PDA

View Full Version : Flatbed v. drum scans for lightjet prints - Microtek 1800F.



QT Luong
17-Oct-2003, 13:25
I would like to hear from experiences that you might have of using a flatbed scanner for making exhibition-grade color lightjet prints, and how the results compare to using a drum scan. What would be the difference in quality ? Neglectible, small, or very visible ?

I am particularly interested about results obtained with the Microtek Artixscan 1800F, but positive experiences with other scanners such as the Epsons would be encouraging, since it appears that on paper the 1800F betters them.

I don't expect the resolution to be a limitation. In the past, i have used a very cheap (now broken !) 600dpi flatbed to make a 12x18 from 5x7 and the print was sharp, easily beating anything done from 35mm. Even if the real scanner resolution is only 1200dpi, one could still make a 4x to 6x enlargement. What I am wondering about is the dynamic range and noise.

Bruce Watson
17-Oct-2003, 14:53
There are of course many variables at play here. As you have seen, much depends on the level of enlargement. My context here is from the point of view of 4x5, BTW.

If you are scanning to make 2x enlargments, probably any scanner you have access to will do the job. Consumer flat beds (Epson 2450 or so) seem fine up to about 4x enlargement to me. Better flat beds do visably better, I think. After about 4x, I can really see a difference when you go to drum scanning, especially if you oil mount.

What is "better?" What I see as you go from consumer flat beds to pro flat beds is smoother tonality, smoother transitions, somewhat better dynamic range, somewhat better apparent sharpness. It's sort of like removing a veil (thin, uniform) from between you and the print.

When you get to drum scaning, what I see is improved sharpness, improved smoothness (especially transitions), and much improved dynamic range. It's like removing a somewhat thicker veil from between you and the print. A lot of this is due to PMTs vs. CCDs. Quite a bit is due to oil mounting on a drum, especially the smoothness.

If you really push your transparencies (and you do, at least what I've seen on your website) and you want maximum quality, I think you are going to be happiest with a drum scan, especially at an enlargement in the 4x to 6x range and larger. The reason is, you'll see more detail in the shadows, and all most all of the smoothness of tone and transitions that you see on the light table.

Of course, YMMV. The easiest way to tell, is to scan the same negative multiple ways and make prints (same size of course). Put the prints on the wall together under the same lights, and have an apples-to-apples comparison. The one you are happy with is the one you are happy with.

Neal Shields
17-Oct-2003, 20:37
To satisfy my own curiosity I did a comparison between how much detail that I could extract from a 4x5 velvia slide with my Epson 1640 and having a professional drum scan done. ( I used a 1/2" tall section of the film and got a 100 meg scan)

The following is a link to a power point presentation I did showing the difference. To my mind it is astounding. The name of the construction company on the crane is 1/10th of a mm tall on the film.

http://truckgenerator.com/subdomain/sueandneal/ppiscansvelvia_files/frame.htm

(It takes a while for the 12000 dpi to load over a phone line.)

I also took the drum scan into Photoshop and backed the resolution down until I got back to 1600 dpi. The difference between the drum scan and the flat bed scan was again dramatic.

My conclusion is that for posting to the web I will use my flat bed scanner but if I decide to go through digital to get prints, I will spend the money to get a drum scan.

I have also tried going above the optical resolution on the 1640 and find that all I get is a bigger file.

Neal

MattO
18-Oct-2003, 00:42
All I see here is a fuzzy image - no powerpoint presentation or any comparisons -

MattO
18-Oct-2003, 00:56
If you are looking for a reasonable alternative to drum scans that could save some money, you will not find it with the <$1,000.00 flatbeds. Period. The Epson 1680 Pro is likely the best flatbed you can buy, but still falls short of drum scans, but only at higher print sizes is it noticeable. For gallery quality, I would not even consider less than a drum scan.

violin
18-Oct-2003, 01:07
i agree with matt, "gallery quality" to me means "highest possible standards" unless your particular asthetic is deliberatley low quality. i did some tests with my local repro house comparing flatbeds, "virual drums" (imacon), tango drum and really high end heidelberg (8900). the difference is real and there to be seen, by even the most untrained eye.

Saulius
18-Oct-2003, 07:54
Tuan, I recall this subject was touched upon in photo.net lf forum at the time this one was not operating. You might try checking this out.

http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=004N6n

Neal Shields
18-Oct-2003, 08:14
Matt,

You should be able to switch between different scans by clicking on the arrows at the bottom of the page, the titles on the left, or by clicking the slide screen and then using your left mouse button. It works for me,but I am still learning to build web pages, I may have screwed up.

Neal

Dan Baumbach
18-Oct-2003, 09:59
I don't think the issue is resolution as much as dynamic range. I scan my 4x5 trannies on an Epson 3200 scanner and the resolution is pretty good. However when I want to make an exhibition print I rent time on an Imacon scanner. The sharpness is definately superior and the increased dynamic range make getting the color I want in the print much easier. There were some images where the highlights were completely blown in the Epson scan while the Imacon got good detail in the hightlights.

- Dan

Michael S. Briggs
18-Oct-2003, 14:23
Re Neal's webpage with scanner comparisons: it won't work in any browser that I regularly use (Opera 6.11 and Netscape 4.8). Looking at the source code, it was written by Microsoft Power Point. It will probably work in Internet Explorer but few other browsers. Microsoft has difficulty writing standards-compliant webpages. I've encountered plain-html webpages written by Front Page that show many dozens of basic html errors when checked by an html validation webpage.

