NAME

rescale - convert 32-bit float to 16-bit integer sound file

SYNOPSIS

rescale [ -P desired_peak ] [ -p input_peak ] [ -f factor ] [ -r ] [ -t ] [ -s input_skip ] [ -o output_skip ] [ -d duration ] [ -e end_silence ] [ -h ] input_file [ output_file ]

DESCRIPTION

This command-line utility converts a 32-bit floating-point sound file to 16-bit integer. It does this by first scaling every sample by a factor, and then chopping off the portion to the right of the decimal point. (So if the factor is 1, a sample of 22091.428915 becomes 22091.)

There are several options that affect what factor rescale uses, where the output goes, and whether to dither before truncating from 32 bits.

OPTIONS

-P desired_peak

Rescale so that desired_peak is the new peak. Ignored if you also specify factor. This is 32767 by default.

-p input_peak

Specify the peak of the input file. This is taken from the input file header by default.

-f factor

Multiply every sample value by this factor before converting to integer. This is desired_peak / input_peak by default.

-r

Replace input_file with the rescaled version. Ignores output_file. By default, rescale writes to a new file. (See output_file below.)

-t

Use dithering algorithm. This is off by default.

-s input_skip

Skip this many seconds on the input file before reading. This is 0 by default.

-o output_skip

Skip this many seconds on the output file before writing. This is 0 by default.

-d duration

Rescale this many seconds of the input file. This is the entire file by default.

-e end_silence

Write this many seconds of zeros at the end of the output file. This is 0 by default.

-h

Print usage summary.

input_file

The name of the input file, which can be either 32-bit floating point or 16-bit integer in any of the formats understood by RTcmix.

output_file

Write rescaled output to this file, which cannot already exist. Ignored if -r also specified. If neither -r nor output_file given, rescale writes to a new file with the same name as input_file, but with ".rescale'' appended.

This file has the same header format as the input file, as long as that is one of the types that RTcmix can write (AIFF, AIFC, WAVE, NeXT, IRCAM). If it's a type that RTcmix can read, but not write, then the output format will be AIFF.

EXAMPLES

   rescale foo.wav

Assuming "foo.wav'' is a 32-bit float file, this command rescales the file so that its peak amplitude is 32767, and writes the 16-bit output to "foo.wav.rescale.''

   rescale -r -f 1 foo.wav

rescales "foo.wav'' using a factor of 1.0 (i.e., it merely drops digits of precision to the right of the decimal point), and writes the output to "foo.wav,'' overwriting the original 32-bit file.

   rescale -P 20000 -t foo.wav newfoo.wav

rescales "foo.wav'' so that its peak amplitude is 20000, and writes the output to a new file, "newfoo.wav.'' Before truncating to 16 bits, applies the dithering algorithm to each sample.

AUTHORS

John Gibson <johgibso at indiana edu>, based on the original Cmix rescale, but revised and expanded.