Compact drumscanner

[pjotr]

Hi everyone,

what is the most compact drumscanner that is available(used)?

I remember seeing some very compact units that were vertical desktop drum scanners, during a visit to cebit(computer fair in germany) in the 1990's. But I am unable to find any information about this equipment on the internet today.

thanks,
pjotr

[vinny]

Imacon is likely what you're thinking of but they aren't drum scanners (ccd).

[Lenny Eiger]

The Premier is pretty small, much small than its predecessors, the 4500, etc. I think the 11,000 is pretty small as well...

Lenny

[Daniel Stone]

The "vertical" drum scanner you might have seen would probably have been an ICG, since you mention a European conference. The other "drum scanner"(as it was touted as, but in all reality, it IS NOT A DRUM SCANNER), is the Imacon/Flextight line of scanners from Imacon(now Hasselblad). They were marketed as a "drum scanner", but drum scanners have PMT's, NOT CCD'S.

Assuming you're probably shooting 8x10 or smaller, the Howtek/Aztek 8000 scanners(similar to the Premier(current model for sale from Aztek)) is very small. About 2'x3" or so. And about 165lbs or less. I love mine. Very good results when used properly, and with 8000dpi available, it handles small format(35mm/120) film sizes very well too!

-Dan

[Tyler Boley]

I remember that scanner, I was very interested in it at first but it was expensive. I've seen nothing about it since, it was a drum scanner, very cool looking, European.. very compact and was vertical... whatever it was, it seems to have disappeared.
Tyler

[pjotr]

hi,

thank you so far. I'm pretty sure it was a real drum scanner not an imacon. But I can't remember anything other than it was really compact and vertical.
p

[EOTS]

Heidelberg Tango is also one of the best drum scanners, and vertical.
But I think it is quite bulky or at least heavy nevertheless.

[Lenny Eiger]

This comes down to the micron settings. The Aztek premier, Howtek 8000 and ICG 380 are scanners that can handle scans down to the 3 micron level. This means that all the parts, system wide, have to be to very tight tolerances. (Not unlike a very large piece of glass - a Rodenstock, Schneider or Nikon 12 inch lens, for example that must hold its focus from edge to edge.) The Tango, and other drum scanners on the market are only capable of 6 microns. The result is that the top three are sharper than the rest.

When it comes down to it, these are very small increments. Some images (and a small print size) may or may not not show a difference. I have one image that shows a marked difference at 8x10 inches, however there are many times that I think the difference would not be visible at all. Yet, when using the words "one of the best" these things should be taken into consideration.

The operator becomes the more important factor. I would never send a scan to a lab, or low-cost scanning service, where the person didn't know me personally and know what I was after.

The Tango is very large, spins very fast. One person I know put one in his garage, then had to eventually pour a cement floor to stabilize it. I don't know if this is normal or not, so just take this as anecdotal (and not a fact, per se). However, compact, it is not...

Lenny

[EOTS]

Point taken, Lenny!
Sure there are even better ones around, especially on the new market, like the Premier!

I just noticed the Tango appearing on the used market a couple of times,
remembered that it was vertical,
and heard that the scanner is better resolution-wise and Dmax-wise than f.ex. my Howtek 4500.

Of course I'm not such an experienced and excellent drum scan operator as you are...
But resolution-wise I guess it always depends on the bugdet and what film one want's to scan...
For example my 4x5 color slides, depending on the image (technique, camera stability, wind, yada, yada),
I found that there's often only little information above 2000dpi.
In my cases I do a 4000dpi scan and notice that the 2000dpi scan would have sufficed and the 4000dpi only gives me some additional grain.
On the other hand, in my rare cases of 35mm scans (shot with top Nikon glass) even the 4000dpi showed a lot of detail still available.
Well, the better Dmax would rock in any case I guess ;-)

Best regards,

[Lenny Eiger]

and heard that the scanner is better resolution-wise and Dmax-wise than f.ex. my Howtek 4500.


