Pages: [1]   Go Down

Author Topic: Some scanning issues  (Read 16518 times)

Eric Myrvaagnes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 22814
  • http://myrvaagnes.com
    • http://myrvaagnes.com
Some scanning issues
« on: July 09, 2005, 04:20:20 am »

Ray,

Thanks for the detailed comparison. Having an Epson 2450 scanner, I have been wondering whether to update to a 4990 or try to justify spending the extra for something like the Nikon in order to get decent scans of my tons of old slides. Your report has helped me save a good bit of cash!

I'm sorry about the slide-mount cropping problem, er, "issue". My sense is that for most of my slides, the amount of cropping done by the mount is roughly equivalent to the cropping done by the camera viewfinder, so it usually isn't an issue (problem?) unless the slide has been shifted in the mount. So most of the time it's, as they say, "no issueo" ("no problemo"?).    

Eric
Logged
-Eric Myrvaagnes (visit my website: http://myrvaagnes.com)

Eric Myrvaagnes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 22814
  • http://myrvaagnes.com
    • http://myrvaagnes.com
Some scanning issues
« Reply #1 on: July 11, 2005, 12:05:40 am »

Quote
Unless your slides are mainly low quality snap shots (technically), I'd have to recommend a dedicated slide & negative scanner such as the Nikon 5000.
Ray,

Thanks again for the clarification, I guess, even though it means spending more money. I guess you're saying there's no such thing a s a free lunch.   

Eric
Logged
-Eric Myrvaagnes (visit my website: http://myrvaagnes.com)

Eric Myrvaagnes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 22814
  • http://myrvaagnes.com
    • http://myrvaagnes.com
Some scanning issues
« Reply #2 on: July 11, 2005, 10:12:31 am »

Very true. My trusty 2450 does a half-way decent job on 4x5 negs and on MF (unless the DR is too high), and marginally ok on some 35s as long as the eventual print size is quite modest (on a few I have gotten away with prints to 10x15"). I am still wondering how much improvement I can expect from a 4990. One plus for the 4990 is that I'll be able to scan my 8x10 negs on it, which I can't do now. I'm reluctant to shell out for a film scanner that will only do 35mm, and the ones that do MF as well are out of my current budget (unless somebody offers to buy my excellent Pentax 67II kit with five good lenses).

your last statment, "Scanning 35mm on the Epson is like using a 4x5 format camera with a 35mm back," really speaks to me.

Eric
Logged
-Eric Myrvaagnes (visit my website: http://myrvaagnes.com)

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Some scanning issues
« Reply #3 on: July 11, 2005, 11:30:09 pm »

If you have a copy of Harald Johnson's "Mastering Digital Printing Second Edition", you may be interested in the material on scanning resolution in Chapter 3 up to page 92.

I am using a Minolta Dimage Scan Elite 5400 operated with Silverfast Ai for scanning colour negatives, and find the results (colour, detail, luminosity range) very satisfactory. But this scanner is only useful for 35mm film. It uses glassless carreirs for both negatives and slides. Both the specs and the performance of the Minolta scanner are impressive with the right software. I bought Silverfast after I found both Minolta's own software (Version 1.1.2) and Vuescan to be rather unsatisfactory for producing balanced colour and adequate contrast from colour negatives, though I hear Minolta's latest software on the Scan Elite 5400 II is better with negatives than was version 1.1.2. Silverfast isn't cheap or intuitive (at first), but it saves mountains of time on post-scanning image adjustment once you get comfortable using it. Minolta's version of ICE does clean-up old cruddy Kodachromes eventhough they say in the instruction book that it doesn't (Johnson's book, page 89, bottom box).
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Some scanning issues
« Reply #4 on: July 12, 2005, 09:20:37 am »

Ray,

The Minolta scanner is an excellent product. I seldom need to use 5400 dpi but good to have it in case. It's ability to reproduce detail from shadows to highlights is very impressive.

I don't know what an analogue gain facility is, but with Minolta and Silverfast software you have extensive control over the histogram by exposure adjustments for either aggregate RGB or channel by channel. There is white point and black point adjustment as well.

For handling images with a very wide luminosity range, one well-done scan usually has enough information for successful luminosity blending in Photoshop.

Both Silverfast and Minolta software give you multi-sample scanning capability to reduce random noise. I scan at 4X.

It is very difficult to reliably profile a scanner for negative film. Best is to rely on the film specification and film type adjustment capabilities in the software and here is where Silverfast shines above all the others.