Bruce M. Herman
18-Oct-2003, 15:14
Tuan,

I have an 1800f, and previously owned a Linocolor-Hell Ultra Saphir II. The 1800f does have a lower noise level than the Ultra, but Heidelberg's Newcolor 5000 software is superior to SilverFast in ease of use and in achieving the correct color balance on the first pass. I've used both to make prints up to 11x14 with an Epson 1280 printer using Epson's ink set. The differences in dynamic range are small, but visible.

I've also had scans from drum scanners and subsequently used these to make inkjet prints on my Epson 1280. I cannot see any real difference between these and the same prints made from 1800f scans.

I would make the following observation: although I have not used Epson's ultrabright inks, the 3rd party pigment inks that I have used with the Epson 1280 are well known to have a smaller gamut than the Epson dye based inks, not to mention the gamut in a LightJet print. If you are printing with these inks, a scanner with a Dmax of 3.4 would probably be more than adequate.

What would I do? For LightJet prints, I would go to the drum scanner. For dye based or comparable gamut inks, I would use the 1800f. For most pigment inks, or for printing on uncoated fine art papers, I would use a less expensive scanner.

One last note - I have found that the 1800f driver is very picky about other applications running when it is installed. If you work on an IBM clone, install in the safe mode. I also use 2 or more scans (a setting in the SilverFast software) to avoid some color artifacts that seem to appear in single pass scans.

I hope that this has been of some help, and that you read it in good health.

Cheers, Bruce

Henry Ambrose
19-Oct-2003, 18:02
I have owned and used the Leafscan45, the Agfa 2500 (no longer made and long since sold by me) and an Epson 1680 for 4X5 film. All give fine results with 4X5 film. I do mostly color negative but did not have any trouble with any of these scanners with transparency film or B&W negs. Actually sometimes the Agfa would fail me on real dense film but you won't buy one of those anyway.

What is missing from your question I think is the size of the prints you want to make. I now regularly use the 1680 to scan 4X5 film and have 8.5X11 prints output from a Durst Lambda with 400 ppi input that I (and very important my architecture clients) find completely wonderful in their quality. A very small reproduction size but I've made lots of them with great success and a little cropping does not hurt them at all so I feel like I have a cushion to work with. The largest print I've seen from my 1680 scanner was a ten foot by ten foot trade show display background that looked fine, and again a satisfied customer. It was not fine art or even close but looked quite good given the output device they used and the purpose intended. From on-screen viewing of the files I get from the Epson I estimate they are great to 11X14 from the Durst printer but I've only made large 4X5 prints on Epson inkjets and they have a way of making not-so-good scans look better than they really are. I can compare the Leaf at 2500 ppi from roll film and the Epson at 1600 ppi from 4X5 and the Epson/4X5 scans are simply amazingly better with both scans being about 120MB file size. For what its worth I've worked extensively as a commercial photographer with digital cameras and film scanners since 1991 and I've seen and worked on lots of scans from most all kinds of scanners since that time.

Anyway, my point is that if you're not making large prints then the Epson 1680 will do a great job. This scanner has a reputation for not having great lenses and I agree. The Leaf is much sharper as a file returned from the scanner, as is the Agfa. However the 1600 pi res of the Epson overcomes the Leaf and Agfa on 4X5 film. The Leaf as it only gives 1000ppi output from 4X5 and soon runs out of resolving ability. The Agfa because it has some trouble with dense film. Though a bit fuzzier, after proper unsharp masking the Epson scans look wonderful. Great color, detail and smoothness. This scanner has a much better range than the Agfa and makes better prints even though it is only 1600 ppi vs. 2500 ppi and of course you're scanning through glass with the Epson. I have not compared the 1680 to the Epson 2450 or 3200 which are much cheaper than the 1680 but the specs lead me to believe that the density range of these cheaper scanners might be lacking. I can not say for sure from experience. A friend has used the 2450 to make big three foot wide Epson prints from 4X5 B&W negs that are superb, so I might be wrong. The Epson 1680 sells for about $1100 with the film adaptor.

I'm with the others about drum scans being better, but if you're making smallish prints you can make them great on a flatbed used skillfully. Since you are using 5X7 I think you will get outstanding 11X14s which will be about equal to my 4X5 at 8.5X11. The repro ratio is in your favor. As print size goes up you may want better scans but then the price of the print should cover the additional scanning cost. If you want, send me a piece of film, and I'll scan it for you on my 1680 so you can get an idea of the quality.

Dick Roadnight
21-Oct-2003, 16:50
Neal Thank you, that is useful infomation, but the 1,200 is not significently better that the 4,000, and would not, IMO show a vast inprovement over the quality 4000, 5000 and 6000 ppi flat-bed film scanners.

Dick Roadnight
21-Oct-2003, 16:54
Has anybody used e.g. a 16 shot digiback and a Digitar macro lens to get LF pictures into a computer?

Lyle Aldridge
23-Oct-2003, 12:14
Neal, your comparison is striking, but I'm wondering if there's something wrong with your 1640. I've been using the same scanner to scan 4x5, and your 1600ppi example looks like the output from mine with about a 2p Gaussian blur applied.



Perhaps I've misunderstood the methodology. How much of a piece of film are we seeing? From your statement that the crane lettering was about 0.1mm, I've inferred that we're looking at an area of about 3mm tall, or roughly 1/8 of an inch.