Canoscan 9950F Am I nuts?

[Kirk Gittings]

After playing with this new scanner (9950f) and comparing it to my 4870, I decided that a big issue related to sharpness on all flatbeds is the noticeable sag in a horizontal 4x5 negative, which plays havok with the level film plane.

So I simply scaned a negative with the scanner on its side. Yes-holding the scanner on its side for ten minutes so there would be no sag and low and behold I have a significantly sharper scan. All I did to fascilitate this was to put a couple of small furniture bumpers on the film holder so the the lid when lowered would pin it in place. Next time I think I will get a long piece of Velcro double sided strap to wrap arould the whole thing to hold the lid securely, so I don't have to hold it.

Would someone else try this and tell me if I am crazy or not. This seems too simple of a solution to a big problem.

This probably puts some odd stress on some gears or tracks and it may be better to tilt it to one edge vs. another and the scanner will probably wear out faster but what the hell. If it works better with a shorter life.

[Kirk Gittings]

I will try this with the 4870 over the weekend and get back to you all.

[Michael Mutmansky]

Kirk,

As I see it, you have to do several things to get the most out of one of these scanners.

The first is you need to determine what the optimal focus depth (distance from the glass) is. This can be done by using shim stock, or thin cardboard and running a sample scan at a few different distances from the glass. You may find that the best distance is right onthe glass, which will be a problem unless you want o wet mount or remove the glass from the scanner.

The second thing you need to do is make a carrier for the scanner that holds the film flat. Some people tape the film onto the carrier, others make a sandwich carrier. There's lots of ways to solve this problem.

There's no reason to be scanning vertically if a better negative carrier will do the same thing. Besides, scanning vertically will not help if the film has an inherent curl, which it seems most of my film coming back from the labs tend to have.

---Michael

[Kirk Gittings]

I was specifically and only refering to 4x5 trans and negs. I get no curl in these but I do get sag. you can see it quite easily with the eye. As a matter of fact in 26 years and some 15000 4x5 sheets, I have never seen 4x5 curl. Maybe you live in a humid environment. I am not a big fan of taping as any unnecessary handling of the film increases the risks of damage to the film. Same thing for wet mounting in oil. I would much rather ruin a cheap scanner than damage my film. One print sold more than pays for any of these $400 to $800 scanners.

I tried it on my 4870 and got the same increases in sharpness. I dismantled the scanner to see if there were any obvious reasons mechanically not to do it and found none. The 4870 runs on a single rail on the right side of the scanner so I tilted it onto the left side. It worked very well with no obvious problems.

I have a local lab testing the same approach.

[Scott Rosenberg]

thanks for the insightful approach, kirk. sounds like a really easy workaround!

would placing the tranny directly on the glass and scanning with the scanner in it's native orientation not also do the trick?

any comments on the 9950 vs 4870? i'm now vacillating between the two after returning my 9900.

thanks,

[Craig Schroeder]

I'm just now getting some time to experiment with my 9950F and have discovered that 120/220 films vary in width. Agfa negatives are just enough wider that they put a tension on the edges that exaggerates the bend in the center of the negative. Some 220 Fuji color negative film I tried last evening was much better and the scans were better, too. The black and white capability is leaving me lukewarm. It could be that the operator needs more experience in the film scanning world, though!

My naiveté on all of this didn't have me prepared for the amount of time that a high-res film scan takes and so far I'm using smaller formats to make the testing time halfway reasonable. It does seems from my limited number of runs that the film flatness issue is a primary factor in the quality trail.

[Kirk Gittings]

You can't place the trans. directly on the glass because the distance from film to the optics is critical and that distance includes the thickness of the holder.

My opinion of the two scanners after much experimentation is this. For the moment I am going to run both scanners, the 4870 for high resoltion 16 bit b&w scans for my artwork (the 9950f has a limit of 10,000 x 30,000 pixels thats a 275 mb 16 bit file which I find a bit limiting ([bad pun!]) and the 9950f for medium quality-high quantity color commercial work on (fresh from the lab) color negatives because the batch scanning interface is superior to the Epson. Better software when it becomes available will solve this pixel limitation I presume. I say "fresh from the lab" negatives because on further investigation I find the Digital Ice of the Epson to be superior to the dust removal of the Canon.

