Drum Scanning for Large Prints

[Scott Rosenberg]

Greetings,

I just received an order for several really large prints – sizes that will require the absolute best quality scans available. In reviewing the scanner comparison on this site, it appears that the ICG drum scanner is the best option, at least of those tested. As such, I’d like to have my originals scanned on an ICG scanner – do any of you work with labs that run an ICG that you’d feel comfortable recommending? For those of you that have experience with such things, are the ICG’s as superior to Tango scans as the comparison would lead one to believe?

thanks,

[Bruce Watson]

I just received an order for several really large prints – sizes that will require the absolute best quality scans available. In reviewing the scanner comparison on this site, it appears that the ICG drum scanner is the best option, at least of those tested. As such, I’d like to have my originals scanned on an ICG scanner – do any of you work with labs that run an ICG that you’d feel comfortable recommending? For those of you that have experience with such things, are the ICG’s as superior to Tango scans as the comparison would lead one to believe?

Scan quality depends on a number of things, not the least of which is the skill and dedication of the operator, the image itself, and the type of film.

The old Tango (Heidelberg exited the scanner market in 2001 IIRC) gained a reputation in the advertising and magazine world which was its target market. It's a good scanner, don't get me wrong. But for the fine art market, and in particular when you need enlargements for large print sizes and/or are using negative films, there are probably a number of drum scanners that will better serve your needs. And the ICG line is among them.

I think the tests here (http://www.largeformatphotography.info/scan-comparison/) are fairly representative given that you have to view a fairly small image on a monitor.

Sorry I don't have an ICG lab recommendation for you.

[Ted Harris]

Scott, the best scan is a combination of the skill of the operator, the quality of the software and the capability of the scanner. The ICG appears to have the highest resolution and dMax/density range capability for all the scanners on the market today but note I said appears to have. In the scanner comparison that Bruce referred you to the ICG did terribly in fact. We have discussed the results and think there may be a lot of operator error but suffice to say it just didn't perform.

OTOH, when Seybold (who do printing industry standard setting and testing) tested a group of 14 high end scanners strictly to determine which performed the best with 4x5 film the scanner that I operate, the Screen Cezanne, came out on top. The Tango,whichscored acceptably, was behind several other scanners, including two other 'high end flat beds' and was tied with the Creo Eversmart Supreme.

See my PM.

[Scott Rosenberg]

Bruce,
that's the test i based my preference for the ICG scanner on. i agree that the ability of the operator can likely mean more than minor hardware differences, so i'm really hoping someone can recommend a lab that has both - first class hardware and top caliber folks driving them!

Ted,
To me, the ICG scans were the sharpest… I’ll have to go back and look again, as I might be remembering wrong. PM received and replied to.

[Ted Harris]

I should have noted above that the ICG did very well in the resolution area but botht he dMax and density range were petty bad ... look especillyat the crop of the little red container ... see how sharply and quickly the shadow detail falls off.

[Michael Mutmansky]

The ICG scan has higher contrast than most of the others, and appears to have a bit of sharpening in the 'unsharpened' file. These two things make it difficult to be sure whether the scan is sharper or not.

The other consideration is the balance between grain and smoothness of tone that the scan operator makes. With drum scanners, it is possible to change the aperture to reduce the grain in the scan. While this reduces the sharpness a little, it can make a big impact on how smooth the print looks later on.

I think the biggest difference in the comparison on this website between the ICG and some of the other scanners is the contrast, though. For example, look at the Microtek 2500 scan samples, which also have considerably higher contrast than most of the rest. Many people who look at a print from that scan think it is sharper than many of the other scanners out there, but it clearly is not, but the contrast along with a little sharpening give the general impression that it is.


---Michael

[George Kara]

May I suggest you not worry as much about scanner brand as quality. Is it possible for you to hunt around your area to find printed images that you consider to be of the quality you need? You could then find out who did the scans and discuss your needs with the scanning company directly? I have found that scan quality when viewed on a screen doesnt directly correlate with the printed image.

My Canon Eos1ds looks significantly better than medium format scans on the screen, but are inferior in the final analog print. I'm not sure why this happens.

Most of the high end scanners will do an admirable job and the brand differences may not have a significant impact on the final print.

George

[Michael Mutmansky]

Scott,

I was scanning some B&W last night, so I did a sample to show you what I mean about balancing sharpness with smoothness of tone.

The image below was made on TXT with a Mamiya 6 camera.