I find the image matching between scan preview and photoshop quite close in RGB color space and less satisfactory in ProPhoto color space. I don't aim for perfection at the scanning stage, because Photoshop is better than all other applications at fine-tuning everything. At the scanning stage I want something close enough to the final result that the adjustments in Photoshop are not major, unless the image has huge luminsoty range requiring blending; then we are into more extensive Photoshop as mentioned above.

Don't buy Silverfast until you know for sure which scanner you will be using it with, because Silverfast, unlike VueScan, is scanner-specific. If you want a cheap sneak preview of what Silverfast can do, check out Ian Lyon's tutorials on the LaserSoft website, try the demo, and even go as far as buying Taz Tally's book from Amazon.com - "Silverfast The Official Guide" - for about 20 dollars you'll see in detail whether it is worth investing the $258 for the whole product.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Some scanning issues
« Reply #5 on: July 12, 2005, 11:06:54 pm »

Quote
Don't buy Silverfast until you know for sure which scanner you will be using it with, because Silverfast, unlike VueScan, is scanner-specific. .
Mark,
I was very pleased that I could also use my Vuescan software, which I'd earlier bought to use with my Nikon 8000ED, with the Epson 4990 flatbed. Vuescan has to be the best value scanning software available.

However, as always there's often no clear winner in all respects. Each software might have a particular feature which is better than another scanning program regardless of price. Nikon Scan which was bundled with my 8000ED has just 2 profile options for slides, Positive or Kodachrome.

I find that scanning Kodachromes with the Nikon software and its Kodachrome setting, produces slightly richer reds that are truer to the Kodachrome effect than Vuescan produces with its Kodachrome setting. On the other hand, Nikon's scratch removal (ICE) is not appropriate for Kodachromes. It produces a double edge on parts of the image, whereas Vuescan's infrared dust and scratch removal works perfectly.

The problem I have with Silverfast is that so far I've been unable to produce a scan that is fundamentally better than a scan I can produce with Vuescan and/or Nikon Scan, except with regard to B&W negatives. Silverfast's SRD seems about the best tool for removing dust and scratches on silver based B&W negatives, but it's not perfect by any means.

Silverfast's SRD is not as straight forward as ICE. One has to make adjustments with a preview which shows the effects of each adjustment. Having made the appropriate adjustments to remove all the visually identifiable sctratches, one can still end up unavoidably removing detail that is not a scratch.

A great new feature of PS CS2 is the 'spot healing brush'. This seems to work like the 'clone stamp' tool but without having to sample nearby textures. One just clicks on any defect and it miraculously disappears. I wonder if this is a better approach than using Silverfast's SRD.

Silverfast seems pretty miserly in their demo software. It's difficult to make any assessment if it's worth the money or not. The demo for my Nikon 8000ED, using my standard Kodachrome test slide, produces 16 bit files with 8 bit quality (ie. serious banding in the shadows similar to an 8 bit scan with Nikon Scan software).

The Silverfast SE version that was bundled with my Epson 4990 will scan in 48bit but will output files only in 24 bit. It doesn't offer the 'long pass scan' feature of Vuescan (digital blending which is more effective than mult-scanning), and doesn't even offer multi-scanning unless I upgrade to SE Plus.

Quite frankly, I get the impression that Silverfast is on a marketing binge to attract customers by offering a lot of bells and whistles, but no real improvement in scan quality.

If professionals prefer it, I can only assume it's because it speeds up their workflow in a fully calibrated system where scanning output goes straight to printer.
Logged

BernardLanguillier

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 13983
    • http://www.flickr.com/photos/bernardlanguillier/sets/
Some scanning issues
« Reply #6 on: July 13, 2005, 05:59:51 am »

You know what, I might buy a cheap Tachihara 8*10 just to check...

But I suspect that you are right, the 4990 should be perfect for that, but I am concerned by Newton rings... using some scanning fluid is probably the solution, if the sealing of the glass of the 4990 is good enough...

Cheers,
Bernard

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Some scanning issues
« Reply #7 on: July 13, 2005, 10:12:46 pm »

Quote
The only reason why I bought Silverfast is that after using Minolta's software and Vuescan I found Silverfast by far the best game in town for scanning COLOUR NEGATIVES

I'll do some more comparisons. I might eventually get to the stage where I'm ready to start scanning proper and work methodically through my collection  . I'm a bit concerned about future pressures to rescan as technology progresses and my expertise in scanning increases. I'm most concerned about spending a lot of time scanning scratchy B&W negatives. The silver that makes ICE ineffectual also bestows great longevity. Is it likely that sometime in the near future there'll be a fix that automatically removes scratches from B&W silver based film without causing blur?

