Compared to ....

When ever I do a review, printer or scanner, I get loads of emails asking which is the better unit, HP, Canon, Epson, Microtek or Lexmark (actually come to think of it I have never had any questions about Lexmark or any press releases). As you will know by now I always refer you back to the reviews and ask you to make your own decisions based on what you see. In this section I am going to do a head to head comparison with the Epson 4870 scanner. Canon, with their years of experience in optics should win hands down, but will they?. I have also scanned the same test image with a Nikon 4000 Coolscan, can the Canon really match a dedicated film scanner? I will perform the test with a totally open mind - I am looking forward to this.

The test image shot with a Nikon F4 and 105mm Nikkor lens - Ektachrome.

The 35mm transparency has been placed in the centre of the glass platen on both the Canon and Epson scanners, this should in theory give the best optical performance. To start the comparisons I am going to use 4000dpi (maximum resolution on the Nikon Coolscan).

Nikon Coolscan Detail	Nikon with USM (100, 1.8, 0)
Canon 9950F Detail	Epson 4870 Detail
Canon with USM (200, 1.8, 0)	Epson with USM (200, 1.8, 0)

The above scans were made without any sharpening or colour corrections, the sharpening on the bottom two images was applied in Photoshop CS. Looking at the first results, as expected the Nikon dedicated film scanner runs away with the gold trophy. The Canon scan seems to be suffering with edge flare and lacks warmth in colour. However, looking at the actual image sharpness the Canon is one step ahead of the Epson. Scan times to produce a 62mb file were similar; 2min 1 sec for the Canon and 2 min 12 sec for the Epson

I also tried the same image on both scanners using VueScan, this produced a slightly sharper image on both the Canon and Epson, but the Canon version needs some more tweaking, there were white lines in the scan - these lines are not present in ScanGear. I guess Ed Hamrick has some more work to do on the Canon version.

Canon using VueScan	Epson using VueScan
Strange white lines in the Canon VueScan version - left, (lines are not present using ScanGear)

What is the Dmax factor is one of the many questions I get asked. I did ask Canon what the Dmax was for the 9950F and was told it was about 3.8. However, things are not that simple as there are so many different ways to measure the exact amount. Some manufacturers may give a bigger DMax figure, but their shadows will just run riot, others may give a smaller figure because they feel the noise in the shadows is unacceptable. So to make things really simple I am just going to use the Levels command in Photoshop and see how much detail is being kept in the shadow area.

Canon shadows	Epson Shadows
Canon with Levels applied	Epson with Levels applied

Both scanners were about the same in this test. I matched the densities prior to scanning, once they were scanned I used Levels in Photoshop to open up the shadow areas. The same values were applied to both scans. The Epson is showing more noise than the Canon. There wasn't a great deal of difference by using 48 bit scans.

Canon 48 bit	Epson 48 bit
Canon 48 bit	Epson 48 bit

Finally, both scanners offer dust removal. Canon has FARE and Epson has Digital ICE. Needless to say that by using the dust removal you are going to extend scanning times. I have used a 2400 dpi setting for this (to keep the time down) Canon produced a full scan in 1 min 10 seconds, the Epson did the same in 2 min 11 seconds (21mb file). To apply the dust removal, Canon used Fare Level 3 set to high and this took 1 min 52 sec. Epson used Digital ICE which was set to ON and it took 7 minutes 36 seconds.

Canon FARE 1 minute 52 seconds	Epson Digital ICE 7 minutes 36 seconds

Both scanners removed the dust and the watermark equally well, but the Canon did it faster.