$80,000 scanner vs. $3000 scanner

"Josh" wrote in news:1111031991.313041.39660
@f14g2000cwb.googlegroups.com:

Hello all,

I currently am not happy with my latest scanning solution. My local lab has a digital workflow that includes a Noritsu scanner that can scan at 4000 dpi and costs about $80,000. Clearly nothing I can afford or would ever buy. I cuurently shoot a lot of color negative and TMX/TMY 6×7 negatives as well as 35mm Portra negs and have him scan them. Obviously the guy who owns the scanner speaks highly of its incredible quality. He charges me very little – if I batch about 100 negatives together he will scan them as TIFF’s onto DVD for about $3 per scan. But I find that often the edges of the histogram are not where I want them. My latest project is to scan all of my own wedding negatives – negs I have printed wet and I know there is wonderful detail in the dress, but I get the entire dress in about 250-255 on the curve on his scans.

Sounds like something to talk to him about. Ever thought of using 16-bit?

My question is that if I get the Nikon 9000 would I notice a significant decrease in quality. There’s got to be some reason why that scanner cost $80,000.

The ability to scan 550 slides an hour versus 1 slide every 40 seconds (if you add a slide feeder, three minutes each if you use all the enclosed software) springs to mind.

But I feel I’d be more in control of the process
since I’d be doing it. And for example I could try and try again to tweak a difficult scan and get more out of it myself, rather than having someone else do it – someone who would never really spend the time to get it right no matter what I paid since he is busy (and it’s not his stuff).

Obviously $3000 is nothing to shake a stick at,

It’s also about $1,000 over B & H price. Does that help your decision?

but since I’m pretty
committed to go the digital darkroom route (starting with film still) I’m willing to sell my Jobo and enlarger etc. to help fund it.
Any advice.

-Josh

"Josh" wrote in message

Hello all,

I currently am not happy with my latest scanning solution. My local lab has a digital workflow that includes a Noritsu scanner that can scan at 4000 dpi and costs about $80,000.

Go to rec.photo.equipment.medium.format and ask. There is a scanner guru who has some definitive tests with scanned, real-life outcomes from MF.

I asked him if he could make 16 bit scans but he stated that the machine is set up to make 8 bit scans and possibly some large change would have to be done that would disturb his workflow otherwise. I’m sure his machine makes 16 bit scans. HIs printer prints at 8 bits so he has no need for the extra data.

As to the price being somewhat based on the speed, that makes me feel better as speed makes no difference to me if I have a few slides to scan.

I’ll look at B&H but I thought the price was recently $3000 for the Nikon 9000 but possibly it has fallen as of late.

Thanks

-Josh

I get the Nikon 9000 would I notice a
significant decrease in quality (compared to a drum scanner)

I’ve used a Nikon 8000 for several years for medium format film, same basic scanner as the 9000. I think it does a good job on slide film but I don’t like it that well for negative film. I also had maybe a dozen or so films drum scanned and have compared the files carefully with what my scanner produces. There’s no doubt the drum scans are a bit better but you’ll probably find the difference is not *that* noticeable, especially in mid-tones and highlights for prints up to say 16×20" (with medium format). The drum (mine were done with a Tango) does pick up more shadow detail.

What I’d suggest is that you have a representative sample scanned by someone with the 9000 and with the drum ($3 per scan is amazingly cheap BTW, the drum scans I got cost $50 each) and compare them directly to see if it’s worth the hassle of doing it yourself. I know that since I’ve moved mostly to digital I absolutely *hate* firing up the 8000 now and doing film scans, compared to the simplicity of RAW conversions 🙂 But at any rate there is no subsititute for directly comparing the same film scanned with both systems since the differences are important to some and inconsequential to others. You need to judge for yourself.

Bill

Thanks for your advice. The Noritsu scanner sits on a tabletop and is about the size of two microwave ovens. I’ve only seen it from a distance the few times I’ve been allowed in the "back" of the store but I do not think it’s a drum scanner because I think they sit on the floor. The scan that got me hooked on scanning was an underdeveloped negative I had made in the British Virgin Islands of a sailboat. It was a bear to print wet as it needed a 5.5 multigrade filter and a ton of manipulation and still it wasn’t perfect, and this is after the neg was selenium intensfied. I had him make me a scan ($10 since it was a single frame) and I was amazed at how much information the scanner picked up. I’ve since printed it at 20 x 24 via Mpix.com and am thrilled – this showed me that I can shoot film and be digital from there. It’s all the other scans I’m unhappy with. I’m sure he could do a perfect scan every time but not without a lot of effort and therefore a lot of cost (i.e. surely not $3, I’m sure he "batches" them now thus the diferent curves based on different exposures.) Also I’m disappointed to not be a part of that pert of the process. I figure if I had some areas that I would normaly burn in because they are dense I could do a second scan for the denser areas and then use both scans in photoshop in layers and use the parts from each I want. Perfect example is a wedding dress (my wife didn’t like the 18% gray dress I had picked out for her, just kidding).

Oh and I did look it up, I was wrong – the 9000 lists for only $2000 which makes me feel a lot better!