I would say that the Howtek would have a higher resolution than the Tango. That's been my experience, I had a 4500, and I had occasion to compare a negative scanned on both. It's just one negative, not an exhaustive test, and once again, to be considered anecdotal. However, I wouldn't toss the 4500 just yet....



I found that there's often only little information above 2000dpi.
In my cases I do a 4000dpi scan and notice that the 2000dpi scan would have sufficed and the 4000dpi only gives me some additional grain.
On the other hand, in my rare cases of 35mm scans (shot with top Nikon glass) even the 4000dpi showed a lot of detail still available.
Well, the better Dmax would rock in any case I guess ;-)
Best regards,

Well, you are are correct in that there are many factors, glass being an important one. I have heard this claim many times, usually "after 4000" rather than the 2000 you suggest. I disagree. I think there is a lot more there. One may not see it on the monitor. I was discussing this with Evan from Aztek the other day and this is what he said:

"You had mentioned you changed your monitor to an LCD (Eizo), beaware that
LCD monitors no matter how good they are cannot replicate the appearence of
grain from your scans properly. To be exact an LCD will over sharpen what
is being displayed and if you are doing an aperture check will tend to make
you push softer than you should.

We have an LCD mode in the setting tab in DPL that is meant to try to
correct for this, however this is only a bandaid. Current LCD technolgy is
limited in this fashion the only way around it is to pull out a CRT and
perform aperture checks on it."

This surprised me, I hadn't considered the LCD monitor to be such a factor, but it would explain a few things around here... There are always plenty of new things to learn. I hope you aren't using AutoTrack in your scans, this could be doing things you don't want. If you aren't getting more than 2000 dpi results, its time for a few little things to tune things up, in my opinion.

Best,

Lenny

[georgl]

Don't get lost in technical data too much, focus on condition, software and availabilty instead. The Tango has a small footprint (66x64cm) but is very heavy - 250kg which also indicates the build quality and engineering.

The drum-mount is machined stainless steel, the drums are balanced, the motor is of high-quality - it runs much smoother than a Howtek and many other small drum scanners.

I have recently acquired a Tango, it replaces my Eversmart and Howtek D4000. The dynamic range and resolution is not too different, the Tango propably has the edge - but my test is based on an excellent-condition Howtek-scan with agressive tone curves applied. My own Howtek is not capable of producing similar results, especially because of a "jitter"artifact.
The aperture size is also an overrated criteria, it doesn't exactly relate to resolution (e.g. 10mikron = 2540ppi) but mostly affects the rendering of grain - even on high-res negatives I set the aperture beyond 10mikron. I will make a post on my blog showing a comparison: drumscan.blogspot.de (http://drumscan.blogspot.de)

[8x10 user]

Right... We went through this before. Lenny doesn't even us the 3 micron setting for anything. It's pretty much for leica users who are using advanced technical pan or some type of high resolution surveillance film. The grain structure for LF films is big for those little apertures... All you get is a bunch of color noise from "grain aliasing". IRL the tango has a small enough aperture for any large format use.

The premier is also limited to a D-max of 3.88.

[8x10 user]

The tango does require sturdy floors... Cement is ideal. You can actually put it on a carpeted floor as well, you only need to find a 1" thick sheet of aluminium to put it on ;)

[Lenny Eiger]

The premier is also limited to a D-max of 3.88.

I wouldn't use this to compare anything. In their own comparison, they used 3.4 vs the Tango at 3.0. Imacon said 4.6, which is beyond all film, and is a ridiculous assertion, based on theoretical numbers vs real ones. Aztek's rating was 2.7 for Imacon.

I would simply say that the Premier has the capacity to dig deeply into very dark area of the film. I'm sure the Tango can as well. Anything with a PMT is going to beat the CCD scanners in this regard as the PMT are very sensitive. Trying to make sense out of any of the comparisons is a futile exercise, I think. There is no independent and verified set of numbers to compare with.

