next up previous contents
Next: Fourrier analysis Up: Programmer guide Previous: Related to data   Contents

Linear predictive analysis

int ARGModel(sig_t *sig, float a, datakind_t kind, int p, vector_t *v, float *s)
Perform generalized linear predictive analysis of order p on signal sig, using the spectral resolution defined by a. Output can be The energy of the filter $\frac{\sigma}{A(z)}$ is returned in s. Note that sig is usually a frame.

int ARConvert(vector_t *v1, datakind_t t1, datakind_t t2, vector_t *v2)
LAR $\leftrightarrow$ PARCOR $\leftrightarrow$ LPC conversions. v1 contains the source coefficients of type t1 which are converted in t2 type coefficients returned in vector v2. v2 is NOT allocated in the function.

int LPCepstrum(vector_t *a, int p, vector_t *c)
computes p cepstral coefficients returned in vector c from the prediction coefficients of vector a.



Guillaume Gravier 2003-05-07