Minolta DiMAGE Scan Multi Pro 4800dpi 4.8 DMAX

[donald_brewster]

This scanner is evidently now in stock and available from B&H for

around $3500, and includes Digital Ice. Looks like it scans up to

6x9cm. Has anyone out there had a chance to look it over or test

it? Almost a quantum leap over Nikon's and Polaroid's newest MF scan

offerings. Hopefully a lot better than Minolta's last Multi offering.

[scott_eaton]

I'd like to know how Minolta achieved DMAX performance higher than your state of the art drum scanner myself.

 

Assuming that the 4.8 figure is exagerated it would be logical to assume the 4800dpi spec is exagerated as well. I dunno...just how good is this scanner?

[Spearhead]

4800 is interpolated for medium format. It's 3200 in real life.

[miles_hecker]

A scanner can be engineered to see through a higher Dmax by using an A/D converter with more bits and turning up the light source used by the scanner. It's all meaningless of course because the real statistic of interest is the scanners range with real film between Dmin and Dmax.

 

Turning up the light source allows you to see through an incredibly dense emulsion but it blows out the Dmin values and highlights! The real limiting factor of course is the signal to noise ratio of the scan. Here photomultiplier tubes excel and CCD's lose. Their blacks are simply cluttered with noise. However, they've come a long way and my Nikon LS-8000 is so good, I rarely find myself needing to pay for a drum scan these days.

 

Last I heard the highest density slide films only have a Dmax of about 3.8, so I wonder what the Minolta was designed to scan.

 

How does the Minolta stack up? We won't know until someone bites the bullet and buys one. The software however is every bit as important as the hardware and Minolta isn't noted for it's software prowness. For that matter neither is Nikon!

[scott_eaton]

Well, we all expect scanners to hedge their numbers, but the CCD engineers are getting a bit out of line when they are claiming a Dmax of 4.8 from a $3,000 desktop scanner. I mean, where am I going to mount the racks to hold the liquid nitrogen tanks to keep the D/A converter close to absolute zero? The resloution specs are more believable since this is more of a mechanical function. Scitex/Everex did it with their flatbeds.

 

Yeah, you can crank the sensitivity up on the CCD array all you want - down to one photon sensitivity if I remember some of my physics classes right. I ran my old LeafScan out of spec all the time to chew through dense chromes, but the DAC hardly delivered a nice, linear out-put when in that mode.

 

Most of the new, higher end desktop units have just caught up with the Imacon Flextight in terms of Dmax, but my old Howtek drum was all over the Imacon unit. Maybe Heidelberg should just go by the new standards from Nikon and Minolta and rate the Tango at a Dmax of 6.0

[rich815]

I can get a Dmax of 5.4 out of my Minolota Dimage Scan Dual II. Don't ask me how, I just do.

[miles_hecker]

Actually its the CCD that needs to be cooled. It can get quite hot. The new high end digital cameras use a magnesium or aluminum body more to get the heat out of the CCD than to maintain structural integrity.

 

Astronomers actually use liquid nitrogen to cool their CCD sensors and get noise free images of faint stars! I don't think this is practical for a field camera. Some new low noise CMOS sensors such as the Leaf Cmost are incredibly quiet and noise free. I believe they will pass film and drum scans within the next few years.

 

See http://www.fillfactory.com/htm/track_record/htm/cmost_samples.htm

 

However for the $$$ the new LS-8000 and Sprintscan 120 are hard to beat. I think they come very close to the Imacon and 90% of the way to a drum scanner for 25% of the cost.

 

See http://wyofoto.com

 

The home page image on my web site had very dense forground greens and blacks. I think the LS-8000 16x resample scan penetrated them amazingly well.

[squareframe]

I can get Dmax 6.3 from a flashlight and a photodiode. have we lost sight of what it is we are trying to grab hold of, or are we simply spouting off numbers and ignoring the physics? I don't know about you, but I am simply attempting to scan my negatives at both extremes and produce a low-noise, accurate, focused sampling within a 16-bit data space. give me a well-focused, low-noise image plane over these over-inflated film scanner specs from warped film-holders and margarita induced marketing hype.

[pete_andrews]

The dynamic range limit is in the CCD sensor itself, not the A/D converter. These things are stuck with handling a brightness range of around 5000:1 max, before 'blooming' or saturation takes place in the highlights and noise takes its toll on the shadows.<br>This is 3.69D in terms of a density range, and assuming a Dmin of around 0.12, you can theoretically get down to a Dmax of 3.8. Only Kodachrome and Velvia actually possess a Dmax greater than 3.8D, and even they bottom out at around 4.2D, so what earthly use a genuine dynamic range of 4.8D would be, I don't know.<br>However, having said that, there's a trick you can do with multiscanning, whereby you combine two scans with different exposures to give a much wider dynamic range.<br>The bottom few bits of the digital information from a long exposure are combined with the top bits from a shorter exposure, to give an expanded dynamic range. It's a fairly easy technique, at the cost of just doubling the scan time and a bit of clever processing. Provided the CCD has reasonable 'anti-blooming' characteristics, the increase in shadow detail can be dramatic.

[bobatkins]

You don't need liquid nitrogen cooling. I use a CCD camera in the lab and it has a thermoelectric cooler on it. There's no reason you couldn't put a cooled CCD in a scanner. Thermoelectric coolers are pretty cheap and the CCDs in these things aren't huge. I've no idea if anyone does cool the CCD in high end scanners, but there's certainly no technical obstacle to doing so, it's well established technology in the scientific imaging field. It's not quite as effective as liquid nitrogen cooling of course, but a whole hell of a lot easier to use!

