View Full Version : Home Testing Flatbed Scanners
Having gone thru like 12 flatbed scanners in three years (4 Canon 9950's, 1 Epson 3200, 2 4870's, 3 4990's, two Microtek 1800f's) and with seven failing in the first week. I have some suggestions to test yours for manufacturing defects when it comes out of the box.
1)Look for crap, dust and fily residue underneath the glass. You can clean this by disasembling it but you shouldn't have to and if you do you will invalidate your warranty.
2) do a 100% max dpi RGB scan first thing of a neg or trans. that has clear blue or gray skies. Look at the scan at 100% and check for alignment by checking whether the scan is evenly sharp on all edges. If the edges are all sharp but the center is out or vice versa that is generally a film flatness issue which can be dealt with rather than an alignment issue. If the corners are not evenly sharp return this puppy. It can't be fixed by you.
Also look for streaking in the direction of the scan. This will not get better.
Also look for stretching of the grain in the corners and ghosting. Stretching has to do with sloppy tolerence in the belts. Both of these are common in the Epsons and you can live with it if you don't print above 16x20.
3) do the same scan with Digital Ice on and check carefully for banding (banding is a repeated pattern) in the skies. This is the most common place that banding will show up. If you see banding return this puppy. It can't be fixed by you. Do this even if you never intend to use DI. This exposes a hardware flaw.
4) Repeat with DI on 2) but apply a steep curve to the prescan. This will also show more subtle banding problems that may inadvertantly creep up on you trying to scan a flat or underexposed negative. Do this even if you never intend to use DI. This exposes a hardware flaw. Return it.
5) Are the scans less sharp than you expect (your expectations might be too high if you are used to drum scans)? Put some paper shims under the film holder and scan again. Do the corners get sharper? If it does the focus is off. Return it! It is a manufacturing flaw! You can't fix this and you don't want to have to use shims every time you scan. Lay a 4x5 directy on the glass emulsion up and repeat the test. Is it sharper? Return it or keep this one for wet scanning! It is a manufacturing flaw! But you can make it work for you if you wet scan.
This testing allowed me to pick up defects in 7 scanners right out of the box and return them till I got a good one. What do you do to test a scanner?
I think it's pretty obvious what you would say about buying one of these used, on ebay :-)
why don't you ask, just to be sure ... i think Kirk has eleven scanners on ebay right now ; )
Thanks Kirk for sharing your knowledge.
Paul, I know you are teasing but just to be clear. When I say for instance I had 3 4990's I mean. I returned two to get one good one (which is now on the fritz also). I currently am using a 4990, an 1800f and a Nikon 8000. All do something I need.
Kirk- if someone told you you had to keep only one of the three you currently have, which would you keep?
My problem is the diversity of my needs. I need high volume 6x9 color neg batch Digital Ice scans for my business from the 4990, high quality b&w scans for my art from a 4x5 that will do a good 16x20 with no dust correction and the occasional huge enlargement from a 6x9 color for clients from the Nikon.
What I really need is an Imacon 949. But if I could only have one of those three it would be the 4990, but it would be a compromise for all my needs. If you don't need Ice look also at the later production Canon 9950. I have not tested the Microtek i800. It should be a contender too.
Kirk, just set up a new 4990 yestday. These defects, can they be seen on a iMac screen LCD or must they be printed? Thanks for the "heads up". Jim
Kirk - a question about focus. Does the 4990, for example, shift focus from the top surface of the glass (as would be used for prints) to a preset point above the glass when the transparency cover is activated by the scanning software?
I still use the older 3200 Epson (after having used earlier Epson models). Based on my casual comparison of print scans vs. film scans, my assumption has always been that the scanner's focus was preset for prints, and that the film holders were a compromise between sharpness and Newton's rings. I get better film-scan sharpness on my old 3200 by scanning with the film, emulsion side down, directly on the glass. Scanning with the emulsion side to the glass (the reverse of Epson's instructions) I seldom have problems with Newton's rings.
I've been thinking about upgrading to the 4990 for 8x10 film scans, but I'm curious if the focus mechanism is still the same as used in the 3200.
I believe the Epson's have fixed focus so that flat art is technically always slightly out of focus.
For 8x10 you really ought to look at the Microtek 1800f. It is designed to lay the film on the drawer glass emulsion up. It is on the plane of focus from my tests.
Ralph, my 4990 works well with the film emulsion down on the glass. There is a choice in the software for filmholder or not. If I remember correctly its available when you are using the 8x10 film area guide.
> If you don't need Ice look also at the later production Canon 9950.
If you use Vuescan, Ice works OK on the 9950. (Current version, not old versions.) That said, I do not much like Ice on any scanner and prefer to use PS to fix problems. But nobody is paying me by the hour.
> For 8x10 you really ought to look at the Microtek 1800f.
Absolutely, esp. if you shoot Tmax 100 - the emulsion side and reverse are both so smooth you get Newton's rings and you do not want to scan through anti-Newton's ring glass.
