FBHALF

Section: User Commands (1)
Updated: 25-Jun-90
Index Return to Main Contents

NAME

fbhalf - halftone a grayscale image

SYNOPSIS

fbhalf [ -fbct<parm> ] [-s<sharpen> ] [ -C<clean> -N<nbr>]
       [ -<type> ] [ size ] < 8bit > 1bit

DESCRIPTION

Halftones the 8 bit images and produces a pbm format output. There are four halftoning algorithms available: Ulichney's "blue noise" dithering, Floyd-Steinberg dithering, Jarvis et al's "constrained averaging", and simple theshholding. The default is blue noise, but constrained averaging works better on line drawings and scanned text (with high values of gamma - near 20). For some images, Floyd-Steinberg gives a tighter, more pleasing pattern than blue noise, but blue noise is more robust and far less likely to produce significant artifacts.

The size argument is a maximum number of pixels. If given, fbhalf will resize the image so the width is the largest multiple of 8 such that the width times the height is less than or equal to the specified size. The default is to produce output the same size as the input.

OPTIONS

-f: Floyd-Steinberg halftoning. Using 7-5-3-1 weights.
-b<noise>: Blue noise, with <noise> percent random weights. As per Robert Ulichney's PhD thesis.
-c<gamma>: constrained-average halftoning. As per Jarvis. Good for line drawings and graphic art with a gamma of about 20.
-t<threshhold>: Floyd-simple threshholding, pixels greater than <threshhold> are set to white, and those less than or equal to <threshhold> are set to black.
<size>: number of pixels, specifies a desired number of output pixels. The actual size used will have an aspect ratio of 1.0 and a width which is the largest possible multiple of 8 so that the image has less than or equal the specified number of pixels. This is a positional argument.
-s<beta>: sharpening, specifies a certain amount of sharpening after resizing but before halftoning. The default is no sharpening.
-C<num>,-N<num>: clean image by flipping isolated pixels. A pixel is isolated if there are fwer than C like colored pixels in the nearby NxN square. This option can get rid of "snow" in line art images halftoned by the Jarvis constrained image halftoning.
-B: face format, as used by Bennet Yee's face program at CMU.
-F: FBM, format (by default, the default). You are guaranteed not to lose information by specifying FBM as the default.
-G: GIF, Compuserve GIF format.
-I: IFF format, interleaved bitmaps (ILBM), used by Amigas.
-P: PBM, Jef Poskanzer's bitmap format.
-S: sun, Sun rasterfiles (not run length encoded).

EXAMPLE

Given an 8bit sun rasterfile, halftone using Blue Noise and sharpen with beta of 8. Output a Sun 1bit rasterfile:

% fbhalf -S -s8 < foo.8bit > foo.1bit

Using Floyd-Steinberg halftoning, convert a color GIF file to a PBM bitmap with about 300,000 pixels. Sharpen slightly with a beta of 2.0:

% clr2gray < foo.gif | fbhalf -P -s2.0 300000 > foo.pbm

BUGS

On the 386i, the sharpening option causes a core dump. As a workaround you can use the fbsharp(1) command prioer to running fbhalf(1).

HISTORY

Copyright (C) 1989,1990 by Michael Mauldin. Permission is granted to use this file in whole or in part for any purpose, educational, recreational or commercial, provided that this copyright notice is retained unchanged. This software is available to all free of charge by anonymous FTP and in the UUNET archives.

25-Jun-90 Michael Mauldin at Carnegie Mellon University: Release 1.0 mlm@cs.cmu.edu
07-Mar-89 Michael L. Mauldin at Carnegie Mellon University: Beta release (version 0.9) mlm@cs.cmu.edu

This document was created by man2html, using the manual pages.
Time: 04:53:25 GMT, August 26, 2022