If one drum is better than another it is likely a very small amount.

Lenny

[sanking]

The aperture size is also an overrated criteria, it doesn't exactly relate to resolution (e.g. 10mikron = 2540ppi) but mostly affects the rendering of grain - even on high-res negatives I set the aperture beyond 10mikron. I will make a post on my blog showing a comparison: drumscan.blogspot.de (http://drumscan.blogspot.de)


The issue is discussed in Tim Vitale’s article, “Film Grain, Resolution and Fundamental Film Particles”

Vitale writes,

“Art of Drum Scanning -- No Film Grain with Higher Digital Image Resolution Often, a drum scanner operator will choose a pixel pitch (ppi) that is much smaller than the aperture. An example: the operator selects a 12 um aperture (113 um2 round) because it is known to eliminated film grain for the film being scanned, and then scans the image at 4000 ppi pixel pitch, which has an equivalent pixel size of 6.3 um, smaller than the aperture size.
The aperture size is 12 um, but the pixel size is 6.3 um; all detail between 6.3 um and 12 um is lost, but the image file has a large number of pixels based on the 4000 ppi pixel pitch. The PMT sees uniformly mixed light through the 12 um aperture, from the region of the film corresponding to the pixel pitch selected (4000 ppi). The light is measured and converted to digital values, and then applied to each of the 6.3 um pixels (each pixel has 40 um2 area).
The large aperture size (12 um) corresponds to a resolution of 2117 ppi, but the pixel pitch is 4000 ppi. Each of the 6.3- um pixel’s is seeing light from an area about 3 times larger (113 ÷ 40 = 2.8) than the pixels. The greater pixel population created by the denser pixel pitch has had the grain removed by the larger aperture size. This creates a digital image file that will make a large print with no grain. The downside is the image resolution is sacrificed in favor removing film grain. In the language of a flatbed scanner operator: the image has empty pixels. Not that there are clones of parent pixels, but that each pixel was made with an aperture that was 3 times larger than the size of the pixel; different process same concept.”


My question is this. If your film has a resolution of 4000 dpi or higher (which would require an aperture of 6 um), don't you essentially throw away this resolution if you use an aperture of 13 um or higher? Even if you use a ppi pixel pitch of 4000, resolution is still only about 2000 ppi, so the scanner is just interpolating data points. Really no different than if you were rezzing up the image file in PS.

Sandy

[Tyler Boley]

Actually I see it as different... the drum scanner, if set to a resolution higher than that aperture, is actually sampling at the higher rate. It's just that there is more wrap around from each sample, and more overlap from each sample, and that could be stating it badly... I see it more as the difference between a point source enlarger head vs a diffusion enlarger head. But the practical and visual differences between this and uprezzing can really only be determined by careful comparison tests. I've gone back and forth on this since day one of getting a drum scanner with a variable aperture.
Tyler

[Tyler Boley]

actually, reading this, I'm uncertain of the explanation of how the samples are combined, a bit over my head on a holiday. Would be interesting to get Tim into this conversation, he's been very approachable in the past.
Tyler

[tjvitale]

actually, reading this, I'm uncertain of the explanation of how the samples are combined, a bit over my head on a holiday. Would be interesting to get Tim into this conversation, he's been very approachable in the past.
Tyler

I'm online, what is it that needs to be added to the discussion?

Tyler came and got me. I know Lenny and Sandy well.

If you don't match the aperture to the resolution, you loose resolution. There may be a need for the extra pixels, but each one will be collected from an area larger than its footprint. Maybe you need the light; maybe you want to make a very large print; there are legitimate reasons.

The quote Sandy gave is a little verbose, but I do not need to correct anything that was said.

I think the issue of aperture size is quite critical because most people forget about it.

The other issue is that many of the early drum scanners don't have the range of aperture options found in the more recent drum scanners.