The Canon does not have Digital Ice so Vuescan cannot drive it. Canon has something similar called FARE but it is not the same and by all accounts not nearly as good. DI is a Kodak patented product and superior in all my testing and everything I have seen written. It is never perfect but it saves us a ton of time. It works slightly better in some setups than others. For instance on the Epson's it works better with Silverfast than Epsonscan, but it takes allot longer. I assume some superior number crunching is taking place. On the Nikon it works better with the Nikon software and is almost instantaneous.
It may work adequately for him that way, but it is not optimum. The plane of focus on the Epsons is about 1/16 above the glass and is fixed. And it does make a small difference. Try it with a 4x5 on the glass and in the holder. That is why the wet mounting people went over to putting an acrilic sheet down below the negative on these scanners to get the negative up to the plane of focus.
> Canon has something similar called FARE but it is not the same and by all accounts not nearly as good.
I am not sure whose accounts, but since it is software controlled, it depends on the software. The Vuescan support has been improved quite a bit in the last month, so I would not depend on old reports. They are both the same process - you generate an infared scan and subtract it from the visible light scan, with different ways of handing it in the software. I do not much care because it does not work on black and white silver negatives. Since you do 120, it seems like the Nikon would be much better than anything else but a drum scan. Certainly better than the flatbeds.
Unfortunately nothing is simple and the Nikon is a great example. The Nikons do a great scan but.....Triple CCD sensor dedicated film scanners (which all the Nikons are and a few others) have a conflict with Digital Ice which produces banding. So you have to slow the scan down by using only one sensor (the super fine setting) to get DI to work without banding. So one of the benefits of this scanner, speed, takes a big hit. For a 40 MB file it goes from 4 minutes to ten minutes. The 4990 takes 20 minutes for the same scan but...... The Nikon is also slower to load and cannot do a 4up batch scan like the Epson's. With the Epson and Silverfast we batch burn four scans which takes 120 minutes but we have SF write the file to another networked computer. We work on each file as soon as it is written. It is a good workflow. All in all we get a better scan from the Nikon but we have a better and quicker workflow with the Epson.
Also DI is not just software, there is a patented hardware component to. I have no idea what that is. It was originally deveoped by Applied Science Fiction. Kodak got ahold of it when it bought ASF and turned it into their in house digital research branch. That was probably the end of their creativity.
AFAIK, it is basically an infrared channel. The defects and dust distrub/block the infrared and the film is pretty much entirely transparent to it. B&W film and Kodacrhrome film block infrared, so these do not work with ICE. I am sure there are lots of details to work out, but those are the basics of it to my understanding.
Note there are a considerable amount of 'software only' solutions, but these usually soften the shot horribly.
Ralph and Kirk,
I tested my 4990 and I get better scans with film flat on the glass - very sharp and completely satisfactory for an inexpensive scanner. This 4990 is a visible improvement over the other two Epsons I've owned in the last few years. I am fully capable of judging the quality of a scanner's output - I've been scanning since 1992 and I've seen and worked with tens of thousands of files from many different scanners of all types and scanning cameras.
I -know- the machines we are discussing (cheap flatbeds) are somewhat variable and each needs to be tested to see what it does. My last Epson was sharpest -between- the holder and glass, which needed testing to show. If the plane of focus was exactly 1/16" above the glass on all Epsons then they'd all be the same and need the same film mounting procedure, wouldn't they? Isn't this what this thread is about - testing your individual scanner to see what it does?
Now as for Digital Ice and other such - don't use it! Clean your film and spot it manually if you want all the info that's in your film. GIGO. There ain't no magic buttons.
Henry Ambrose wrote:
"Ralph, my 4990 works well with the film emulsion down on the glass. There is a choice in the software for filmholder or not. If I remember correctly its available when you are using the 8x10 film area guide."
I wonder if the 4990 does something to change the plane of focus in the scanner hardware, if there's a choice not to use the film holders in the software?
"Now as for Digital Ice and other such - don't use it! Clean your film and spot it manually if you want all the info that's in your film. GIGO. There ain't no magic buttons.:
Henry, I appreciate the sentiment. I don't use ICE on my personal work, but with 50 images a week to scan for our commercial business, it is a necessity.
I'd like to be able to scan 8 x 10s at home. But everytime I get close to buying the 1800f I read a post like this and it makes me think $900 is a lot of paper. With 3000 sheets of AZO left I guess I will wait a little longer.
I do appreciate the info for when I do buy one!
It depends on how large you aim to print. Henry Ambrose is no slouch and he gets by with his 4990 on 8x10.
"I'd like to be able to scan 8 x 10s at home. But everytime I get close to buying the 1800f I read a post like this and it makes me think $900 is a lot of paper. With 3000 sheets of AZO left I guess I will wait a little longer."
