Analysis/Synthesis Team,
IRCAM 4 Janvier 96

12.1 Sound file format

For sound file input, Xspect now uses the Audio File Library. Therefore all the format supported by this library are available. In particular, Ircam's sound file format (.sf) is fully supported, short, float, and from any big/little endian machine. Usual AIFF and AIFF-C sound files not compressed are also supported, as well as WAV and SND formats.
Files with other formats can be examined provided that that the correct "Default Raw Sound File Format" is set. A sound file format is defined by the following keywords and corresponding values:

type short/float
rate 'value, e.g. 44100'
byteorder big/little
channels 'number, e.g. 1'
dataoffset 'number of bytes from begining of file to begining of sample data'

The default format is

type short
rate 44100
byteorder 'the byte order of the host'
channels 1
dataoffset 0

The default format can be changed by using the "Raw format default " menu item in the "Context" menu.

Formats can also be specified by giving them in an ascii file as above (with any separator between the different items) and giving that file name after the -fileFormat argument on the command line.


xspect -fileFormat myFormat soundFile.raw

where the file myFormat contains for instance:

type float
rate 44100
byteorder big
channels 1
dataoffset 'size of my header in bytes'