[pete_andrews]

I don't think any commercial scanner uses Peltier cooling. The benefit would only be in slightly reduced noise, not increased dynamic range. <br>The heat pump energy has to be dissipated somewhere, which means you could realistically only expect cooling to about 10 degrees below ambient; roughly halving the noise. This is about the same reduction as a 4 times oversample.<br>Besides, makers are reluctant to put an adequate lamp warm-up delay in their software, let alone a 20 minute cool-down delay for the CCD.<p>I'm afraid I don't have any expectation of manufacturers doing the technically 'right thing' anymore, even if they knew what it was. You're lucky if the kit just works to spec these days.

[brucelin]

err... why is everyone arguing over what the theoretical dmax is when the thread is about whether the new minolta scan multi-pro is any good? when i click on the thread, i was expecting to see whether anyone has already gotten a hold of this brand new scanner, and if it really lives up to its specs. maybe y'all should start another thread debating the quality of ccds?

[patrick_drennon]

I like my Polaroid Sprintscan 120. It's a true 4000 dpi. With a

good subject the output is amazing compared to my old Agfa

Duoscan. I also managed to save some poorly developed

transparencies with it, great tool. Dimage is probably at least as

good, if not better.

[struan_gray]

There's an in-depth review on the imaging resource site:

 

http://www.imaging-resource.com/SCAN/DSMP/DSMA.HTM

 

For what it's worth, Imacon have upgraded the Flextight Precision recently, with a bump in Dmax to a 'true' 4.2. They also have a combined flatbed and tranny model based on the same imaging system.

 

Test before buying to avoid disappointment.

[donald_brewster]

Thanks Struan - that was a helpful link. Here is another link I found -- a side by side comparison of the three MF scanners, Nikon, Polaroid, and Minolta. http://www.megapixels.co.uk/merchant/film_scanners_mediumformat_compared.htm

[frank_ward1]

I've been working with the Minolta pro for about a week now and I think it's a solid scanner.

 

First, according to the Minolta tech rep. it is a stitched (not interpolated) 4800 dpi on 120 scans. In other words, it has a 35mm wide section that scans up to 4800. This does two passes over a MF film and stitches the two passes together. Otherwise, you can do 120mm scans faster at 3200 and one pass.

 

My own experience confirms this 4800 dpi strip as I use it to scan my panoramic Noblex negs.

 

As for the software, It has ICE, ROC and GEM and a lot of automated options. I've tried them all with the exception of autofocus. The manual focus system makes so much sense that I don't want to bother with autofocus.

 

I scanned an old slide from Afghanistan and did all the adjustments myself and thought I got a pretty good image. Then I let the machine decide on what to do. Damn, with ICE, GEM, ROC and autocolor and auto exposure it did a much better job than I did. Mountains appeared out of the backlight that I didn't even notice on my scan and can hardly see on the slide.

 

I've been scanning in 16 bit with 16X sampling and 3200 dpi (I am not printing big enough to use 4800 dpi on 120 film at this time and the 200-300 mb files provide all the info I need). The scan times are long on 6X7 color negs. 30 plus minutes gives me a scan that puts a smile on my face. I had been using a friend's flatbed Scitek. No need any more.

 

If I get a chance I'll run a test scan next to another friend's Nikon 8000. My gut feeling is that unless you are printing for Grand Central Station; the Polaroid SS120, Nikon 8000, or the Minolta Pro are all great machines.

 

Frank

 

http://www.culturalvisions.com

[adri_de_groot]

Thank you for participating in the discussion on the Minolta Multi Pro scanner. I have considered buying either this one or the Polaroid Sprintscan 120, or the Nikon 8000. The SS120 doesn�t have ICE, which may come in handy especially on large surfaces like 6x9 film, and I hear all kinds of horror stories on the Nikon (banding, software, installation, etc.). The Minolta unit has gotten excellent reviews.

 

There is one problem however for me, and that is that I have lots of 6x12 and 6x17 images as well. So what to do?

 

Big question: can one conceivably scan a 6x17 in this scanner in 2-3 passes?

 

This coincides somewhat with the question: when you put your medium format images in archival sleeve pages, you cut them like this: three 6x6 images

Two or three 6x7 (depending on page width)

Two 6x9 images

One 6x12

One 6x17

 

So when you scan a 6x6 or a 6x7, do you have to cut the strip into individual frames?

 

Or is there a way to leave them uncut and scan one of them? How much can the holder allow film that is not being scanned to stick out of the scan area? Can it work for 6x17? Or at least 6x12???

 

Thanks.

[adri_de_groot]

http://www.kenrockwell.com/minolta/mp.htm

 

for an excellent review!

 

As to the question whether this scanner can be used for scanning 6x12 and 6x17: yes, you don't have to cut the film to 6x9 size, if that's what you are worried about. The holder can accomodate filmstrips up to 18cm in length (three 6x6, two 2 6x7, one 6x12, one 6x17, and a bunch of 6x4.5) All you have to do is move the film through the holder. A 6x12 is best in two overlapping passes, and 6x17 is ebst in three overlapping passes. Stitchting is done at high magnification, and after cuting off some of the edges of each scanside that will be stitched. The glassholder does not cause Newton rings or noise. Make your own mask and scan X-pan or other 24x70mm film on it!