SMOD(1)

NAME

smod - modify intensities in a stream of 16-bit signed short integers

On this page

SYNOPSIS
DESCRIPTION
OPTIONS
EXAMPLES
SEE ALSO
AUTHOR
COPYRIGHT
BUG REPORTS

SYNOPSIS

smod [-a add] [-s sub] [-m mult] [-d div] [-A] [-e exp] [-r root] [file.s] > output.s

smod reads a stream of 16-bit signed short integers, applies a sequence of arithmetic operations to each value, and writes the modified stream to standard output. It is typically used to scale or bias signal data such as that produced by d-i(1) or consumed by the other sig utilities.

Any number of operations may be given and they are applied in the exact order in which they appear on the command line. For efficiency, smod precomputes a lookup table covering every possible input value (-32768 to 32767) before processing the stream.

If an operation produces a result outside the representable range of a 16-bit signed short, the value is clamped to 32767 or -32768. The count of values clipped high and low is reported on standard error when processing finishes.

If file.s is not given, smod reads from standard input. Both input and output must be redirected; smod refuses to read from or write to a terminal.

OPTIONS

-a add: Add add to each value.
-s sub: Subtract sub from each value.
-m mult: Multiply each value by mult.
-d div: Divide each value by div. A divisor of zero is an error.
-A: Replace each value with its absolute value.
-e exp: Raise each value to the power exp.
-r root: Take the root root of each value (equivalent to raising to the power 1/root). A root of zero is an error.
-h, -?: Print a usage message and exit.

EXAMPLES

Example 1. Bias and scale a signal

smod -a 100 -m 2 data.s > scaled.s

Reads short integers from data.s, adds 100 to each value, then doubles the result, writing the modified stream to scaled.s. Operations are applied left to right, so the bias is added before the scaling.

Example 2. Compute signal magnitude in a pipeline

d-i < samples.d | smod -A -m 0.5 > half_magnitude.s

Converts double-precision samples to shorts with d-i(1), takes the absolute value of each sample, and halves it before writing the result. Any value that would exceed the short range is clipped and counted on standard error.