Quote
Brushing the negs with a "Visible Dust" brush (the same one recommened on L-L for sensor cleaning) before scanning helps reduce this work to only the residual crud and imperfections that can't be blown or brushed away.

Have you tried wiping negatives with isopropanol first?
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Some scanning issues
« Reply #8 on: July 14, 2005, 01:24:22 am »

Quote
I'm afraid of damaging my negatives by applying chemical agents to them. Some damage make take a long time to reveal itself. Do you have long-term experience with this approach and is it better than using an electrically charged brush?
Ah! Good point! I wonder if there's a thread somewhere on the net that deals specifically with this issue. The chemical I use is simply called 'Film Cleaner' and the ingredients are 90% isopropanol. I don't know what the other 10% is. I've been using it, off and on, for around 8 years and I've recently rescanned slides  that would have been cleaned with this fluid around 8 years ago, which do not seem to have been detrimentally affected in any way. What more can I say!
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Some scanning issues
« Reply #9 on: July 14, 2005, 08:20:14 am »

Sounds to me like experience speaks for itself.  :D
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Some scanning issues
« Reply #10 on: July 09, 2005, 01:59:00 am »

My attention has recently returned to scanning some of the many old slides and negatives I have, archived in plastic and paper sleeves and Kodak slide mounts.

I was vaguely aware that slide mounts would crop a small margin around the frame, but I'd never taken a ruler to a Kodak slide mount and measured the opening. I'd had no reason to, until I got suspicious that some of the shots I'd taken many years ago, after scanning on my Nikon 8000ED, seemed to contravene even a basic sense (I won't use rules) of composition that I would have been aware of in those days, like not chopping off the tips of someone's fingers.

I was quite shocked to discover that all these old Kodak slide mounts reduce the 24x36mm format of 35mm to around 22x33mm. Well, perhaps that's not too bad if the slide is perfectly centred in the mount. One and a half mm off each end of the width and just 1mm off each end of the height might be no big deal. But, I presume the whole purpose of this cropping is to allow for error in the automatic mounting process. I have some slides that have been shifted as much as 3mm or more to one side in the mounting process.

I now feel obliged to remove every slide from its mount before scanning and carefully position it in the Nikon negative holder which doesn't appear to crop any part of the image (give or take 1/4 of an mm).

A preferrable alternative would be to replace the mount with another type of mount that doesn't crop the image. I haven't found such a mount yet.

Yet another alternative, is to use my Epson 4990 flatbed which comes with a negative holder which accommodates 4 strips of 6 negatives (or positives) with no divisions along the short edge, and the width is marginally greater than 24mm.

I'm still trying to determine if my Nikon 8000ED can produce superior results to this (apparently) extraordinary good value flatbed. The fact that I'm having difficulty in arriving at a definite conclusion in this regard might be interpreted by some that I simply don't know what I'm doing. And that might well be the case  :) .

The problem (oops!, issue) is that it's difficult to separate the performance of the hardware from the performance of the software. The Nikon software allows use of a 'scanner RGB profile' which I find produces better results than scanning 'into' any of the offered colour spaces (ARGB, Wide Gamut RGB etc). But, regardless of this factor, Nikonscan cannot use ICE satisfactorily with Kodachrome. There's a double edge. Vuescan, however, can use  its equivalent of ICE on Kodachrome, producing superior results.

Vuescan with the Epson 4990 is as bad as Nikonscan in this regard. Back to double edges or similar; in fact occasionally green edges. But Silverfast SE, which is packaged with the 4990, is superior to Vuescan in this respect and produces results, with the 'difficult' slides I've tried so far, that are very similar to what I achieved on the 8000ED. It's serious pixel-peeping to tell the difference in terms of resolution and shadow noise.

One major advantage of Vuescan is that it offers the option of digital blending. I guess most of us by now have heard of this technique of increasing the dynamic range of digital cameras by taking 2 or more different exposures of a scene and blending the results in a manner which extracts all the relevant detail from both (or all) shots.

Vuescan can do something similar when you 'photograph' your slide in your scanner. In fact, the scanner, I would have thought, lends itself to this technology. You have either a stationary target  and moving scan light, or a staionary scan light and moving target. In either case, the stage is ideal for double (or triple) passes of varying exposure.

