Drum scanning technology: Why so expensive?

[Ed Richards]

Ted - no arguements at all on color, I only do black and white.

Bruce - that is all correct, of course, but isn't that why we do zone stuff to match the negative to the output, even if it is a lower dmax scanner rather than a lower dmax paper? You can certainly pull a lot more range out of a negative with a drum scanner, but isn't the real question how much of that you can put back on paper?

Since all I can get on the paper is 256 shades, if that, my task is to get enough out of my negative to fill the range of the paper, not to get everything out that is possible. My problem, which is the same as those who print in conventional means, is to do my best to get my negative to match the range of my output media. The consumer scanner will capture more dynamic range than I can get on paper, which is all the really matters, if the negative is right.

Sharpness, OTOH, is a bigger issue and one where the drum will clearly do better and where the only thing I can do with the negative to comensate is to take pictures of things with no fine detail.:-) So there will be an increasing gap in sharpness between my scan and a drum scan as the print gets larger, if the viewer stands close, and if the image depends on fine detail to work.

My bigger concern, and one that I think gets lost in this crusade for techical perfection, is that very few memorable images depend on techinical perfection. We can tinker with technology, and buy better technology, but we cannot buy inspiration and timing. Those we only get by shooting a lot of pictures so we are ready when it happens. I do not think this is an issue for Ted or for Bruce, but I read a lot of posts from folks on the list who are new to photography or who have not found their way yet, and I worry that they will let these discussions of the ultimate image distract them from shooting pictures and concentrating on what is in those pictures.

[Brent McSharry]

Since all I can get on the paper is 256 shades, if that, my task is to get enough out of my negative to fill the range of the paper

Does anyone know how many shades of grey the eye can differentiate between in a b/w print under standard illumination? This would seem to make a big difference to what capabilities b/w photographers/printers look for in both input and output media.

[Ed Richards]

Brent - Adams and all the other folks working on controllling exposure and development have written on this. You can see Adams, The Print, or for a more modern treatment (with all the technical detail you could want), Way Beyond Monochrome. Bottom line, 256 is about the best case, and that is not under standard illumination - you need a bright, point source light, and, ideally, you are printing on high gloss film. Step back to a mat paper, or turn off the halogen light, and actual visible range contracts a lot. Mostly you create the illusion of range by contrast, such as the great pictures by Paul Strand where you this luminous white object, which is really only luminous because it is against a very dark field, which in turn makes the dark, darker.

You want more dymamic range in the negative to give some room for adjustment, because no one gets the range match exactly right. But the more closely you can match the negative density to the print, the less range you need the scanner to extract from the negative. The BTZS folks really push this for printing, but it is the same for scanning.

OTOH, I am sure I have some negatives that would be printable if I drum scanned them, but are not now. I prefer to concentrate on the ones that are printable. Thus, if I had a few thousand to burn, I would spend it on drum scanning my difficult negatives, not my good ones. Tried some of the good ones, and found that I did not see enough difference to matter.

[Bruce Watson]

Bruce - that is all correct, of course, but isn't that why we do zone stuff to match the negative to the output, even if it is a lower dmax scanner rather than a lower dmax paper? You can certainly pull a lot more range out of a negative with a drum scanner, but isn't the real question how much of that you can put back on paper?

I don't think so. I think the real question is: How do we minimize the degredation of the image at each step of the workflow? If you want a first class print, you have to start with a first class image on film. If you underexpose your B&W negative and you don't record the shadow detail, no scanner is going to retrieve it for you. Therefore it won't be on your final print.

Every part of the workflow degrades the image. I want to do the best I can to minimize the damage at every step. So I try to make the best capture to film I can, the best scan that I can, the least amount of manipulation in Photoshop that I can, and do the best setup of the printer (linearize it), using the inks that give me the widest range and smoothest tonal transitions. Every step in the workflow is important.

IOW, if you don't have the information in the file, you can't put the information on paper.

Since all I can get on the paper is 256 shades, if that, my task is to get enough out of my negative to fill the range of the paper, not to get everything out that is possible. My problem, which is the same as those who print in conventional means, is to do my best to get my negative to match the range of my output media. The consumer scanner will capture more dynamic range than I can get on paper, which is all the really matters, if the negative is right.

I disagree. While individual pixels are a given shade, real images aren't seen by the eye/brain system as bits of individual shades. Most of what we see is the transitions between shades. And this gets you into the sharpness/accutance debates. Human perception seems to be geared toward edge detection. This is certainly true in both vision and hearing. IOW, I don't think discussions about the number of shades we need to use to print have much meaning to how we view a print.

The difference between scanners (even between drum scanners) can be described as a matter of clarity. Say we stand you in a room and put an image on the wall. Then we hang a veil between you and the image. The thinner the veil, the more detail you can see - also the more dynamic range you can see. As the veil gets thinner, the blacks get blacker and the highlights become less blocked. The amount of detail you see also improves.

Scanning puts a veil between you and image. What we are talking about here is how much of a veil. Drum scanning arguably has the least amount of veiling (but it's not zero). CCD scanners create more veiling.

My bigger concern, and one that I think gets lost in this crusade for techical perfection, is that very few memorable images depend on techinical perfection. We can tinker with technology, and buy better technology, but we cannot buy inspiration and timing. Those we only get by shooting a lot of pictures so we are ready when it happens. I do not think this is an issue for Ted or for Bruce, but I read a lot of posts from folks on the list who are new to photography or who have not found their way yet, and I worry that they will let these discussions of the ultimate image distract them from shooting pictures and concentrating on what is in those pictures.

