wb_vad.h revision 5d5c3a132bb446ac78a37dfaac24a46cacf0dd73 
1/*
2 ** Copyright 2003-2010, VisualOn, Inc.
3 **
4 ** Licensed under the Apache License, Version 2.0 (the "License");
5 ** you may not use this file except in compliance with the License.
6 ** You may obtain a copy of the License at
7 **
8 **     http://www.apache.org/licenses/LICENSE-2.0
9 **
10 ** Unless required by applicable law or agreed to in writing, software
11 ** distributed under the License is distributed on an "AS IS" BASIS,
12 ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 ** See the License for the specific language governing permissions and
14 ** limitations under the License.
15 */
16
17
18/*-------------------------------------------------------------------*
19 *                         WB_VAD.H                                  *
20 *-------------------------------------------------------------------*
21 * Functions and static memory for Voice Activity Detection.         *
22 *-------------------------------------------------------------------*/
23
24#ifndef __WB_VAD_H__
25#define __WB_VAD_H__
26
27/******************************************************************************
28 *                         INCLUDE FILES
29 ******************************************************************************/
30#include "typedef.h"
31#include "wb_vad_c.h"
32#include "voMem.h"
33
34/******************************************************************************
35 *                         DEFINITION OF DATA TYPES
36 ******************************************************************************/
37
38typedef struct
39{
40    Word16 bckr_est[COMPLEN];              /* background noise estimate                */
41    Word16 ave_level[COMPLEN];             /* averaged input components for stationary */
42    /* estimation                               */
43    Word16 old_level[COMPLEN];             /* input levels of the previous frame       */
44    Word16 sub_level[COMPLEN];             /* input levels calculated at the end of a frame (lookahead)  */
45    Word16 a_data5[F_5TH_CNT][2];          /* memory for the filter bank               */
46    Word16 a_data3[F_3TH_CNT];             /* memory for the filter bank               */
47
48    Word16 burst_count;                    /* counts length of a speech burst          */
49    Word16 hang_count;                     /* hangover counter                         */
50    Word16 stat_count;                     /* stationary counter                       */
51
52    /* Note that each of the following two variables holds 15 flags. Each flag reserves 1 bit of the
53     * variable. The newest flag is in the bit 15 (assuming that LSB is bit 1 and MSB is bit 16). */
54    Word16 vadreg;                         /* flags for intermediate VAD decisions     */
55    Word16 tone_flag;                      /* tone detection flags                     */
56
57    Word16 sp_est_cnt;                     /* counter for speech level estimation      */
58    Word16 sp_max;                         /* maximum level                            */
59    Word16 sp_max_cnt;                     /* counts frames that contains speech       */
60    Word16 speech_level;                   /* estimated speech level                   */
61    Word32 prev_pow_sum;                   /* power of previous frame                  */
62
63} VadVars;
64
65/********************************************************************************
66 *
67 * DECLARATION OF PROTOTYPES
68 ********************************************************************************/
69
70Word16 wb_vad_init(VadVars ** st, VO_MEM_OPERATOR *pMemOP);
71Word16 wb_vad_reset(VadVars * st);
72void wb_vad_exit(VadVars ** st, VO_MEM_OPERATOR *pMemOP);
73void wb_vad_tone_detection(VadVars * st, Word16 p_gain);
74Word16 wb_vad(VadVars * st, Word16 in_buf[]);
75
76#endif  //__WB_VAD_H__
77
78
79