pixdsplit — disentangle chars from doubles in a stream


pixdsplit [options] [infile]


pixdsplit separates out interwoven characters and doubles from a single input stream. It treats the stream as consisting of successive "pixels", each of which is made up of n unsigned chars followed by m 64-bit doubles, and writes the chars to one output stream and the doubles to another. (n and m are described below for the -# option.) The pixels are read from infile if it is specified; otherwise, pixdsplit reads from the standard input.


The command-line options and their meanings are:

-c charfile

Causes pixdsplit to write the chars to the specified file. If charfile is the empty string, then pixdsplit does not write out the chars. Default is "-", so the chars are written to the standard output.

-d doublefile

Causes pixdsplit to write the doubles to the specified file. If doublefile is the empty string, then pixdsplit does not write out the doubles. Default is ``'', so the doubles are not output.

-# n.m

Defines a pixel to be made up of n unsigned chars and m 64-bit doubles, where n and m are positive integers. Either of n and m (but not both) may be omitted and if m is omitted, then the separating period, too, may be omitted. The variable for which a value was omitted retains its previous value. By default, n = 3 and m = 1, so the input stream is assumed to be in pixd(5) format as produced by the -d option of rt(1).


The command

pixdsplit file.pixd | pix-fb

extracts the pix(5) data from file.pixd, sending it for display to pix-fb(1), while

pixdsplit -# 3.2 -c '' -d coords datafile

ignores the RGB information in datafile but writes its pairs of doubles to coords.


rt(1), pix(5), pixd(5)


Paul Tanenbaum


This software is Copyright (c) 1996-2016 by the United States Government as represented by U.S. Army Research Laboratory.


Reports of bugs or problems should be submitted via electronic mail to <>.