Cheers


Tim
510-594-8277

[Tyler Boley]

Hi Tim, I guess the question Sandy posed is of interest. What different might there be between creating more pixels by collecting more with the resolution of the scanner, or "making up" info by up-rezzing methods.
And I don't mean interpolation by the scanner software, I mean higher res mechanical scanner settings, higher than the matched aperture setting... it seems to me it's real sampled information, not made up, even if not necessarily contributing detail.. actually the more I type the the less clear I am...
TYler

[polyglot]

Think of it as "stuff out of focus". You can sample that at higher resolution and not gain (much) more information with some very narrow exceptions.

Think of the output image as being the input image convolved (http://en.wikipedia.org/wiki/Convolution) with the aperture - as the aperture gets bigger, the output gets softer. However if the aperture behaves like a perfect step function, then application of some postprocessing can deconvolve the softening effect to some degree and increase the resolution (not just the acutance) of the image. In the simplest case, an unsharp mask (with radius set to the aperture radius in pixels) applied to an image captured with a large scanner aperture, can restore some of the sharpness. Deconvolution is a messy and imprecise thing because there is some information loss in convolution, but the point is that capturing with a larger aperture does not automatically imply that you're just creating "empty pixels".

Another term worth googling for is superresolution, but that applies only when there are multiple passes and there is jittering between the passes. The maths behind it is based on deconvolution though - you could think of a high-res soft image as being a regular array of slightly-offset lower-resolution images with larger pixels.

[georgl]

I'm preparing a post regarding different resolution and aperture settings for my blog (http://drumscan.blogspot.de). But here is a "sneak peak" ;-) of a Portra 400-scan with different aperture settings (-5 is smallest, so 10µm, 0 is standard and +5 is the biggest setting) @6000ppi:
I used a different upload service this time, I hope everybody can see it:
http://www1.picturepush.com/photo/a/8661279/img/8661279.jpg

[tjvitale]

Seems to me this has been discussed before.

Opening the aperture softens the image, we know the results. I'd make a series of tests on a relevant area of a particular negative (or several representative negatives) with a grain structure you face often and critically observe the affects of "aperture size on output."

Remember that grain is dependent on image density, and image detail [in the specific area of the image] above and beyond the film emulsion itself. Film type is the main factor influencing grain.

Then, grain on a specific piece of film will also be influenced by (a) lens used [resolution limits], (b) f-stop which limits resolution, (c) developing [e.g., (d) grain shaping using Pyro in B&W], (e) hand holding, (f) shutter speed [jiggle] and (h-m) some other factors that slip my mind at the moment. Point is: the "grain" target is moving and relevant to a specific bit of grain structure you need to eliminate, modify or soften.

I have looked at grain under the microscope. It just keeps getting smaller and changing as you increase the magnification. Those Kodak authors had it right, appreciation of grain depends on how you are observing it. And, remember, our visual system and brain are in the mix, we can only resolve at about 270-300-330 ppi or 6-7 lp/mm. Our brains are hardwired to find patterns in everything. Our nature is not that of a machine, but rather a helpful system that facilitates finding order so we can make sense of things. Its a feature, not a bug. We have machines that can be dispassionate and precise if we use them as such.

On the other side of the problem, I have never tested the issue of who has the best resampling math. I believe, however, that it come down to which algorithm "you" like the best, or "best suits" the issue you are trying to overcome. Sometimes, it comes down to which methodology "solved a particular problem in the past" (when you were paying attention to that bit of detail), while most of the time we stay out-of-the-weeds so we can get "other" stuff done.

Adobe has several resampling methods in Photoshop. There must be something written on the topic. Or maybe one of the developers, who is online such as Jeff Scwaee (sp?), or one of the very few others I have seen from time to time, has wisdom to offer on why PS has included one or more of those methods of resampling. Pity about Bruce Fraser, he could have helped with this one!

I always say "do the experiment" and evaluate the results. Tell us if you have time. This is why "I" can be very slow at my day job, I'm always experimenting, until I'm forced to get some "work that pays" out the door.

Cheers

Tim