|
|
|
|
Microtek ArtixScan 1800f - Another class of scanner?..!
Talking about scanners, one usually thinks about Epson and Canon, Nikon, Konica/Minolta and some other names, known from their portfolio of digital cameras. MICROTEK do not sell digital SLR cameras like Canon or Nikon or Inkjet printers like Epson, but they offer a pretty comprehensive portfolio of scanners for "prosumers" and professionals. I was able to get their ArtixScan 1800f flatbed scanner for some time on loan to take a closer look, work with it and write about it. When the scanner was delivered by FedEx, I was feeling sorry for the guy who had to carry that large package on his shoulders up to the third floor where I do my work - it is a large scanner, wow! Especially when you compare it to the well known Epson 3200, 4870 and 4990 models - I would guess, the ArtixScan 1800f is about two times the footage and four times their weight. But as soon as you have touched it, you feel, that this is not fragile plastic (nothing against Epson!), not cheap stuff, that is quite some heavy duty equipment. Well it is not cheap, that scanner, but you can get it just under $ 1000.00 (as of July 2005), which is about two times the market price of the Epson 4990. Just to be correct - Microtek also offers a similar, lower priced scanner, the i900 for about half the price and that one is probably the competition to the Epson 4990 and similar models from Canon and others. A short note on working with the ArtixScan1800f and
various operating systems - the scanner works of course in Windows and Mac OS,
but also in Linux OS. Very high Dmax of 4,8 plus 48bit color depth and high real optical resolution So, what makes that ArtixScan 1800f so attractive and justifies its higher price - let me summarize it from my own experience up to now: It is an excellent scanner, no question. Not only it's Dmax of 4,8 and 48bit color depth, which is significantly better than what you get for around $ 500.00, the whole system is very well designed, very robust and delivers consistently excellent scans. Just think about the many other scanners which provide a Dmax of 3,9 or 4,1 or so - the bandwidth of tonality achieved with the ArtixScan is much better due to the high Dmax (maximum optical density). The sensor is a 43,200 elements tri-linear CCD (14400x3) and scans at an optical resolution of 1800x3600ppi - which is a real resolution, not an interpolated one (which goes up to 14400x14400ppi). Multiple sampling scanning comes as standard option with sampling rates up to 16x, resulting in visibly reduced noise, artifacts and signal fluctuation - which in real life means that your scans get clearer details and colors and a less grainy background. Separate work areas for reflective and transparent targets Let me start with the design concept: This flatbed scanner separates the work area for transparent targets from the one for reflective targets and this is a very good move. It is a different task to scan a photograph, printing, newspaper page or similar item than scanning a film negative, a slide or any other transparent target. I would say, to separate these two workflows has got a lot of merits - the film workflow (transparent target) is inside the equipment which makes it less vulnerable to dusting and other negative environmental influences. All transparent targets are loaded into the scanner by means of a "drawer" which contains either masks for various film formats or it is a blank glass plate for all other films, targets and tasks. And that drawer mechanism works well - it fits very precisely, has got a defined position and it's positioning accuracy is very good (I took the drawer with a slide mounted on it out and in and out and in and compared the position of the slide - no noticeable changes at all on the preview scans).
On the photo above one can see the two separate scanning areas - the usual one for reflective targets and the drawer type for transparent targets. As there is no illumination system built into the lid, it is light in weight - another advantage of this concept. As one can imagine, this concept requires a more sophisticated illumination/imaging path than conventional concepts - but the advantage of glass free scanning outweighs the more on mirrors and components for imaging. Impressive scan speed The ArtixScan1800f can be connected to a computer by
either Fire wire (IEEE-1394) or USB2 - I tried both and decided to stay with the
Fire wire (1394) interface. Scanning is done fast, very fast, but - and this one
needs to keep in mind - the computer needs to have enough available capacity to
benefit from the scanner's fast scanning ability. Glass-free scanning Usually the scanning sensor head is positioned underneath
the top glass plate (flatbed scanner) - which means that all scans are done through that
glass-plate. Using a scanner a lot, it can happen, that one makes little
scratches on that top glass plate - well, all scans will be done through those
little scratches, no way out (unless you get the glass plate exchanged). What is shown below is the transparent target tray with the insert for four medium format sized slides/negatives up to 6x9cm. The long vertical slit is required for calibration and the automatic scanner adjustment
The film templates Together with the ArtixScan come several film templates -
plastic trays, fitting into a master frame holder. These s.c. SnapTrans film
templates are for 35mm slides, 35mm film stripes, medium format film stripes up
to 6x9cm and 4x5inch film.
