A few scanning tips


Threshold is the key for Line art

In Line art mode, every pixel has only two possible values. Every pixel will be either black or white. The Line art Threshold control determines the decision point about brightness determining if the sampled value will be a black dot or will be a white dot. The normal Threshold default is 128 (the midrange of the 8-bit 0..255 range). Image intensity values above the threshold are white pixels, and values below the threshold are black pixels. Adjusting threshold is like a Brightness setting, to determine what is black and what is white.

Microtek ScanWizard 5 converts the histogram graph to be the threshold control when in line art mode.

Umax VistaScan 3.5 uses the Brightness control for line art threshold. MagicScan 4.2 uses the Highlight control for Line art threshold, and both Preview images show the result visually.

Other brands normally do the same. Normally, threshold will be the only control enabled in line art mode, the only one that has any effect, regardless of what it is named. Some show the graph, and some don't, but all show the Preview.

Setting the scanner's Line art Threshold control lower is often required to lighten colored or dirty magazine backgrounds like this one from a magazine. One quick click to lower Threshold from 128 to about 85 will greatly improve this type of scan. Once you've done it a time or two, you won't even have to look. And needless to say, it would change OCR on this document from impossible to fine (hint, hint). The OCR software wants to make characters out of all those black dots, but it can't know what is text and what is noise.

From this with Threshold = 128

to this with Threshold = 85

The meaning is that the page had gray smudges of enough intensity that the line art sampling decided they should be black instead of white. The higher threshold is a higher hurdle, so that this decision remains at white. For examlple, samples at 100 remain white instead of go to black. The images above are 120 dpi for smaller video size here. The original was scanned at 300 dpi to copy on the laser printer for my wife, and she much prefers the copy to the original.

Microtek ScanWizard presents the Threshold control in the form of a histogram with only a threshold black point. The histogram graph is well explained in the A Simple Way pages later. Basically the graph is a barchart that shows the count of every pixel tone value in the image, and how the count of pixels is distributed from Black at 0 on the left end to White at 255 on the right end.

The small clips above were from the top right column of this document, about the worst part. Normally a white background will be a huge pixel count at the far right end of the histogram. A colored or gray backgound has a much lower luminance value. This histogram peak was at 135, which is the much lower tone value of the colored background, and you can already see in ScanWizard's upper right hand thumbnail image that this scan will be quite dark, before you scan. The default Threshold value of 128 picks up quite a lot of the colored background pixels and turns them black, and we can easily see that right here. The left peak is of course the black text, and the large right peak is the background, so it's a very easy decision where the threshold should properly be.

Moving the threshold to the left (grab the black triangle with the mouse and slide it over), back to about 85 (where I have added the black dot marker, well down on that flat floor), gives a very nice scan with plain white background and black text. In Line art, every histogram value to the right of the threshold is necessarily white, and every value to the left is necessarily black, because that is the definition of threshold. So the effect here is to change all tonal values from 85 to 128 from black to white. Moving the threshold is our choice to make, but we have a lot of help to make it. Try it, experiment, it will be very clear and easy.

Copyright © 1997-2010 by Wayne Fulton - All rights are reserved.

Previous Main Next