by Leigh Perry, Sydney, for largeformaphotography.info
Added 8 Dec 2009: Adrian Tyler iQsmart3, Ian Mazursky Howtek 7500, Lenny Eiger Aztek Premier.
[ This page contains a lot of images and will load slowly over a modem connection. ]
Out of curiosity, I initiated a global scan-around. The idea was to circulate a suitable photograph that could be used to assess the relative performance of all the scanners that I could enlist. Comparisons would be made of 'actual-pixels' detailed crops from several areas of the image.
The scans compared at the bottom of this page represent the results of several months of scans contributed by interested scanner owners from several countries. Scanners included so far are alphabetically.
Appeal: If you have a scanner that you'd like to add to the comparison, please email me.
Scans have been contributed by (alphabetically) Tim Atherton, Danny Burk, Rafe Bustin, Charles Cramer / Bill Atkinson / Joseph Holmes, Creo Israel (Avi Bachar & Eyal Harel), Lenny Eiger, Dave Eisenlord, Geoff Eldridge, Eric Fredine, Jeff Grant, Ted Harris, Mike King, Ian Mazursky, Lee Middleton, Michael Mutmansky, Leigh Perry, Daniel Portnoy, Ed Richards, David Rogers, Raúl Sá Dantas, Paul Schilliger, Neil Snape, Leon Strembitsky, Guy Tal, Adrian Tyler, Ernst Vegt.
To kick-start the comparison, I shot a Velvia Quickload with a Rodenstock Apo Sironar S 150/f5.6 at f/19. For maximum sharpness, the actual-pixel crops used for the scanner comparison were all taken from the plane of sharp focus.
With the scans being performed by different operators, this was always going to be a subjective comparison. Therefore there was no attempt to compare colour rendition, which is heavily influenced by the operator and the choice of scanner software. So long as consistent guidelines for the setting of black and white points were followed, however, it is possible to draw useful preliminary conclusions about a scanner's resolution, its ability to make out shadow detail, and its noise characteristics in the shadow areas.
The comparison was based on the assumption that a scanner's potential owner might want to produce a large digital print, so the scan resolution is 2400 ppi. This allows a 10x enlargement when printed at 240 dpi, yielding a 40x50 inch print. Some scanners, such as the Imacon Precision II could not achieve this resolution, so their scans have been interpolated to 2400 ppi from their highest native resolutions.
No sharpening was (knowingly) applied during the scanning process, although it can sometimes be difficult to know what processing is performed by scanner software. After the crops were made, they were sharpened using Photoshop's unsharp masking (USM). The actual-pixel crops required a different sharpening radius and amount for each scanner. Some scanners required quite a lot of sharpening; some required little. In particular, the drum scans were already very sharp and only required a small amount of USM.
The bottom crop is a shadow performance test suggested by Paul Schilliger. It examines how much detail the scanner is able to pull out of deep shadow, and how much noise is in there. To see what was happening in the shadows, I used Photoshop to apply the steep curve that you see on the right. To my eye, some of the scans (such as the SprintScan and the Eversmart) appear to have had noise reduction applied to the shadow areas -- they have a 'smoothed' look. This may be a feature of scanning software, or I could just be wrong (it's happened once before).
The scanning protocol was simple. There was no spotting or cloning of dust particles. Black and white points were set to match the darkest and lightest tones, without any clipping. No colour correction was performed.
Apart from some overall notes above, the results of the scans are presented here for you to draw your own conclusions. Over time, more scans will be added.