-Josh

"Josh" wrote in message

Hello all,

My question is that if I get the Nikon 9000 would I notice a significant decrease in quality. There’s got to be some reason why that scanner cost $80,000. But I feel I’d be more in control of the process since I’d be doing it. And for example I could try and try again to tweak a difficult scan and get more out of it myself, rather than having someone else do it – someone who would never really spend the time to get it right no matter what I paid since he is busy (and it’s not his stuff).
-Josh

Josh:
If you go to this link
http://www.marginalsoftware.com/HowtoScan/DiscussionsTone/sc anning_color_negative_film_4.htm you’ll see a Kodacolor II negative scanned with a Scitex EverSmart Supreme (a $55k flatbed scanner) and an LS-8000. Unfortunately, it had to be scanned at a rather low res, but still, IMO, there’s a pretty big difference.

Dane

In , Josh wrote:

My question is that if I get the Nikon 9000 would I notice a significant decrease in quality. There’s got to be some reason why that scanner cost $80,000.

Yes, there is. A couple of reasons, really–the ability to gang scan images is one, but the most compelling reason, in two words, is: dynamic range.

Many people mistakenly believe that you can judge the quality of a scanner by its resolution–a fact manufacturers of consumer scanners have preyed upon, because increasing a scanner’s resolution is easy to do.

However, one of the best benchmarks of a scanner’s quality is its dynamic range–the total range of tones from light to dark the scanner can "see." A scanner with a poor dynamic range "sees" everything above a certain lightness as pure white and everything below a certain darkness as pure black, meaning that subtle detail in hilights and (especially) shadows is lost.

Dynamic range is measured on a logarithmic scale from 0 to 4. Because the scale is not linear, a scanner with a dynamic range of, say, 3.9 can capture far, far more detail in shadows than a scanner with a dynamic range of, for example, 3.7.

Most people consider a dynamic range of 3.5 to 3.7 to be the bare minimum acceptable for professional-quality work. To give you an idea of what that means, most consumer scanners have a dynamic range of about 2.
4. What that means is that most consumer scanners have a really big
problem with rendition of shadow detail, especially with challenging originals, such as underecposed transparencies.

A very high-end drum scanner, such as a $320,000 Linotype Hell drum scanner, has a dynamic range greater than 4.0, which means its dynamic range is superior tothat of transparency film; it is capable of capturing the entire tonal range of film and then some.

Now, a home scanner costing a thousand dollars or so, or a transparency scanner costing a couple thousand dollars, may give you results that are just fine to you much of the time–but the real test is in originals with a great deal of shadow detail, such as a transparency that’s underexposed by a stop or so. With such a challenging original, the high- end scanners really shine.
—
Art, shareware, photography, polyamory, kink:
http://www.xeromag.com/franklin.html

In , Josh wrote:

Thanks for your advice. The Noritsu scanner sits on a tabletop and is about the size of two microwave ovens. I’ve only seen it from a distance the few times I’ve been allowed in the "back" of the store but I do not think it’s a drum scanner because I think they sit on the floor.

There are tabletop drum scanners.

The diffeence is that a drum scanner works by taking the original to be scanned and mounting it to a glass cylinder, or "drum," which then spins at high speed. As it spins, the scanner element–a device called a "photomultiplier tube" or "PMT," scans across the surface of the spinning drum.

Flatbed scanners work by mounting the original on a flat sheet of glass and running a light source and scanner element beneath it; slide scanners work by passing light through a mounted slide and passing a scanner element over the other side of the slide. Both of these types of scanners use a sensing element called a "charge-coupled device," or "CCD. "

CCDs are cheap and easy to manufacture. They are also not terribly sensitive to light, and they have a linear response to light, which reduces their effectiveness in low light situations. PMTs are difficult and expensive to make, but they are far, far more sensitive to light– some PMTs can detect just one single photon!–and they have a logarithmic response curve, making them far more sensitive and discriminating in low light levels. Much of the extra money in a drum scanner is in the PMT, which gives the drum scanner a far wider dynamic range, and far superior detail rendition in shadows.

Compared to an ordinary scan, a drum scan made by a skilled operator is crisper and cleaner, with greater detail; the CCD scan by comparison looks muddy and flat. However, getting the most out of a drum scanner requires a skilled operator–you can’t just stick the original on the drum, press a button, and get a good scan.
—
Art, shareware, photography, polyamory, kink:
http://www.xeromag.com/franklin.html

Wow! I’m amazed that a flatbed scanner has that type of resolution – I know it’s an expensive one but I thought there were limitations as to the transmission throught the glass and film flatness.

That website had a great tutorial on bits and noise that helped me a lot. I figure with a multiple scan option using Hamrick’s VueScan I could minimize noise and tweak out a little more usable dynamic range if I go with the 9000. I don’t mind if it takes me a few hours to get a scan right – remember I used to take a single negative and disappear into the darkroom for 6 hours and come out with a print if it was a difficult one – I just want to be more involved with the process and I figure the 9000 will be an adequate tool.

-Josh

I’ve done some research and found out the Noritsu QSS-32 scanner is LED based so it’s probably just a really good film scanner and not a drum scanner. Thanks for your post

-Josh