So just because the conversation might confuse newbies we shouldn't have it? I don't believe that, and I don't think you really believe that either. What this discussion should mean to newbies is that they still have much to learn and that it's not as simple as pushing a button and getting everything done for you. There are decisions to be made; education will give them the information they need to make them. Maybe this discussion will inprire some to think and go off searching other sources to learn more, which is always a good thing.

[David_Senesac]

Drum scanners will always cost a lot more than consumer flatbed scanners because of the higher precision of the electronic and mechanical parts and high engineering costs due to the relatively few hardware or software electronic engineers that have a command of the optical engineering that is required. This all forces these products to be expensive low volume products. With just a few photo multiplier tubes, ie for red, green, blue, and luminance, such designers can put a large amount of the design and electronics behind every PMT in order to deliver the best dynamic range and most linear color characteristics by using expensive parts throughout the few sensors. The result is clearly superior especially versus consumer CCD arrays. When scanners are compared, most consumers just think in terms of resolution. It was several years after CCD flatbed or film scanners came out that the manufacturers would even mention dynamic range aka Dmin Dmax. That was because their products just sukked in such specs and they had to be prodded while kicking and screaming several years before they would even blurt the terms out. But there are a lot of other non-linearities involved in color optical characterizations and getting the hardware to reduce all that during the sensing before generating the raw data is the only way that software lookup tables have any chance of getting it close to right.

My higher end consumer flatbed CCD scanner does a fair job of scanning my 4x5 transparencies with resolution and noise a couple notches below drum scanner results. Makes an ok 16x20 print with in the ballpark color fidelity. However trying to get the color hues to critically match the transparency colors beside my computer on the light table is sometimes simply impossible because there are multiple dimensions to the color that if corrected at one hue/frequency are likely to become worse at others. So anytime I want a real scan I send it in for drum services. Its a good thing the current explosion of smaller format digital cameras can't touch view camera sheet film resolution, quality, and cost performance because it guarantees at least another decade of continued need by high end photographers using film for drum scanner services that is not likely to disappear like services for the smaller film formats will. ...David

[Ed Richards]

> I disagree. While individual pixels are a given shade, real images aren't seen by the eye/brain system as bits of individual shades.

Ah, but I am not wiring this into your brain directly. I am putting this on paper, and it does not matter at all how the brain processes that input - whether you start with a Dmax 4.0 ultimate scan or one that barely gets 2.0, they both end up as pixels with limited range on paper and there is no way to tell where they came from before that. There is no history in the information in those pixels.

> If you underexpose your B&W negative and you don't record the shadow detail, no scanner is going to retrieve it for you.

But I do not underexpose my negatives, I fully expose them and I limit development as necessary to keep the highlights from blocking. I make sure I get all the information into the negative. Frankly, I do not even need to do that much on compression because negatives do not develop that much dmax unless you really over develop them. My bet is that expansion of flat scenes is more important.

Most of this dmax discussion would make a lot more sense if we were talking about chromes. So would most of the other stuff about micro contrasts. Color is just a lot harder to do, and really shows the benefits of a great scan.

>So just because the conversation might confuse newbies we shouldn't have it?

Not at all. I do think we should be aware that the absolutist rhetoric can be very disorienting to folks who do not have enough experiece or technical sophistication to keep it in context. Modern film like Tmax, with something like dilute Xtol, competently shot and scanned on an optimized consumer scanner, can produce prints that are TECHNICALLY as good as any AA master print. There are ways to do better in terms of sharpness and maybe even tonality, but for me, just getting to AA standard is good enough while I work on getting to his vision.

[Bruce Watson]

Ah, but I am not wiring this into your brain directly.

Say what? You lost me.

...scanned on an optimized consumer scanner, can produce prints that are TECHNICALLY as good as any AA master print.

You might think so. I've never seen it. Scans made from CCD scanners, particularly consumer scanners like the Epsons, seem too veiled to me, even at small enlargements. I find them lacking, TECHNICALLY, when compared to good darkroom prints, or good drum scans.

But maybe that's just me. Wouldn't be the first time. So if you are happy with your scans, good on ya. Carry on. It should all be about the art, and if you are happy with your art that's all you need.

[Ted Harris]

Bruce, I have to quibble with you regard all scans made on all CCD based scanners. Have you seen scans from the Creo Supreme II, the Screen Cezanne and the very few other comperable high-end flatbeds? For the most part, the differences I have seen at the high end depend far far far more on the operator than on the machine.

[Bruce Watson]

Bruce, I have to quibble with you regard all scans made on all CCD based scanners. Have you seen scans from the Creo Supreme II, the Screen Cezanne and the very few other comperable high-end flatbeds? For the most part, the differences I have seen at the high end depend far far far more on the operator than on the machine.

He didn't say professional flatbeds. He said consumer flatbeds.

I'll agree that the professional flatbeds are quite good these days. For much LF work, a pro flatbed can rival a drum scanner, this is true. And at this level operator skill and dedication plays an ever important roll.

[Ted Harris]

LOL .... Bruce, then, as usual, we are singing fromt he same song book. As for consumer flatbeds your comments are kind. No doubt that they now do a good enough job to get a decent scan for smaller enlargements but I am now starting to do some one on one comparisons of prints from Cezanne scans v. prints from 1800f scans ... smaller prints ... 1x14 and 16x20 and while the differences may sometimes be subtle there is no doubt they are there.