Scanner recommendations please

[AndrewEllinas]

I want to digitise 30 years worth of film taking on 35mm and 6 x 6 formats, negatives, slides and prints.

Probably 75% of it is 'family album' material.

From what I've read here and an eslewhere, it's impossible to beat a dedicated film scanner.  Flatbeds, like the Epson V750 pro, don't appear to be that good.  Nikon Coolscans get great reviews but are very pricey and don't seem to be available.

What would you recommend? 

Andrew

[Mark D Segal]

What is the percentage of 35mm versus medium format film? Which is more important to you? roughly how many images do you think you may end-up scanning? Given the market situation for scanners these days, answers to these questions would help inform appropriate advice.

[AFairley]

The Epson V700 and V750 actually will give a dedicated film scanner a good run for the money.  I found that head to head with the Epson V700 and a Coolscan V you will see a difference when pixel peeping.  In practice unless you are doing large prints, not so much.  For "family-oriented" stuff, certainly Epson would me more than adquate, IMO. 

[Mark D Segal]

This reminds me Andrew - magnification also makes a difference. Perhaps you would like to say something about the sizes you intend to print the pictures at, and whether you envisage doing a fair amount of cropping. When you crop photos, for any given size of output, the magnification increases according to the amount of remaining image material, such that the intrinsic quality of the scan would start to matter more. AFairley's recommendation could make sense, but I'd like to learn more about your needs and intentions before saying whether I would concur unreservedly.

[AndrewEllinas]

Thank you for your posts.  In reply:

1.  80% 35mm, 20% 6 x 6.  Some of my best photographs were taken on the 6 x 6 format, but some of the most the important, in terms of family archive, are on 35mm.

2.  I rarely print larger than 10 x 8.

3.  I want to scan as many as I can, but the reality is that I doubt I will do much more than 2,000/3,000.

Regards

Andrew

[Mark D Segal]

Andrew,

That's a lot of scanning. One option a fair number of people have tried, successfully, is not to scan their own stuff, but send it to a commercial service such as Scancafe. Their basic package is reasonably priced, and from what I've seen of their output, it's pretty good. Between their basic and pro packages you can most likely get the quality you need. Their pricing is arranged so that to a considerable extent you pay for what you wish to keep, not necessarily everything they scanned. 2000 of their scans at a mixture of pro and basic pricing would cost you the value of a good scanner, but the time saving would be immense. Worth thinking about.

If you wish to do all the work yourself, which is the only way I would handle my stuff, because I want total control over the image at every stage of the work, the Epson V750 (this is the model with the fluid mounting accessory and a better version of SilverFast than what comes with the V700) would be the way to go for handling both the 6*6 and 35mm. It's resolution is considerably better for the 6*6 media than for 35mm. For 35mm, it's passable, but you'll get noticeably better quality using the Plustek 7600i. It has very decent resolution for the price (about USD 450). This website has published my review of the Plustek - you may wish to have a look.

If you want to be a bit extravagant and buy two scanners, a combination of the Plustek and the Epson would set you up very well for what you wish to do. You can also consider buying the Plustek and sending the medium format to Scancafe, or buying the Epson and scanning it all at home, or sending batches of the 35mm to Scancafe.

I would also recommend going to the used equipment sites and looking for a Minolta Dimage Scan Elite 5400 model 2. This 35mm film scanner is gem. (Their model 1 is also excellent, but very slow.) The resolution is 5400 PPI - higher than Nikon's flagship 5000 and the lens quality first-rate. While the second-hand Nikon scanners sell at a huge premium over their original retail prices new, the pricing on the Minolta is far more reasonable and the quality excellent. Both are discontinued, but servicing on both will probably remain available until the service outlets run out of spare parts. You'll be able to drive them on the newest operating systems because SilverFast and Vuescan have up-dated drivers, which the original manufacturers have stopped doing. Minolta also had a comparable model - Dimage Scan Multi Pro, a very high quality dedicated film scanner for handling both 35mm and 6x6. You would do very well to pick-up one of these on the after-market as well, but they can be expensive.

Well, there's a range of thoughts for you. No one resounding answer here; just a number of suggestions - it is a difficult area these days because so much good equipment has been discontinued.

[AndrewEllinas]

Hi Mark,

Thank you for your very informative post.

I think buying a Plustek 7600i for my 35mm films and sending out my 6x6's to a UK equivalent of SacnCafe seems to be the best solution.

Regards

Andrew

[Oldfox]

Check that the scanner has a batch mode. With your amount changing the negs/slides one at a time will be 'pain in the ...'.

[Mark D Segal]

Andrew,

The following post may also help you decide on dedicated v flat bed scanners http://www.luminous-landscape.com/forum/index.php?topic=20803.msg163653#msg163653 - reply 14.

As Mark has mentioned, 35mm scanning on the Epson V750 is only passable.

