fanart.tv forums

I have noticed that a large proportion of scans performed by other people have tears in them in the form of dropped rows of pixels. They are not rare.

For example,
https://i.imgur.com/KD2Kgca.jpg
https://i.imgur.com/FXFEyWn.jpg

When an image like that is processed with a descreen filter, the missing row disturbs the pattern, is spread out, and gets very noticeable and ugly. What is the cause of this and why is this issue so common?

If a sound card or a CD ripper dropped samples like so, the system would be considered broken and not be used. I have owned 2 scanners CanoScan 4400F and 5600F, and have never encountered this. I am able to scan a CD at 1200 dpi without the scanner stopping. Anything wider than that and the scanner does have to stop, but it is nevertheless able to restart from the precise position. I sometimes see the problem on 300 dpi scans downloaded from the internet.

Does it depend on the drivers? Or the interface method? If this issue comes up, I would like to be able to advise on how to avoid it.

This once happened to me when the scanned object was moved.
This site suggests it can also be related to mechanical problems: http://www.scantips.com/faq1.html
Not sure if the scanner stopping and restarting can cause this, but it does sound likely.

It seems likely that a mechanical problem could cause the motor to skip a step. It would be quite hard to move the object by this small amount on purpose. Once it overcomes friction, it would move more than that. When my computer can't take the data quckly enough, my scanner stops, rewinds a little, and seems to be able to resync perfectly. At least for now. (knock on wood)

I'm not getting relevant results in google search because I'm probably not using the right technical terms.

I recall reading in a guide somewhere that the item should be placed so that its width is minimal, to avoid issues with its internal buffering (or something like that). But in order to minimize visibility of folds in the paper, particularly if the paper is shiny, I find that the object must be placed so that the fold is orientated along with the direction that the sensor travels, which means that the scanner area has a landscape shape for inlay and booklet.