libjulius/src/wav2mfcc.c File Reference

Calculate feature vector (MFCC) sequence (non on-the-fly ver. More...

#include <julius/julius.h>
#include <sys/stat.h>

Go to the source code of this file.

Functions
boolean	wav2mfcc (SP16 speech[], int speechlen, Recog *recog)
	Extract MFCC parameters with sentence CMN from given waveform.

---

Detailed Description

Calculate feature vector (MFCC) sequence (non on-the-fly ver.

)

Parameter vector sequence extraction of input speech is done here. The supported parameter is MFCC, with any combination of all the qualifiers in HTK: _0, _E, _D, _A, _Z, _N. Acoustic model for recognition should be trained with the same parameter type. You can specify other parameters such as window size, frame shift, high/low frequency cut-off via runtime options. At startup, Julius will check for the parameter types of acoustic model if it conforms the limitation, and determine whether other additional processing is needed such as Cepstral Mean Normalization.

Functions below are used to convert fully buffered whole sentence utterance, and typically used for audio file input. When input is concurrently processed with recognition process at 1st pass, in case of microphone input, the MFCC computation will be done within functions in realtime-1stpass.c instead of these.

Author:

Akinobu Lee

Date:

Sun Sep 18 19:40:34 2005

Revision

1.2

Definition in file wav2mfcc.c.

---

Function Documentation

boolean wav2mfcc	(	SP16	speech[],
		int	speechlen,
		Recog *	recog
	)

Extract MFCC parameters with sentence CMN from given waveform.

Parameters will be computed for each MFCC calculation instance in the engine instance, and stored in mfcc->param for each.

Parameters:

	speech	[in] buffer of speech waveform
	speechlen	[in] length of speech in samples
	recog	[in] engine instance

Returns:

TRUE on success, FALSE on error.

Definition at line 85 of file wav2mfcc.c.

Referenced by j_recognize_stream_core().

Here is the caller graph for this function:

---