[naturephoto1]

I do not have the experience to comment on the comparison of other high end scanners with the Heidlelberg Tango. I will have to agree that the operator and the software are probably as important as far as the quality of any of the high end machines. Bill Nordstrom has been my master printer for the past 10 plus years and we have been preparing what most consider outstanding output on Fuji Crystal Archive Photo paper from files scanned with the Heidelberg Tango (as well as what we used prior to the unit). Bill is quite knowledgeable on the usage of the machine and has kept it current with current software.

Rich

[Michael Mutmansky]

The red area in the image above is shown at actual pixel size below, with four different scanning apertures.

The sharpest is also the grainiest, so if you like smoothness of tone, that would probably be a problem for you.

This same condition happens with slide film, and especially with color negative film as well.

The scans are all 4064 SPI scans, FYI.


---Michael

[Michael Mutmansky]

I just noticed that the second image looks a little more grainy in the thumbnail, but that is an artifact of the site software used to reduce the image for the thumbnails. If you click on the image, you will see properly.

The difference between the first two is subtle, but is obvious for the others.


---Michael

[Bruce Watson]

The red area in the image above is shown at actual pixel size below, with four different scanning apertures.

The sharpest is also the grainiest, so if you like smoothness of tone, that would probably be a problem for you.

This same condition happens with slide film, and especially with color negative film as well.

Your four samples are interesting, but I suggest that anyone who wants to learn the truth of this should make prints. The reason is two-fold. First, the difference between the monitor (light source) and paper (reflected light). Second, the resolution difference (monitor from 72ppi to around 100 ppi, printer from around 240 ppi to 360 ppi).

To figure this out years ago when I was learning to drum scan, I scanned the same film (5x4 Tri-X) three consecutive times. The only thing than varied was aperture. The first one was what the software ordered (6.25 microns), the second and third were opening up one step (12.5 microns and 25 microns).

I printed each of these scans (each print being a 30cm square out of what would have been a print 125 x 100 cm, around 11x enlargement, just under 4000 spi). No sharpening. For those interested, Epson 7600, Cone PiezoTones, StudioPrint RIP, Hahnemhule Photo Rag.

I put the three prints on my "proofing wall" under solux 4700K lamps. Then I got people to look at them (friends and neighbors, no scanner operators). I asked just one question -- which one do you like the best?

IIRC, 7 of 10 liked the 6.25 micron print best. Several volunteered that it seemed "sharper" to them. Three of the 10 liked the 12.5 micron print. One volunteered that it looked "smoother." No one picked the 25 micron print. What can I say about the observers? They were all right!

While the difference between the 6.25 and 12.5 micron scans was large in Photoshop, it was small in prints. And what people could see had them choosing the sharper print.

So what's the moral of my story? Like most tools, manually overriding aperture size is a tool that should be used sparingly. While you can vary aperture size to trade off sharpness for smoothness, the difference isn't that great in prints, and the average viewer often likes sharpness over smoothness.

[Michael Mutmansky]

Bruce,

My point is that a drum scanner can produce multiple results with only subtle adjustments to the input settings.

I know people who do not like to see film grain in their prints, and when I scan their work, we will scan a bit softer to help avoid that.

I prefer to see the grain (it is, after all, a film source) so I scan sharper for my own images.

The important thing is to understand what the scanner is able to produce, and then work with that to get a result that is most like what you want. It's a complex set of variables, and it takes a lot more than blindly sending a negative to a service bureau to get the best result for your needs, because they don't know your requirements or preferences.



---Michael

[Gordon Moat]

http://www.theslideprinter.com/scanning-services.htm

Denver Digital Imaging Center

I have not tried them out, so I have no idea on whether or not they are running their ICG to the best of its capability. However, I have bookmarked their site as one place that runs an ICG, just in case I want to try out a scan on one. Considering the cost, I would reserve that for images that needed the darker tones to come out really well resolved.

Ciao!

Gordon Moat
A G Studio (http://www.allgstudio.com)

[JW Dewdney]

My results with A&I in Los Angeles have been quite good. They only charge 35 cents per megabyte, too, if your scan is 200MB or above. I assume you're in that range.

[Henry Ambrose]

I think at this point you would be best off trying a couple of places to see who can deliver your job as you want it.

Chromatics in Nashville www.chromatics.com has really good equipment and operators. They can make great big Lambda prints for you too. As far as that goes they can handle all your needs from development to shipping big mounted prints.I've been consistently pleased with their work for years.

