Minolta Elite Scan 5400 image artifacts

[cpc]

Here is a 100% crop from a 5400dpi scan with no sharpening (grain dissolver on) with Minolta Dimage Scan latest version (1.1.5 for Win) on Win7. Note the micro lines, both horizontal and vertical, which look like digital artifacts.
sample 100% crop

Same thing with Vuescan (demo version). And same thing at lower resolutions. So it smells like hardware/driver issue. Only got the scanner recently, so a question to other 5400 users: is this normal? Any solutions/suggestions?

 

[lewis44]

Try Grain dissolver off. When I had one, never used it and never had a problem.
Could be something else, but rule that out first.

 

[cpc]

No difference with GD off. It doesn't really look like an optical effect. I wonder if dust/dirt on the sensor can lead to similar artifacts.

 

[aldobonnard]

I am a 5400 user, but these alleged micro-lines are not obvious to me, I cannot quite see them in the sample which you have provided.

 

[cpc]

I am a 5400 user, but these alleged micro-lines are not obvious to me, I cannot quite see them in the sample which you have provided.

Look really close. Small lines, 1 pixel wide, a few pixels long, sometimes dotted. Lots of these.
Do you, by any chance, happen to have a 100% crop from a scan?

 

[Michalm]

I have seen blurred details with ICE turned on on some frames scanned with coolscan V.

 

[j j]

If that is a full-resolution scan at 100% those are tiny. Would they show at all at normal sizes?

Might the scanner be picking up imperfections in the film rather than a software/hardware issue? The emulsion is surely the most delicate item in the chain.

 

[aldobonnard]

Look really close. Small lines, 1 pixel wide, a few pixels long, sometimes dotted. Lots of these.
Do you, by any chance, happen to have a 100% crop from a scan?

Not here. By the way I'm still using XP for compatibility reason with Minolta s/w.
That could have happened to me when the ICE and /or ICE grain reduction was kept on with another scanner (LS5000), and with any scanner when the grain reduction feature from vuescan is kept on.
Bad jpeg compression can also do that sort of artefacts.
Also, in deep shadows, depending on the developer used (BW), it can happen that the scan does not look smooth, viewed at 100%.
Sorry, that may not help you greatly.

What about colour scanning ? Have you tried with and without ICE activated? Same results?

 

[cpc]

If that is a full-resolution scan at 100% those are tiny. Would they show at all at normal sizes?

Might the scanner be picking up imperfections in the film rather than a software/hardware issue? The emulsion is surely the most delicate item in the chain.

Because of the somewhat regular orientation they do affect the downsampled image. There is a feel of slight oversharpening/aliasing in the downscaled image even though no sharpening is applied.
I think emulsion micro-scratches wouldn't appear as perfectly horizontal lines, no?

Not here. By the way I'm still using XP for compatibility reason with Minolta s/w.
That could have happened to me when the ICE and /or ICE grain reduction was kept on with another scanner (LS5000), and with any scanner when the grain reduction feature from vuescan is kept on.
Bad jpeg compression can also do that sort of artefacts.
Also, in deep shadows, depending on the developer used (BW), it can happen that the scan does not look smooth, viewed at 100%.
Sorry, that may not help you greatly.

What about colour scanning ? Have you tried with and without ICE activated? Same results?

They are also seen in the tiff files, so it isn't jpg compression.
Also seen in the color scans, although the color and color noise sort of mask them better. Will give ICE a try, just to exhaust the options.
(This is Tri-X in TMax 1+9, btw, not a very harsh combo.)


I was checking some 100% crops on flickr from different scanners. Occasionaly there are similiar lines, although those might have been overcompressed. Couldn't find 100% crops from the same scanner model though.

I am contemplating opening and cleaning whatever is accessible (the mirror, probably; possibly the sensor).

 

[j j]

A dirty mirror can and does make a difference. When I had a dusty mirror in a scanner it diffused the image and created glowing highlight halos.

For what it's worth I read that the mirror in the 5400 is vertical and so less likely to get coated than a horizontal mirror.

 

[cpc]

So I googled a couple of old reviews of the scanner, both featuring 100% crops.

These are ruined by heavy compression and not very telling.

These have much better compression quality but lower contrast. Nevertheless, I seem to notice similar tiny defects, especially in the first scan (from the Agfacolor slide).