Hi Nick,

That was an excellent piece you wrote back then. It did prompt me to mention something I was about to do anyhow for Andrew - he needs to think about the software every bit as much as the hardware. The Plustek does a pretty good job of replicating grain, dust and scratches too. Not quite as sharply as the Nikon, but very much there. Fortunately, the dirt and scratches aspect has become completely manageable while retaining fine image detail and quick scan times using it with the bundled SilverFast version http://www.silverfast.com/product/Plustek/719/en.html. I think best value is to buy the bundle with the full featured SilverFast product including the IT8 profiling target, because this scanner benefits from custom profiling, as mentioned in my review on this website. iSRD is a much better scratch and dirt tool than ICE ever was, in terms of the user-control it allows and its usability on Kodachrome. Dealing with grain is another matter. SilverFast does have a grain filter, but I'm less enamoured with that tool. I think it's better to leave both the grain and sharpening out of the scanning process altogether and use specialized plugins to Photoshop for that. My choice for grain mitigation this days is ToPaz DeNoise (used on a duplicate image layer) and then for sharpening - PK Sharpener Pro. Yes, it means investing in a couple of pieces of software, but for the volume of work Andrew wants to do, and assuming he wants great image quality, this is the way to go.

Hi Andrew - your strategy makes sense.

[AFairley]

Andrew, one thing I do recommend is that you borrow a scanner and try scanning some stuff yourself before committing yourself to any particular path.  My experience with scanning is that it is incredibly time consuming, between cleaning the media, tweaking color/exposure adjustments, and (in the case of the Coolscan) tweaking focus if needed.  I can easily spend 15-20 minutes to scan one slide.  I imagine that people who scan more can go a lot faster, but I would think that 5 minutes per transparency would be about the fastest you can go.  (If someone can prove me wrong, please post and give me hope!)  Obviously if you are scanning strips of negatives, things will go somewhat faster, but even so you are looking at a lot of time.  Actually doing some scanning to find out if it's you're cup of tea might be a good idea.

[Mark D Segal]

Andrew, one thing I do recommend is that you borrow a scanner and try scanning some stuff yourself before committing yourself to any particular path.  My experience with scanning is that it is incredibly time consuming, between cleaning the media, tweaking color/exposure adjustments, and (in the case of the Coolscan) tweaking focus if needed.  I can easily spend 15-20 minutes to scan one slide.  I imagine that people who scan more can go a lot faster, but I would think that 5 minutes per transparency would be about the fastest you can go.  (If someone can prove me wrong, please post and give me hope!)  Obviously if you are scanning strips of negatives, things will go somewhat faster, but even so you are looking at a lot of time.  Actually doing some scanning to find out if it's you're cup of tea might be a good idea.

Indeed, I alluded to the time issue in reply #1 above, but you have put a finer point on it. No matter how one goes about this, it is time-consuming. And scanning negatives, which I do extensively, if anything can be more time-consuming. But yes there is hope. Firstly, I spend the minimum amount of time cleaning media - a few brush strokes with a bespoke soft-haired brush and an air blower - a few seconds of time. I let SilverFast iSRD look after the rest. Saves lots of time and the results are fine. Secondly, for the scanner's we're talking about there is no manual focus, so no time spent on that. Thirdly, for hue and tone adjustments, the primary time-saver is to make sure the colour management is set-up correctly. That little investment will save scads of time. Fourth, don't waste too much scan-time with multi-sampling - not worth it. Using SilverFast Multi-exposure which is a two-pass is enough. But once you use iSRD in SilverFast, and a good noise/grain reducer in Photoshop, the need for multi-anything is reduced even further. So what remains is (i) getting the film into the scanner, (ii) hitting the pre-scan, (iii) cropping the image, (iv) hue and tone adjustments, and (v) scan. compared to a digital workflow from a digital camera, items (i), (ii) and (v) are time-added. (iii) and (iv) you need to do anyhow, whether at the scan stage or the post-scan stage. So bottom line: yes, image processing takes time no matter what the route, but with efficient procedures, scanning need not add *that much* relative to what you'd spend on a non-film digital workflow. But there is an unavoidable minimum that it does add. One does this as a labour of love, not goal-driven speed.

[BradSmith]

Andrew,
I've just been down this path.   Scanning all the images yourself will clearly drive you insane.  Or you'll quit before you finish.

Background:  I have an excellent dedicated scanner - the previously mentioned Konica Minolta.  I have over 6,000 slides covering nearly 40 years from before I went digital.   I wanted to digitize the keepers, defined as important enough family pictures plus those that I might print for framing at 8x10 up to 16x20.  I plan on producing photo books (Blurb) for the family photos, plus I want to pass on the digital files to my son.  I also print landscapes that I like for hanging on my walls.

What I've done and am doing (it's a process  is:   I reviewed about a 1000 on a light table.  Threw away probably 60-70%.  The remaining 300+ I sent to Scancafe as a trial.   Got them back and was satisfied that they'd work, with minor tweaking for my Blurb book purposes.  After reviewing the Scancafe scans, I chose about 20-30 that I might work on for printing for my walls.  I plan on scanning the ones that I'd want to print on my Epson 3800 with my Konica Minolta to get the very best possible file I can produce.   For that relatively smaller number, I'm willing to take the time.