Any concern that the Dmax of the 4990 might not be quite as great as that of the 8000ED becomes irrelevant. 4.00 is very impressive anyway.

Absolute resolution is not something I've had time to test conclusively yet, but I'm going to search for some shots I took a few years ago with my Minolta 50/1.4. So far, resolution seems on a par with the Nikon 8000ED, give or take a small margin. Since the Nikon 8000ED was a really expensive item for me, the purchase of which was difficult to justify at the time, you might think I should have a personal bias to demonstrating the superiority of the Nikon. Not at all  :) .
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Some scanning issues
« Reply #11 on: July 10, 2005, 07:55:27 pm »

Quote
I have been wondering whether to update to a 4990 or try to justify spending the extra for something like the Nikon in order to get decent scans of my tons of old slides. Your report has helped me save a good bit of cash!
Whoa! Hold your horses! This wasn't a comprehensive review of the Epson 4990, just some first impressions. I've now had time to scan a number of slides and negatives in both 35mm and MF formats. There's no getting away from the fact that this flatbed produces softer scans than the Nikon 8000. I would characterise the difference as being similar to the difference one would expect between a zoom lens and a prime lens.

There's also an issue with the deepest shadows in film such as Fuji Velvia, and negatives where the shadows are respresented (on the film) as almost complete transparency. The Epson tends to produces patches of total black with no detail whatsoever, whereas the Nikon seems able to reveal at least some detail in such areas, whether useful or not. However, these sorts of factors depend a lot on the skill of the operator and are not necessarily a problem. In any case, a print of any high contrast scene usually needs a few areas of solid black to look realistic.

Unless your slides are mainly low quality snap shots (technically), I'd have to recommend a dedicated slide & negative scanner such as the Nikon 5000.
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Some scanning issues
« Reply #12 on: July 11, 2005, 12:45:42 am »

Quote
I guess you're saying there's no such thing a s a free lunch. 
Well, not quite. As computer technology progresses, I think all of us can find examples of a recent product that is not only cheaper than an older product, but as good or even better.

Dmax and resolution are the two most significant specifications of a scanner that are published. These two specs seem to be far higher in modern scanners than they were just a few years ago and scanners certainly appear to be getting better as well as more affordable.

It just seems that the specs are exaggerated, or at best misleading.

I could make a perfectly true statement that two lenses, say, are able to resolve 100 lp/mm, but one lens could still be sharper than the other because it has a higher MTF (more contrast) at 100 lp/mm. I can only assume that the lens in the Epson 4990, fine as it may be, is resolution limited due to the larger area it's required to scan. Scanning 35mm on the Epson is like using a 4x5 format camera with a 35mm back.
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Some scanning issues
« Reply #13 on: July 11, 2005, 11:05:42 pm »

Quote
your last statment, "Scanning 35mm on the Epson is like using a 4x5 format camera with a 35mm back," really speaks to me.
Eric,
I've never used a 4x5 camera with or without a 35mm back so that statement is just speculation. The 4990 might be better than that. If one assumes Epson is applying the same degree of exaggeration to their specs across all their flatbeds (a big assumption) then the 4990 would have to be an improvement over the 2450.

Generally, the larger the format the lower the resolution. (In terms of lp/mm, in case anyone picks me up on that  :) ).

I see no reason to think that the 4990 would not be ideal for 4x5 and 8x10 scanning, but I wonder if there might be a need to scan at higher resolutions than one would normally use with a better quality scanner. I might try a few experiments to see how resolution decreases when scanning at lower resolutions. At some point it obviously will, but I haven't yet compared 2400 dpi scans with 4800 dpi scans.
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Some scanning issues
« Reply #14 on: July 11, 2005, 11:32:05 pm »

If you have a copy of Harald Johnson's "Mastering Digital Printing Second Edition", you may be interested in the material on scanning resolution in Chapter 3 up to page 92.