[Doug Keyes]

Scott,
I use a great resource here in Seattle. His name is Dick Busher and he runs a small outfit called Cosgrove Editions. He is very experienced and does alot of scan work for very hi-end photography books. I've had 6x7 transparencies drum scanned at 4000ppi (which is the max they do). I know other people that have been happy with his work as well. (Chris Jordan, who writes on this forum is one)

You can contact Dick Busher at 206-524-6726

One note on scanner quality, I agree with most people here that the operator is a large part of the equation. For me it's the biggest part. I'm very picky when it comes to scans. My gallery and museum clients expect the same quality.

Good luck,
Doug

[David_Senesac]

While I've had various scans made for many years, there are a number of people on this board with much more scanner knowledge than mine. My sense the last few years has been that for large format film scans, any drum scanners being used by higher end labs can easily capture as much detail as is on our film. If I have my Provia scanned higher than 2400 ppi, I just see increasing grain. At that size one will end up with a 32x40 inch print at about 300 ppi that is considered all the resolution necessary to view detail at arms length. Of course all these 16 bit capture drum scanners have resolution and Dmin/max specs well beyond that so what is the point if an ICG can capture 12000 ppi while a Tango only 8000? Maybe of some use with the highest resolution black and white films but not for color films. Isn't the quality bottleneck in quality with printing and their inks and not with the film or scanning? And as others have said the real key to getting good output is to have someone that has the experience to make good scans for fine art purposes. I learned long ago that simply delivering film to a service bureau that has similar drum scan equipment does not guarantee quality. If such labs are busy they may slap one's film on a drum with film from several other business customers with generic settings that do little in terms of customizing the machine settings for one's target film. ...David

[Gordon Moat]

In general (and very general terms) there is very very little benefit to scanning beyond 6000 ppi. However, there are a few good reasons to do just that. One thing to consider is that scanning alignment happens in horizontal and vertical alignment, so off axis details might be a good thing to oversample. The other aspect is tonality of slight colour gradients; simply having more colour sampling points is not a factor of resolution, though it can provide more points of data to create a smoother print. Cropping and editing in post are other reasons to oversample.

With Dmax figures, obviously if you match or exceed the capability of your film, then you can probably feel satisfied of capturing as much information as possible. I slightly disagree on the Tango to ICG comparison, in that I do not feel that a Heidelberg Tango can capture the densist film area to the same ability as a newer ICG drum scanner. If you shoot dark images, night images, or images with complex details in darker tones, you might want that extra information.

Remember all this is operator dependant. A well run Tango will provide better results than a poorly operated ICG. In fact, a well run high end flatbed scanner could also provide better results than a poorly (or so-so) operated drum scanner. The machine is only one part of the results. While internet postings of results are interesting in comparisons, my opinion on the best way to judge scanning results is by seeing those results in print; when the print meets your expectations, then you have a good scanning solution.

Ciao!

Gordon Moat
A G Studio (http://www.allgstudio.com)

[lutherasmith]

[QUOTE=Michael Mutmansky;227259]The red area in the image above is shown at actual pixel size below, with four different scanning apertures.

The sharpest is also the grainiest, so if you like smoothness of tone, that would probably be a problem for you.

Michael,

What apertures are you using? These are drum scans?

Thanks,

Luther

[Michael Mutmansky]

Luther,

Too long ago to be absolutely certain, but probably 6.25, 12.5, 25, 50 microns. There's issues of focus in a drum scanner as well, but I left that variable out of this equation. Fyi, the best that can be achieved from a consumer flatbed will probably look somewhat like the third image (25 micron). There's a awful lot lost between the second and third images.

When I scan Tri-X 120 for my own purposes, I'll normally scan at 12.5 microns. It reduces the grain speckle a little, but has no real negative impact in the apparent sharpness of the film. I believe that's because the grain is slightly larger than the 4064 SPI size.


---Michael

[Jim Jirka]

Scott,
Give a call to Danny Burk. I know he does a wonderful job with his scanning services.
Just a thought.

[lutherasmith]

Luther,

Too long ago to be absolutely certain, but probably 6.25, 12.5, 25, 50 microns. There's issues of focus in a drum scanner as well, but I left that variable out of this equation. Fyi, the best that can be achieved from a consumer flatbed will probably look somewhat like the third image (25 micron). There's a awful lot lost between the second and third images.

When I scan Tri-X 120 for my own purposes, I'll normally scan at 12.5 microns. It reduces the grain speckle a little, but has no real negative impact in the apparent sharpness of the film. I believe that's because the grain is slightly larger than the 4064 SPI size.


---Michael

Michael,

Thanks, I am still learning to use my drum scanner. I mostly use the auto aperture setting.

Luther