I had an opportunity to test Canon's new affordable 4000 PPI (Pixels Per Inch) film scanner FS4000US for a weekend  prior to decision to buy one (thanks to J. Jussila). Here I present some results from quick tests I made. I was mostly interested in if Canon's version of infrared based dust removal, FARE (Film Automatic Retouching and Enhancement), would perform as advertized, and what impact it would have on image sharpness. I also wanted to test effect of anti-newton glasses in 3mm Gepe mounts I use to decide weather it's worth the trouble to unmount slides for scanning. Unfortunately I cound't locate my test slide for film scanner comparison (found here), so at the moment I cannot directly compare FS4000US to previous film scanners I've used.

First I tried to scan with default settings but soon I noticed that good results from slides with dark parts could only be had when "Colormatching" was disabled from Canon's software FilmGet, and scans were done using 42-bit colours (14-bit per channel). Otherwise dark areas would saturate to black and/or histogram would be spiky to begin with. Because of that, all scans on this page were done with "Colormatching" set to off and 42-bit colours, and all my comments only apply to this combination of settings.

This is fine with me as I prefer doing all adjustments in Adobe Photoshop and with FS4000US I can get full 42-bit image into it. Moreover, Norman Koren's page (see link below) reports problems with spiky histogram also when using 42-bit colours and doing adjustments in FilmGet. And on top of that, if I'm going to be using ICC profile, then I couldn't do any adjustments dependent on the image to be scanned in FilmGet anyway.

 

Figure 1. An example of FARE (set to normal) in action with dusty slide. Images scanned using 2000 PPI and small portions shown 1:1 (100% magnification). FARE is obviously working wonderfully! I also tried using FARE with Kodachrome slides, but as advertised, it is NOT working with them.

Figure 2. Comparison of  effects of FARE and slide mount with anti-newton glasses (3mm Gepe AN) to image sharpness. Images scanned using 4000 PPI resolution and shown here 1:1. FARE was set to normal when used. Levels adjusted and sharpening (Unsharp mask, amount 250%, radius 0.5 pixels, and threshold 0) performed in Adobe Photoshop. Photo of peacock taken with AF Nikkor 85mm f/1.8 lens, aperture not recorded, support used for camera, film Fuji MS 100/1000 @ 100 ISO. Effect of both FARE and anti-newton glass is very small, but noticeable.

 

Figure 3. Parts of same images as above shown at magnification of 300%. Here it is easier to spot the difference. Scan without glasses or FARE is sharpest, but the effect of glasses and FARE is still quite small, even when combined. In upper right image there is rectangular error genereted by FARE.

Figure 4. More similar errors like seen in Figure 3 shown at magnification of 400%. I couldn't find these errors from any other scans I made. At the moment I don't have any idea what might cause these problems. It'd be interesting to know if anyone else has had these kind of problems.

 

Figure 5. Effect of anti-newton glasses seen in 4000 PPI scan of light part of slide. I haven't really studied this problem with film scanners I have used before, but I don't remember noticing anything strange when scanning slides mounted in anti-newton glass frames using 2000 PPI with Polaroid SprintScan 4000. It seems that at least when doing 4000 PPI scans it is better to unmout slides.

 

Figure 6. Noise pattern of FS4000 is quite uniform, with only very slight striping visible. Using 42-bit colours I was able to get very good results from dark or underexposed slides. Manual adjustment of gain resulted easily in blown out whites, so it's not really usable option if the slide has also light areas.

Table 1. Comparison of full 35 mm frame scan times using 42-bit colours. Use of FARE resulted in heavy use of virtual memory with system having only 256 Megabytes of memory. Clearly SCSI and fast processor is needed to achieve acceptable scan speed. Times taken from pressing of "scan" button to image appearing in Adobe Photoshop.

 

 

Conclusion: Based on these quick tests Canon CanoScan FS4000US looks like a quite nice scanner. My test with negatives were also very promising, colours were usually pretty accurate after adjustment of gamma in Adobe Photoshop, something I have never managed to do with any other film scanner I've used.

Before drawing any further conclusions bear in mind that since I don't own this scanner, my experience with it is very limited. Any comments regarding to this scanner test can be sent to Petri.Kekkonen@oulu.fi

For more comprehesive reviews of this scanner see the following links:

• Canon CanoScan FS4000US film scanner first impressions by Norman Koren

• Canon CanoScan FS4000US Film Scanner Review by Taylor Hively

• Steves Digicams - Canon Canoscan FS4000 US - User Review

• Canon FS-4000 Scanner Review, Information, and Specifications from Imaging Resource