat more expensive, but well worth it I think.

 

I'd also like 5000 dpi and a higher DMAX if I could get it, but solving the noise and the flare problems are the two most important improvements that could be made in the current Coolscan 5000.

There was a discussion thread on this same problem in the summer of 2004, I believe.

 

I've had a similar problem that has never been resolved -- you get one run of GEM/ROC, but on the very next run, the post processing error occurs just as you have described it.

 

Closing the application and restarting gives you another GEM run. But then the next one fails. On my box, this pattern is absolutely consistent regardless of the film type chosen.

 

One possible cause that was mentioned in 2004 was having more than one USB device attached to the computer. IIRC, there we several other possible causes and cures mentioned. These solutions worked for some people, but none had any effect on my box.

 

As I've gotten more into digital scanning, I find that I rarely have a need to use the GEM/ROC feature.

" .... How large can I print from these files without seeing pixels or getting an unsharp or blurred image? ...."

 

Using my Coolscan 5000, I scanned a Kodachrome II slide from the early 1970's and took the initial 16-bit TIFF output file to an expert in a shop with a large Colorspan inkjet printer.

 

Using Photoshop CS and Photokit Sharpener, he resampled the image up to 24" x 36" with exceptional sharpness and detail. We then used Laminax gloss film over the paper, which caused the whole image to pop, so to speak.

 

Most everyone who sees the enlargement can't believe it came from a 35mm frame, or that I used a desktop scanner to capture the initial TIFF file.

 

Naturally, not having the experience to really accomplish this work myself, having an expert do the work made all the difference.

After you turn it off, let it cool for a couple of minutes, and then put it in plastic bag. I use an old Circuit City shopping bag, and it works just fine for keeping the dust out of the unit.

 

I find that for getting the best out of a critical slide, use manual focus to choose a good compromise focus point.

There was one film jam with the SA-21 in 1-1/2 years while scanning hundreds of strips. I cleared the jam with no damage to the film or to the adapter.

I've had the same experience as Les Sarile with my Coolscan 5000. ICE softens the image too much. For work where I'm scanning a decades-old 35mm Kodachrome 25 slide, and I'm trying to get as much clear detail as possible out of it for enlargement to 24 inches by 36 inches, I just bite the bullet and take the time needed to clean it properly in Photoshop. For my particular purposes, the results have been well worth the effort.

With my Cooolscan 5000 and contrasty Kodachromes using 16x sampling, there is no visible improvement over 1x in reducing noise levels within dense areas of a slide.

The kind of flaring exhibited in the Nikon 8000 scan is very similar to what occurs in my Nikon 5000 with similar subjects, especially if the media is Kodachrome. In my experience, the problem doesn't occur as noticeably with negative films, but it is still present to some extent.

 

After reading previous discussion threads about this problem, I'm fairly convinced that the "CCD window" is the root cause of it.

I've experienced exacty this problem with my Coolscan 5000, as have others who post to this board. I've not been able to solve it. One suggested remedy from Nikon was to disconnect all other USB devices from the computer except the scanner. That has apparently worked for some people. But alas, not for me.

I've read enough posts from people who have problems with flare and with digital noise using their Coolscans -- exactly the same issues I have seen in my own CS 5000 -- to know that there are definitely technical problems with the design and/or manufacture of these Nikon scanners. These problems are especially apparent when scanning Kodachrome.

My Nikon 5000 had the streaking/flaring problem right out of the box. I suspected then -- and now see confirmation of my suspicion from the technical experts --- that the problem is with the CCD's construction, i.e. that it has a window that covers the sensor and interferes with the CCD's proper operation.

 

Technical issues like these are what makes the Nikon less of a professional-grade tool than it otherwise might be, the other issue being digital noise when scanning Kodachrome. I now have to wonder if the CCD window plays some role, direct or indirect, in the digital noise issue.

The trial version of GF that I recently downloaded was only slightly better at interpolation than Photoshop CS, but on my XP box, GF took possibly fifty times as long as CS to accomplish the same job.

