The Minolta DiMAGE Scan
Elite 5400 film scanner

A readers review by Mike Nunan

Page 3.

DELTA 3200

First of all, a bit of catch-up. In my last report I mentioned that I had tried scanning a frame of Ilford Delta 3200. What I found in the first instance was that the grain was very prominent indeed, much more so than with the Delta 100 scan. Compared to the LS-30 scan of the 3200 frame, the 5400 version looks like a pointilist painting. This led me to try a few experiments in grain softening. The first thought that struck me was to try a bit of manual de-focus. This sorta works, but it's a bit unpredictable, because you have to move the focus knob by feel. There is no feedback to indicate the amount of de-focus you have applied, and no way of setting the absolute focus position numerically via the software. This is an oversight IMO.

Minolta 5400	Nikon LS-30
Minolta 5400 detail	Nikon LS-30 detail
Minolta 5400 DeFocus	Minolta 5400 Grain Disolver

However, it's not really a major problem, because it turns out that a small amount of Gaussian blur (in the range of 1.2-1.8 pixel radius) produces an image that has very much the character of the LS-30 scan. Applying a bit of USM afterwards restores a bit of snap, but because the blur has spread the grain about a bit, it comes back looking a lot more natural and "analogue". A good result. [The great thing about the 5400 is that it has a lot of resolution to spare -- enough to give you 600ppi on A4 from a full-frame scan, or 450ppi on A3 -- so you can use blur to trade off a bit of spatial resolution for smoothness and tonality.]

HORRIBLY DENSE B&W NEG

For this next test I used a frame of Fuji Neopan 1600, shot at 1600, and badly overexposed. Neopan 1600 is majorly contrasty anyway, unless you use a two-bath or compensating developer such as Emofin. This roll was just given to the local lab without special instructions, and even the good frames are hard as nails with ultra-dense highlights. The frame I used for the test appears on the contact sheet as almost pure white; it's not possible to discern the subject matter. On the LS-30, you get some kind of an image if the manual gain is moved to the extreme left side of its range (-2 stops) in Nikon Scan, but the histogram shows significant highlight clipping. The scene contains trees against sky, and the white area of the sky is "eating" into the branches. There is clipping in the shadows too. I wouldn't dream of making a print out of it, but you can at least see what the shot is of.

Minolta 5400	Nikon LS-30
Minolta 5400 detail	Nikon LS-30 detail

The 5400 has produced a file that I still wouldn't count as printable, but it's much better than the Nikon's effort. The histogram is perfectly smooth and not clipped at either end. However, the scan did take a long time (about 10 minutes) presumably because of the long exposure time that was required to get anything back from such a dense neg. Even though the histogram is not clipped, the shadows appear a bit posterised. I think I've hit the limits of this scanner's abilities with this torture test, which was a bit unfair because I think trying to produce a print from this neg using an enlarger would be a lost cause anyway.

KODACHROME SCANS

Ok, this is where it gets really interesting. The 5400 is *really* good at scanning K-Chrome. Where the Nikon images look like they are part of Picasso's blue period and have shadows that take an early bath, the 5400 comes up with a much more neutral colour balance straight off the bat and seems to be able to get satisfactory shadow detail out of even some of my darkest slides. I really hope I can arrange with Vincent to get some of these images uploaded tomorrow, so you can see first hand what I'm on about. One particularly evil Kodachrome II specimen from 1981 shows my nephew Chris aged approx 18 months, lit in silhouette by a bright window directly behind him. He's wearing a wooly top, and the light has haloed the wool and his hair. Most of the rest of the frame is black, except for a bit of smooth tone on the ceiling of the room, directly above him. Through the loupe, a wood panelled door is just about visible to one side. The LS-30 cannot pick up any detail in the door whatsoever, even if I crank up the gamma in Levels. Doing this does have the effect of posterising the tones on the ceiling however. The 5400 reproduces the door panels perfectly, and with a 1.5 gamma boost I have what I would consider a reasonably printable image. For reference, I also have a scan of this slide that I created on the Minolta stand at the Focus on Imaging exhibition, using a Multi Pro. It does not capture the door panels, and starts to posterise when I apply the 1.5 gamma boost. A seriously impressive result for the 5400.

Minolta 5400	Nikon LS-30
	Nikon LS-30 detail

Another "bete noir" K-chrome II slide in my collection is a scenic shot taken in the Swiss Alps. This shot has the sun full in the frame at the top left hand side, with mountains in the distance and various trees and chalets strewn around in the middle distance. The foreground areas are quite dark, with certain parts in shadow. The LS-30 renders this frame excessively blue (as it seems to do with all K-chromes) and loses almost all the dark foreground details. The 5400 produced a really usable scan, which can tolerate a 1.5 gamma boost without any hint of problems. I haven't tried a print from this scan yet, but I will do tomorrow and I anticipate a great result judging by what I see on the screen. I'm very happy with this.

Minolta 5400	Nikon LS-30
Minolta 5400 detail	Nikon LS-30 detail

Having been impressed by the way the 5400 handled my two worst Kodachromes, I proceeded to try three more recent frames that should pose less difficulty, although they do still contain some quite dark areas. All three shots are of classic (50s) racing cars, taken in paddock garages at last year's Goodwood Revival Meeting. I won't try and describe them in detail, except to say that (a) the LS-30 made a mess of them and (b) the 5400 turns them out beautifully. Any attempt to boost the shadows on the LS-30 scans produces instant posterisation, while the 5400 scans can take a 2.0 gamma boost without breaking up.

I should also mention that grain was very minimal in all the K-Chrome scans, although I didn't try any KR-200. I didn't feel the need to enable the Grain Dissolver with any of the scans.

© Mike Nunan 2003 www.photo-i.co.uk
Please use the Forum to post your questions and views.