Operations > OP_SCANFIX > Virtual Bulb |
The Virtual Bulb process emulates a dropout scanner, converting 24-bit input images to 8-bit output images. Unlike a dropout scanner, multiple hues can be chosen, as well as varying ranges for each hue. Hues are specified by center point, with the range of 0 to 255. 0 indicates pure red, and increasing values move through the spectrum through green, blue, and back to red at 255. The specified hue also includes a radius. The colors of full brightness and a hue within the inner half of the radius will drop to white, and in the outer half of the radius will fade to shades of gray. Colors of less than full brightness (the dominant color component having a value of less than 255) will drop to a shade of gray corresponding to the brightness. To emulate a red bulb dropout scanner, for example, a hue of 0 with a radius of 170 would be used. Unlike a dropout scanner, the algorithm also allows multiple hues to be selected and dropped, as well as allowing an inverted operation that extracts only the selected hue and drops all other colors to white.
Figure 3 A Virtual Bulb drop centered on blue, with a narrow radius converts blue hues to white, all other saturated hues to black. Settings shown, hue 166, radius 40.
For images where the data to be dropped out is not full brightness, the brightness correction option can be used. This ONLY works on images consisting of colored ink on white paper, and it attempts to determine the average level of brightness of the given hue, and will adjust the dropout operation to push those hues up to make the average hue become pure white. Use this option if there are "shadows" of the dropped color left in the output image. This operation does not work on other image types, and will give undefined results. It should also not be used with extraction.
Since variance is common in scanned images, it is possible to specify that a given hue is to be adapted on an image-by-image basis. For example, if dropping red, hue 0, then setting SCANFIX_VBULB_ADAPT will tell the algorithm to analyze the image, and change the value of the hue up to 21 in either direction (1/12 of the spectrum) to match the nearest peak in the hue histogram of the image. This is most useful with narrow drop widths, where a slightly off-center hue can leave an undesirable amount of data on the image. Use of this adaptive feature is slower, however, as it requires analysis of each individual image before processing to find the appropriate peak. A larger radius, if it can safely be used, will provide better performance.