I am using a Minolta Dimage Scan Elite 5400 operated with Silverfast Ai for scanning colour negatives, and find the results (colour, detail, luminosity range) very satisfactory. But this scanner is only useful for 35mm film. It uses glassless carriers for both negatives and slides. Both the specs and the performance of the Minolta scanner are impressive with the right software. I bought Silverfast after I found both Minolta's own software (Version 1.1.2) and Vuescan to be rather unsatisfactory for producing balanced colour and adequate contrast from colour negatives, though I hear Minolta's latest software on the Scan Elite 5400 II is better with negatives than was version 1.1.2. Silverfast isn't cheap or intuitive (at first), but it saves mountains of time on post-scanning image adjustment once you get comfortable using it. Minolta's version of ICE does clean-up old cruddy Kodachromes eventhough they say in the instruction book that it doesn't (Johnson's book, page 89, bottom box).
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Some scanning issues
« Reply #15 on: July 12, 2005, 01:11:55 am »

Mark,
I'm still considering whether I should get the Minolta 5400 ll; Dmax of 4.8 and res of 5400 dpi. Wow!

Assuming the 5400 ll is similar to the earlier 5400 in basic operation, does it have an analog gain facility, perhaps by another name? I find this feature very useful on the Nikon, particularly with negatives that are overexposed so that the histogram is pushing against the right vertical, or underexposed for that matter. It's an ideal tool for creating dual scans for blending purposes and shadow noise reduction.

Vuescan has an automatic dual scan feature but with no control over the length of each scan. It seems more effective than multiple scans in reducing shadow noise, but not always as effective as Nikon Scan's analog control from -2 to + 2 EV (but clearly less hassle).

I'm holding off buying Silverfast's full version for either my Nikon 8000 or Epson 4990 because neither scanner is custom profiled. Silverfast seems to be virtually replacing most of the Photoshop adjustments one might use, but that's of no use if the colours in the Silverfast preview do not exactly match the image when opened in PS.

How do you get around this?
Logged

BernardLanguillier

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 13983
    • http://www.flickr.com/photos/bernardlanguillier/sets/
Some scanning issues
« Reply #16 on: July 12, 2005, 09:51:54 am »

Quote
Whoa! Hold your horses! This wasn't a comprehensive review of the Epson 4990, just some first impressions. I've now had time to scan a number of slides and negatives in both 35mm and MF formats. There's no getting away from the fact that this flatbed produces softer scans than the Nikon 8000. I would characterise the difference as being similar to the difference one would expect between a zoom lens and a prime lens.
Ray,

After your first test, how do you feel about the relevance of my guess that a 6*9 slide scanned with the Coolscan 8000 will be very close to a 4*5 sheet scanned with the 4990?

Just wondering.

Cheers,
Bernard

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Some scanning issues
« Reply #17 on: July 12, 2005, 11:59:33 pm »

Quote
After your first test, how do you feel about the relevance of my guess that a 6*9 slide scanned with the Coolscan 8000 will be very close to a 4*5 sheet scanned with the 4990?
Bernard,
I think you might be right  :D . Unfortunately, I haven't got any MF shots of test charts so it's all guess work. I've seen expressed on the web that moving up from MF to 4x5 is too marginal to be worth the trouble. It most certainly depends on the final print size.

If you were asking me if the Epson 4990 could produce a more detailed print blown up to the size of a wall, than a Nikon scanned 6x9cm, I'd make a guess and say, yes.

But maybe the ultimate print quality the Epson 4990 can achieve is with 8x10 film. I'll stick my neck out and declare that would be noticeably better quality than a Nikon scanned 6x9cm  :D .
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Some scanning issues
« Reply #18 on: July 13, 2005, 09:10:19 am »

Ray,

The only reason why I bought Silverfast is that after using Minolta's software and Vuescan I found Silverfast by far the best game in town for scanning COLOUR NEGATIVES, which is almost everything I do with legacy files from pre last October(since then I am 100% on a Canon 1Ds - what a relief!).

Except for massive amounts of dust/scratches in places with quite uniform non-detailed colour (such as sky), I spot my images manually, in PSCS2 mainly using the spot healing brush. It can be labour-intensive, but it is definitely the surest way of not losing anything you don't want to lose. Brushing the negs with a "Visible Dust" brush (the same one recommened on L-L for sensor cleaning) before scanning helps reduce this work to only the residual crud and imperfections that can't be blown or brushed away.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Some scanning issues
« Reply #19 on: July 13, 2005, 11:24:20 pm »

I'm afraid of damaging my negatives by applying chemical agents to them. Some damage make take a long time to reveal itself. Do you have long-term experience with this approach and is it better than using an electrically charged brush?

It is really impossible without being in the manufacturing industry to say whether or when new technology may resolve the conflict between automated dust removal and image detail, so until we have something better than what we have, I do it the slow and safe way.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."
Pages: [1]   Go Up