If I were to chose one scanner right now it would be the Epson 4870 with some strong reservations. It is sharper but the batch scanning interface, the auto exposure, frame recognition and color balance is superior in the Canon and I need all that for the commercial work. If Silverfast comes out with software for the Canon it would be more viable for large files. Neither one is perfect, each has its strengths.

[Kirk Gittings]

Check this out:


http://www.photo-i.co.uk/Reviews/interactive/Scanners/Canon_9950F/page_1.htm (http://www.photo-i.co.uk/Reviews/interactive/Scanners/Canon_9950F/page_1.htm)

The above reveiwer has officially given the boot to this Canon 9950f scanner as it is his third one to malfunction.

Mine has worked flawlessly through about 100 scans varing from 50 to 230 mb, but I may be the lucky one.

I would not buy it yet if you were thinking about it.

Go for the Epson 4870.

[Henry Ambrose]

I'm going out on a limb - only speculating - do you think turning the scanner on its side works because in this position the stepper motor and lead screw is "loaded"? In other words does this remove "slop" from the scanner's mechanism? I'm guessing it might.


If film sag is the culprit then some parts of the film would be in sharp focus and others would not. Do you see this in your normally oriented scans? Or overall unsharpness?


You may be seeing an increase in sharpness from a tighter running mechanism rather than the elimination of film sag.

[Kirk Gittings]

I'm just guessing=you may be right, but I don't think so. I would think that slop in the drive mechanism would create some form of banding in the scan.

On close inspection I would say that yes the edges are clearly sharper than the center. Sag is an issue in an enlarger film holders too, but depth of field makes up for it. I don't think that there is an effective depth of field in a scanner and that is why film flatness is so important.

By the way, I have a backgroung in the repair of high tolerance manufacturing machine repair and maintenance (in a pervious life 30 years ago) so I pulled the 4870 apart. There is nothing mechanical to prohibit setting the scanner on its side.

I have now done about 40 scans this way now and it works like a charm.

[Julio Fernandez]

Kirk: A bit late but perhaps not too late: Yes, film curvature is an inherent problem with flatbeds because in dry scanning the film is not secured to a uniform plane. Accordingly the plane of optimal focus of the scanner is at odds with the film plane through much of the area of the film. The reason for the film's curvature or warp has nothing to do with whether it is set vertically or horizontally. It is simply due to an unbalanced construction consisting of the hydrophillic emulsion on one side and a moinsture insensitive layer on the other. The most effective solution to the problem is to restrain the film to a perfectly flat plane which is at the optimal plane of focus of the scanner. You can do that with wet mounting, the same technique used with drum scanners but much easier and practical with flatbeds since the introduction of turnkey kits for all flatbed scanners. Not only you have greater and more uniform sharpness but your slides acquire a brilliance that until now was the preserve of drum scanners, and if that was not enough, it eliminates dust and scratches even from Kodachrome, B&W, etc. In so doing it enhances the image. You would have to spend thousands more to get a scanner that would give you those benefits. To learn more about wet mounting see http://groups.yahoo.com/group/SCANMAX/

[Ed Richards]

> the 9950f has a limit of 10,000 x 30,000 pixels

You must be using the really crummy Canon software, which is esp. terrible for B&W. Get a copy of Vuescan at once and quit wasting your time and turn your scanner back over.:-) Silverfast is also good, but I like the price and update policy for Vuescan (cheap and free). Plus I like to tinker and Vuescan gives you a lot of control. Seriously, depending on how the mechanics are set up, scanning on the side might not matter or might kill it in hurry.

I find that the scanner actually has a little depth of field and that sagging only matters if the film touches the glass and I get Newton's rings. When this is an issue, I use a shim - file folder with a cutout for the negative - under the carrier and it works great. The negative size is nasty - Tmax is perfect, but I had some old Chromes and negatives that I had to trim to fit the holder. Make sure that the sag is not caused by film being a little big for the holder.