Below is shown the 4x5inch x2 tray in scanning position:
The clear glass holder To provide a flat glass plate with usable area of more than 8x10inch is a very attractive feature! Not only that one can scan 8x10inch film, I take a lot of photos with my 5x7inch Lotus View camera and on such a tray, I can mount two film sheets at once and can scan them one after another. Of course, if you want to scan 4 sheets of 4x5inch film, no problem either. A wonderful device, if - and here comes the big IF, Mr. Newton would not exist. Guess what you can see on almost all scans - Newton rings!! Of course, why not - today's sheets films are usually very glossy on both sides, they do not show anymore the emulsion side as rough and matte, so as soon as you put your sheet of film on the glass plate, I can bet with you, you will see Newton rings - sometimes more, sometimes less, but they are there. The photograph below shows the clear glass holder with a mounted 5x7inch negative:
The Newton rings are unfortunately not only existing, they are also visible on scans as this crop out of a 6x7color slide scan below shows clearly:
Now what to do - one can scan wet, yes wet. I did it and it worked however the differences in wet vs dry scanning are very minute and I am not sure if wet scanning makes sense anyway with the 1800f (unless for the reason of eliminating the Newton Rings) because the 1800f uses that E.D.I.T. technology. Below there is a side by side comparison of a tiny crop from a dry scan with the same crop from a wet scan - it reveals a slightly more intensive color, but is it worth it?
What else can one do - well there is
Anti-Newton spray - it can be bought, but I learned that it is a HazMat item, must be
transported on ground only, must not be used in unventilated rooms etc etc. - so
no real fun either. And spraying my slides and negatives with that spray, I am
not sure if I like to do that or not. Anti Newton Glass - the solution Much better is the following solution - there is the
company
FocalPoint in Florida and they
provide Anti-Newton glass (for scanners, enlargers, printers) inserts for the master frame tray (the one which
accepts the before mentioned SnapTrans templates) and that is such a nice
solution. The AN glass for the 1800f is a standard item, is shipped with UPS 2nd
delivery and fits perfectly into the open frame. One gets a very large scan area - which is even slightly larger than the
one of the clear blank glass plate - and the Newton rings are history. That
glass insert is not cheap (about $ 90.00) but worth the money. Here is a photograph of the Snap Trans Carrier with the Focal Point Anti-Newton glass insert - and this really works well!
Batch Scanning One of the tasks I performed with the 1800f was batch
scanning. I am a very active and passionate photographer and as film is getting
so cheap, especially if one buys slightly out dated film, I tend to shoot a lot
of images. Which means of course, I also have to scan a lot of images as well.
Shooting with a Pentax 645 as example, in 4.5x6cm format on 220 film means 30
images per roll and 10 rolls are done quickly. Here I want to mention that for most of the evaluation and
review, I used the scanning software which comes from Microtek, with the package,
the ScanWizard Pro software. Additionally to that software, Microtek ships the
well known and excellent SilverFast 6i scanning software and a lot of ICC profiles and their data files. I used
ScanWizard Pro as I am of the opinion that if a company provides a scanner, they
also have to provide a proper software with the scanner and not only rely on
third party offerings and having used ScanWizard Pro, I can conclude that it is
a good software, not outstanding but good and working properly. Ok, now back to the task of batch scanning. Drawing a frame around the 15 shots is done quickly and then one should select all of them and pick the most critical one for color fine tuning. After that and if one does not want to prescan and fine-position the frames, batch scanning can start. The batch gets a name, the numbering is automatic and after a short while the batch is done. As mentioned before, it took me about 12 minutes for 15 medium format slides with a size of 33MB each, which is quite fast and very convenient. If one uses the before mentioned scanning templates, then batch scanning is getting even faster as the correct number of frames in the preset sizes can be loaded from the internal data file - just align the frames, correct whatever seems to be necessary and scan, that is all. There are some minor sources of discomfort linked to these preset frames - the preset position is not accurate enough to skip the fine tuning of the location and their orientation needs to be done manually as well. Nothing major, just it is not as fast as it could be if the software would have been evaluated and optimized by someone who works in daily routine with it. The photograph below shows the before mentioned SnapTrans frame with the Antinewtonglass insert and several 4.5x6cm slides ready for a batch scan. The slides have been mounted with a black tape which is usually used for Whiteboards - it sticks enough to hold the slides down and lets itself be removed without traces.