I'd suggest you start by giving Scancafe a try on a sample of your collection.  You'll see if it is good enough for your purposes.
Brad

[PeterAit]

The V750 actually does a great job with negatives and slides. It is hideously slow, taking over 5 min to scan a tray of 20-35mm slides, but the results are good enough for all but the most critical work.

[Mark D Segal]

The V750 actually does a great job with negatives and slides. It is hideously slow, taking over 5 min to scan a tray of 20-35mm slides, but the results are good enough for all but the most critical work.

Could you get a bit more granular about "great job", and "good enough"? What do you mean?

[Steve House]

Could you get a bit more granular about "great job", and "good enough"? What do you mean?

Exactly my thoughts as well.  To me, "great job" would mean when scanning 35mm original, either colour transparency or negative or B&W negative, and printing at a typical display print size, say around 11x14 up to 16x20, when viewed normally the resulting print quality would be indistinguishable from a classic conventional print made from the camera original by traditional wet-chemistry darkroom methods.

[AFairley]

Exactly my thoughts as well.  To me, "great job" would mean when scanning 35mm original, either colour transparency or negative or B&W negative, and printing at a typical display print size, say around 11x14 up to 16x20, when viewed normally the resulting print quality would be indistinguishable from a classic conventional print made from the camera original by traditional wet-chemistry darkroom methods.

There are lots of comparisons floating around out there.  Their worth depends on the rigor of the testor.

http://www.rangefinderforum.com/forums/showthread.php?t=45878

http://forum.silverfast.com/topic6097.html

I find the results between my v700 and Coolscan IV closer than above -- but my Coolscan is lower res that the examples above.]

BTW, what no one has mentioned so raf is that a good DSLR and macro lens with the right setup will do an excellent job on B&W negs and color slides and is blazing fast to use (compared to scan times) -- color negs are tricky because of the mask.

[Mark D Segal]

-- color negs are tricky because of the mask.

In case you missed it, this may be of interest:

http://www.luminous-landscape.com/techniques/scanning_colour_negatives_raw_or_not.shtml

[Mark D Segal]

There are lots of comparisons floating around out there.  Their worth depends on the rigor of the testor.

http://www.rangefinderforum.com/forums/showthread.php?t=45878

http://forum.silverfast.com/topic6097.html

Neither of those tests in the referenced links are determinative, due to methodological issues. The appropriate way of making these comparisons is to use a bespoke resolution target such as the SilverFast version USAF 1951 target. Having done so several times with three scanners, I confirm it remains the case that a high quality dedicated film scanner (such as a Nikon Super Coolscan 5000ED and even the Plustek 7600i) will produce a higher resolution result than an Epson V750. I also note that measured resolution for every scanner I've tested is well below the manufacturers' stated optical resolution. When testing with real photographs, I have found the Epson V750 is "passable" for prints at least up to 11*16.5 inches (the largest I've tested for), but noticeably not as crisp. When I say "passable" I mean the image detail is not fuzzy, it just isn't as "crisp". I shall not be able to provide quantitative data because it is proprietary and still under review. However I have made rigorous comparisons in this article: http://www.luminous-landscape.com/reviews/scanners/plustek-pdf.shtml

[Bart_van_der_Wolf]

Neither of those tests in the referenced links are determinative, due to methodological issues. The appropriate way of making these comparisons is to use a bespoke resolution target such as the SilverFast version USAF 1951 target.

The USAF 1951 target isn't really reliable for the absolute determination of limiting resolution when used with a discrete sampling detector like a CCD scanner. It is sensitive to aliasing artifacts and accidental alignment with the sensel grid. A displacement of half a sensel can make a pattern go from fully resolved to totally gray. This target will give a visual impression, but is not suited for an absolute numerical evaluation (which partly explains the hugely different findings that are reported on the internet). On analog systems (such as film) it works as advertized, but not on e.g. CCD sensors.

That's why the ISO determination of sensor resolution uses a slanted edge type of target for MTF determination to determine the limiting resolution (e.g. at 10% MTF).

Cheers,
Bart

[Mark D Segal]

The USAF 1951 target isn't really reliable for the absolute determination of limiting resolution when used with a discrete sampling detector like a CCD scanner. It is sensitive to aliasing artifacts and accidental alignment with the sensel grid. A displacement of half a sensel can make a pattern go from fully resolved to totally gray. This target will give a visual impression, but is not suited for an absolute numerical evaluation (which partly explains the hugely different findings that are reported on the internet). On analog systems (such as film) it works as advertized, but not on e.g. CCD sensors.

That's why the ISO determination of sensor resolution uses a slanted edge type of target for MTF determination to determine the limiting resolution (e.g. at 10% MTF).

Cheers,
Bart

Quite possibly, but I've been able to replicate similar readings on different occasions and they all seem to produce the same order of quality between the units tested. I've worked with slanted edge material (in real photos) for various kinds of resolution evaluation, so I hear you.