Go to the first, previous, next, last section, table of contents.


Speech Tools File Types

The following file types are supported in the speech tools library. The table gives the name of the file type used in the library and the column on the right gives an explanation of this file type.

Track file types

`esps'
Entropic (ESPS) FEA file, used for storing times series of vectors.
`xmg'
File format for CSTR's xmg interactive speech display program.
`xgraph'
File format for X Windows xgraph program.
`snns'
Stuttgart Neural Net Simmulator normal pattern file.
`varsnns'
Stuttgart Neural Net Simmulator variable sized pattern file. This is used for storing variabnle sized patterns for recurrent neural nets etc.
`htk'
Cambridge HTK format.
`ascii'
Ascii data, one frame per line

* although all types (LPC, MFCC_E_D and so on) can be read, only type FBANK is written

Waveform File I/0

`nist'
NIST Sphere waveform format
`esps'
Entropic ESPS SD waveform file format
`snd'
Sun/NeXT waveform file format
`riff'
Microsoft .WAV
`aiff'
Apple form
`audlab'
CSTR Audlab waveform file format
`ulaw'
8K ulaw format, no header (this is officially what audio/basic on the Web).
`raw'
Raw waveform format (no header)


Go to the first, previous, next, last section, table of contents.