Converts the selected image to a different number of bits per pixel. This option only works on entire images. When working with indexed-color images, it is advisable to use this option to convert them to color images before filtering them (see ``Filtering color images'') or using other operations that assume the image is color. Images can be converted from any depth (8, 16, 24, or 32 bits per pixel) to any other depth.
If the starting image is an 8-bit image with a colormap, the colormap colors are used to select the appropriate colors for the new image. These can be easily changed (see Change colormap ).
If the starting image is a color image, and it is being converted to 8 bits/pixel, the currently-selected palettization method is used to calculate the colors. The default method is ``quantization''. This can be changed (see Color Settings ). 3D images are always converted using the closest-fit method.
Once converted, writing the image to disk will cause it to be saved in its new depth.
Warning: For grayscale images, it is essential to maximize the contrast (``Color .. Contrast .. Maximize Value.'') before changing image depths, converting to color, or most other operations. Otherwise, a 10-bit image converted to an 8-bit grayscale image may appear completely dark, while copying a 10-bit image into a 32-bit image buffer (for example) may cause it to appear with excessive contrast.