For my work, I would have liked to see preset frames for the most common medium formats, again, as before mentioned - 4.5x6, 6x6 and 6x7 are the most common formats and then comes 6x8 and 6x9 (for which a preset frame exists). I do hope that the Taiwanese manufacturer of the 1800f reads to such suggestions and will provide these improvements which would make scanning with the ArtixScan1800f even more pleasant. Color Correction and Color Management Working with the 1800f does not mean that one is required
to operate in RGB only. The software offers also LCH (lightness, chroma, hue)
and CMYK as color modes and the respective color correction works quite fine and with
a pretty high degree of consistency. Corrections, if wanted or needed to apply after the
built-in automatic color correction has done its job, can be made either on
every single slide or on a batch of slides with a few key strokes or mouse
clicks. All kinds of color profiles are available and almost all common tasks
can be done and achieved. The software does profiling of transparent and
reflective targets as well - it is a bit simple and misses some minor features
like target orientation (no flipping, no mirroring ) but it works and it works
fine and the results are repeatable. Scan results I have scanned color slides and film strips, color negatives and prints in various formats and coming from various film brands and my summarizing conclusion is that the scanner is doing an excellent job. It is of course pretty difficult to reproduce all the fine details and tonalities on a website as the inherent resolution limits deteriorate the visibility of the scans significantly, but I tried to crop small areas out of larger scans to show the performance as good as possible. First comes a comparison of a reflective target scan with its original (as good as it gets) - the original is also a scan and comes from a Kodak Portra 160NC 5x7 inch negative and I took that shot in Colorado at the Druide Hills. That place is an accumulation of weird looking huge monolithic boulders - their size can be estimated by the size of the trees which grow between and on top of them - I scanned the original, printed it on 8x10inch and scanned the print at 300ppi - what you see below is a crop, about 25% of the total image
As you can see, the only more visible differences can be found in the darker areas, which are less detailed in the print scan than in the original scan from the negative - if you look closely, you can see some greenish vegetation growing in these dark areas which are significantly darker in the print and also not that well defined. However, one should not forget that the original file size was about 500MB...so that one can see the finest details is no wonder. I used 300ppi for the print scan because this is a rather standard resolution which most printers and systems can do. Does it get better if I increase the resolution - that depends totally on the print. If the print was done on a normal printing system, then it would not get better and I visually compared the scanned print with the print itself and can confirm, there are no more details in that print (the print came from a professional lab, using Kodak paper and a Noritsu system). What options do exist - to print it at home or in a special printing store where the printer is not limited to 300dpi. How does this turn out - better? More details - well here are the results. The scans were done at 720ppi as were the prints on an Epson inkjet
Well, the difference between direct scan and crop and scan-print-scan and crop has become very small, some contrast issues in the sky can most likely also be solved with some more fine adjustments, but what I wanted to show is that with the 1800f, you can get scans from prints which are excellent and match the original to a very large percentage, I would rate it 9.5 out of 10. ArtixScan1800f - does one need it or should one want it? Of course, everyone confronted with two scanners - one costing around $ 500.00 and the other one about two times that amount is asking the legitimate question - is it worth spending that amount of $$ more to get the ArtixScan1800f? Would the i900 or the Epson 4870 or 4900 or another one in that price range not be sufficient? Do I really see a difference? Well let me speak with images, scans and
tasks instead of data and marketing buzzwords. If you are into film photography
or want to digitize your film based archive, you want to consider the
ArtixScan1800f if your final task is either to get the absolutely best possible
scan results at a still affordable price (if you spend 10x-20x that amount on
money, your scan results are probably slightly better - if you know how to
operate such equipment properly), if you want to go into pretty large prints or
if you want to do it faster than with less expensive scanners.
But for getting really excellent scans, the ArtixScan1800f is certainly a first choice equipment. Does one see the differences - again, it depends on your output size, output media and expectations. I am very picky regarding sharpness, color rendition and overall tonality bandwidth - I used an Epson 3200 Photo and found that the scan lines, which you can see if you look closely enough into tiny details, were too much visible as was the limited Dmax. I also used and still use an Epson 4870 and have to say, it is a very good scanner for what it costs, but if I scan my 4x5inch or 5x7inch negatives and slides (yes the 4870 Photo can scan 5x7inch, but just - ), then I see some shades and details in the negatives or slides, which cannot be reproduced in the scans and this disappoints me as I drag that heavy equipment all over the places to get the best possible image and not to loose some details later during scanning. I show one example, it is a B&W shot, taken at the Taylor Lake in Central Colorado in the afternoon, shortly before a rather impressive thunderstorm began. The camera was the 5x7inch Lotus View, the lens a Schneider-Kreuznach Symmar 210mm, the film speed 100ASA, the exposure time around 1 second with a red filter. Of course I had to downsize the image for web reproduction dramatically - the original file size in 16bit BW is about 340MB