As Kirk noted, quality of the final print is related to magnification factor. If you don't enlarge more than about 3X you should get excellent results from the 4990. One of my scanners is the Epson 4870 and I am enlarging from it up to 3X and the print quality is excellent for 5X7" negatves up to about 14X20". I personally doubt very much that you will see any differnece in quality between this scanner and an Imacon at magnification of 3X or less. I have not, and have carefully compared results with both scanners. Of course, in scanning knowing how to get the most out of the scanner makes a huge difference.
However, for AZO, the question is, how are you going to make your negatives. If you make them on an inkjet printer there is likely to be a loss in quality (in grain appearance ) when compared to an in-camera negative, even of the same size. For alternative processes on art papers, probably not.
On my Mac using the standard Epson software through Photoshop I get these choices:
Film (with Film Holder)
Film (with Area Guide)
Whether or not anything moves when using the last two settings I can't say for sure.
I do know that on the Microteks and the older Agfa versions that the mirror flips over to use either the flat glass on top or the drawer. I'd hope that Epson did not just put a functionless button in to make us monkeys feel better about our purchase. But they could have......................
I am a slouch a good bit of the time. And I've only shot 6 sheets of 8x10 in my life but I do have fine results with 4x5 originals at 11x14 from the 4990 and some pretty good smallish prints from 6x7 and 6x6 film.
I think Sandy's 3X rule is a good one to apply to these scanners - its a result most folks can count on. To exceed it requires great effort, dealing with a load of frustration and the ability to swallow that you won't always get what you want. It sounds like you have that part figured out. I wish I had 50 architecture scans going out each week!!
Thanks for your insight. I have made a few imagesetter negatives for AZO and the results are quite good. I standardized that for grade 3 paper but may change to 2 because I have more grade 2 paper. Fortunately most of my negatives are in camera.
I pulled one of mine apart to clean the inside glass. It looked to me like there is no mechnism to change focus except a general mechnical adjustment for the height during assembly. It relies on depth of field. The mask I believe is just for positioning and the button tells the sofware to look for it there or some crap. There has been allot of discussion about this when it came out on other forums. The plane of focus varies slightly depending on the sobriety of the assembler (or robot?) and that is why people shop these to get a good one sometimes.
Though Epson never says it anywhere, every review I can find refers to it as a fixed focus scanner that focuses 1mm above the glass just as the 3200 and 4870 did.
I don't know if the 4990 is exactly the same as the 4490, but I checked mine with a glass resolution target (several hundred 1951 USAF targets in a 2x2 inch area), tilted at a very slight angle. IMO, the DOF is sufficiently large that there should be no difference between emulsion up or down. The area of best focus was from the surface of the glass to a few film thicknesses above. Now, I understand that people get different results emulsion up vs emulsion down, but I have to wonder if it's due to something other than DOF.
Kirk, I've seen you mentioning that you scan 6x9 in the Epson 4990. Based on the rule-of-thumb saying you can output at 3x with excellent results from this scanner, does this mean your commercial 6x9's are delivered just at 18x27cm (7x10.5")?? Is there a substantial loss in quality if enlarged/printed to 3.5x/4x? (4x is all I need from 6x9 and 3x from 4x5"!)
I got to your forum by searching to see if anyone had developed a holder for 8x10 film for use on the Epson 4990. This is because my main problem has been with Newton Rings when scanning 8x10 B&W Ilford Delta 100 negatives. I switched to scanning emulsuion side down on the glass, and I thought this would solve the problem, but after days of retouching deep inside these immense scans I can still see them on a smaller scale, and they are impossible to retouch out. Years ago I tried wet scanning on a 2450 and all I did was make a big mess, so I am loathe to try it again, and don't want to have to go searching for Kami fluid.
As for shasrpness, I didn't do official tests, but noticed no loss of shaprness scanning directly on the glass. In fact, I would swear that the emulsion side down scans directly on the glass were the sharpest of all.
You got me worried about the scan lines issue, but I never noticed scan lines in my scans with the 4990, and I ALWAYS saw scan lines deep inside the PRO scans I used to make as an operator of an Isomet 455 drum scanner and a Scitex Eversmart Pro Scanner. When printing up to 13x 19 on an R2400 I have noticed lines in midtone sky (not high highlight or deep shadow) areas, but immeditaely attributed these to the printer. I may do some scans in a perpendicular direction to the previous to prove if it is the scanner or the printer, but I have always been convinced that all artifacts in the scan should be noticeable on the screen with a quality graphics card.
When putting 8x10 film down for scanning using the film area guide, I used transparent polyester film tape (just like in my drum scanning days) to secure it. The film will start to bow when heated by the scanner. But I noticed that I could only use the tape on the long side (top and bottom) and not on the left and right. This is because the scanner takes an initial reading of the base emulsion for minimum density on the egde and uses that reading as a reading for the whole row of that portion of the scan, in both preview and scan mode! (Using Epson Scan on Mac OS X at least) Using tape to hold it down on the short side back, I got a lighter band all the way across the scan where I used the tape, which had a different base reading from where the scanner saw the clear film first.
Powered by vBulletin® Version 4.2.1 Copyright © 2013 vBulletin Solutions, Inc. All rights reserved.