A few weeks ago, I added another USB device to my computer and moved the Coolscan 5000's cord to a different port. Upon doing so, the scanning speed slowed dramatically and the scanner's motor sounded as if it was working at quarter speed. It was just about as loud, but it was almost as if you could hear each individual pixel being shot. There was no difference in the quality of the scans. When I moved the cord back to its usual port, everything went back to normal.

For all my griping about about my Coolscan's problems with Kodachrome and the extra work it causes, scans of well-exposed Kodachrome 25s taken in the late 1970s and early 1980s are breathtaking in their overall sharpness and clarity -- especially if you are willing to forgo ICE, clean the slide thoroughly before the scan, and clean up any remaining dust in Photoshop, thus avoiding ICE artifacts and ICE image softening.

 

In general, I've noticed that my historical Kodachrome IIs and Kodachrome 25s scan easier and with much less noise than do either the current or the previous Kodachrome 64s. Had I been able to look into a crystal ball in 1978 and see what the future held, I'd have used Kodachrome 25 in preference to Kodachrome 64 for almost all of my outdoor shooting, as opposed to mixing and matching subjects with film.

 

I've since dropped Kodachrome altogether for current work because it just doesn't scan well enough for current projects when using a desktop scanner. However, if Kodachrome 25 was still available, that decision probably wouldn't have been nearly so easy to make.

I went through a phase early-on where I did everything the same way thinking there was some magic formula for the 5000 that would work most of the time for most of my images.

 

But I now think there is much to be said for taking the time and the trouble to immerse oneself in the nuts and bolts of digital color theory and practice, and for learning how the tools can be used individually and together to create an acceptable (or more than acceptable) final result.

 

Here is my general approach as applied to (1) my historical Kodachromes, and (2) EliteChrome 100 and Fuji Super HQ 100.

 

 

(1) Historical Kodachromes:

 

When working with my historical Kodachromes where I want the best image that I can get out of them, I generally rely on a combination of experimental scanner settings, post-scan processing with Photoshop, and judicious use of Neat Image when necessary.

 

For the most difficult Kodachromes, sometimes it takes as many as three experimental trial scans and post-processings in Photoshop and Neat Image before I get a final result that's acceptable in terms of sharpness, contrast, color balance, and minimized noise.

 

For most of these chromes, I don't use ICE, DEE, GEM or multi-sampling at all. I use ROC once in awhile, sometimes SIE depending on the slide, and play with Analog Gain and the brightness control to balance the recovery of shadow detail against noise --- knowing that I want to minimize the application of Neat Image later on to as little as possible for any individual slide.

 

For the Kodachromes, I always use manual focus. All changes to curves and contrast are done in PS. I clean up the dust in PS too, even though it is time-consuming and a pain. Why? Because ICE, especially in combination with DEE and too much Analog Gain, softens Kodachrome images substantially and seems to encourage both flaring and bleedover from contrasty areas.

 

(2) EliteChrome 100 slide film and Fuji Super HQ 100 print film:

 

Here I normally use ICE and sometimes SEI. The 5000 does a remarkable job of lifting the correct color balance and contrast off of these two films without too much intervention, reducing the amount of post-processing considerably in comparison with what is needed for Kodachrome. At least for my purposes anyway.

The next important question to be answered is this: how well does the new 5400 II perform with Kodachrome --- better, worse, or about the same as its predecessor?

 

My Nikon Coolscan 5000 does very well with the latest E-6 and color negative films in the ISO 100 range, but my historical Kodachromes require a lot of tweaking and post-scanning cleanup work.

 

If the new 5400 II performs as well or better with Kodachrome as does the previous model, then it is time to add a Minolta scanner to complement my Nikon. (And to complement my collection of 1960s Minolta SRT-101 cameras, too. Which are still my workhorse cameras.)

Thanks Will. What I think you are saying essentially is that for professional prepress work, better results are obtained if the graphic arts specialist has complete control over the process from beginning to end, imaging through separation.

 

My Nikon 5000 does an excellent job for my purposes with E-6 and negative films. But this does not include prepress. Contrasty historical Kodachromes are the issue. Neat Image helps, but some of these chromes still suffer from bleedover and flare issues.

 

Some of them are damaged as well and I've had to spend some bit of time with them in Photoshop to clean them up, even after ICE was done with them.