I scanned this negative with the Epson 4870 Photo, 4x sampling at 2400ppi and it took about 1hour to complete the entire multisampling process (and my PC is not slow, it is a 3GHz with 2GB RAM) and then with the ArtixScan1800f, same settings and same multisampling rate - the scanning process took about 5 minutes to complete. Now that is quite a difference!. After scanning I processed both scans the same way in Photoshop CS2, cropped a small part out of the right middle and compared the results after I optimized the histogram on both crops. Please be assured that it was not possible for me to achieve the same visual impression with the cropped 4870Photo scan than with the 1800f - in the upper crop, there were not more shades and details than I show here:
The lower scan is from the 1800f and the upper one from the 4870Photo - in the lower scan, there are more shads of gray, more nuances and details which still can be enlarged. To prove this, I took again the center part out of these images and put them side by side:
Again, the lower image is from the 1800f and the upper one from the 4870Photo. Now one can see what I meant - the scan is at the level of the film grain but in the lower part there is more definition, no white-outs and the overall impression is better for the 1800f scan. Of course, this crop is a very extreme example to show the differences on the web as any more or less normal size of an image would not be different due to the limitations of resolution and size of images on the web and on any monitor. But if you print these scans at maybe 12x18inch, you can see the differences already nicely. And that is what I meant - the ArtixScan delivers visibly better scans, but it depends on what one is planning to do with them if it is a "nice to have" or a "must get". Do these differences also show in color scans - yes they do and it applies the same as I said before for BW scans - it depends on what you are planning to do with the scan and what your quality level expectations are. I have scanned a wildflower meadow on the Kebler Pass (near Crested Butte) - and of course again reduced its file size to match the web requirements (the original file size is about 500MB) - here is the total scan of my 5x7inch Kodak Portra 160 negative with the marked crop area:
Again same procedure - same treatment of both scans, the one from the 4870Photo and the other one from the 1800f. At an 4x6, 5x7 and even 8x10inch print, you won't see a difference, maybe the overall impression is slightly different, but if you enlarge further or print at higher resolution than commercial 1hr photo services provide (300dpi-350dpi), then you can see that the 1800f scan contains more details, more shades of colors and overall makes a more complete impression. Here are the two comparison crops:
The left scan comes from the ArtixScan1800f, the right one from the Epson 4870Photo, both have been scanned using SilverFast 6i as for this rather small details and therefore necessary precise adjustments of the scan parameters, SilverFast is the better software for color critical work and for all those applications where fine tuning of scanning parameters with real-time preview is required - to be clear, all those control functions do of course exist in Scan Wizard Pro, but they are not as refined as in SilverFast. Regarding the two scans above - even at screen resolution one can see that the left scan contains more fine details - look in the center at the larger leaves and you see that they are better defined in the left scan than in the right one. Not much of a difference, but for an enlargement of 16x24, this difference is very visible. A good reason to select the ArtixScan1800f over other scanners at half price, even from the same brand, is scanning process speed. The 1800f scans such a 5x7" negative at 48bit depth with 2200ppi resolution in about 4-5minutes (about 500MB), whereas the very good Epson 4870Photo needs much longer for the same scan. As briefly mentioned before, for color critical work, I used the included version of SilverFast 6i instead of the ScanWizard Pro. In my opinion, the results with SilverFast are better regarding color rendition and it is easier and more accurate regarding color correction than the ScanWizard Pro program. The release of SilverFast 6i, which was included in the scanner, I got on loan, was not the most recent one, but I am sure, that with a new delivery, the most recent release will be supplied as well (Of course, what one always can do is getting an online update of the release from the SilverFast website). Last but not least, another scan of a 5x7inch negative (Kodak Portra 160NC), scanned using the SilverFast 6i into a 550MB TIFF file and for the web compressed into a 350kB JPG file. The result speaks for itself...
If you like to see the details in the left lower corner - click on the image and you will see a mother duck and several young ones ......
The ++'s, The 00's and the --'s What is positive, rather not too positive and what is negative is certainly a question of personal work style, material, one works with and other circumstances, so the list below represents my viewpoint and does not necessarily mean that the majority of users will see it the same or even similar way. The ++ - the impressive speed of
scanning, the high degree of reliability and predictability of scanning results,
the flexibility to be able to accommodate not only all available sizes of
targets up to 11x14" (reflective scan) or 8x10" (transparent targets), but also
the ability to optimize the scanner and its routines to the personal requirements and
necessities, the high real and true optical resolution, the very impressive Dmax
of 4.8 and the overall impression of the scanner as a heavy duty professional
piece of equipment Conclusion The Microtek ArtixScan1800f is a scanner in a
different class of performance compared to the Epson 4870, 4990,
and comparably priced scanners. On top of its better scanning quality which is
visible as soon as one targets larger sized prints at higher resolution, color
critical targets or targets which require a very large dynamic range to properly
reproduce, the most striking advantage is the scanning speed and the
repeatability and predictability of results. If one uses SilverFast as scanning
program, what you see in the prescan is exactly what your scan will look like
and color correction, brightness and contrast adjustments and other optimization
steps work very well and produce scans of outstanding overall quality. At this point I would like to thank the Microtek company for letting me have the scanner for evaluation and review. I do appreciate the efforts they made to make it available despite the fact that there is a large demand for this demo scanner - it was an exercise very worth performing!
|
|
Send mail to
Webmaster with
questions or comments about this web site.
|
