cont_adseg.c

WinCE平台上的语音识别程序

/* -*- c-basic-offset: 4; indent-tabs-mode: nil -*- */
/* ====================================================================
 * Copyright (c) 1999-2001 Carnegie Mellon University.  All rights
 * reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 *
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer. 
 *
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in
 *    the documentation and/or other materials provided with the
 *    distribution.
 *
 * This work was supported in part by funding from the Defense Advanced 
 * Research Projects Agency and the National Science Foundation of the 
 * United States of America, and the CMU Sphinx Speech Consortium.
 *
 * THIS SOFTWARE IS PROVIDED BY CARNEGIE MELLON UNIVERSITY ``AS IS'' AND 
 * ANY EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, 
 * THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
 * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL CARNEGIE MELLON UNIVERSITY
 * NOR ITS EMPLOYEES BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT 
 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 
 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 
 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE 
 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 *
 * ====================================================================
 *
 */
/*
 * cont_adseg.c -- Continuously listen and segment input speech into utterances.
 * 
 * HISTORY
 * 
 * 27-Jun-96	M K Ravishankar ([e-mail]) at Carnegie Mellon University
 * 		Created.
 */

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <assert.h>
#include <math.h>

#include "s2types.h"
#include "ad.h"
#include "cont_ad.h"
#include "err.h"

/*
 * Segment raw A/D input data into utterances whenever silence region of given
 * duration is encountered.
 * Utterances are written to files named 0001.raw, 0002.raw, 0003.raw, etc.
 */
int
main(int32 argc, char **argv)
{
    ad_rec_t *ad;
    cont_ad_t *cont;
    int32 k, uttno, ts, uttlen, sps, endsilsamples;
    float endsil;
    int16 buf[4096];
    FILE *fp;
    char file[1024];

    if ((argc != 3) ||
        (sscanf(argv[1], "%d", &sps) != 1) ||
        (sscanf(argv[2], "%f", &endsil) != 1) || (endsil <= 0.0)) {
        E_FATAL("Usage: %s <sampling-rate> <utt-end-sil(sec)>\n", argv[0]);
    }

    /* Convert desired min. inter-utterance silence duration to #samples */
    endsilsamples = (int32) (endsil * sps);

    /* Open raw A/D device */
    if ((ad = ad_open_sps(sps)) == NULL)
        E_FATAL("ad_open_sps(%d) failed\n", sps);

    /* Associate new continuous listening module with opened raw A/D device */
    if ((cont = cont_ad_init(ad, ad_read)) == NULL)
        E_FATAL("cont_ad_init failed\n");

    /* Calibrate continuous listening for background noise/silence level */
    printf("Calibrating ...");
    fflush(stdout);
    ad_start_rec(ad);
    if (cont_ad_calib(cont) < 0)
        printf(" failed\n");
    else
        printf(" done\n");

    /* Forever listen for utterances */
    printf("You may speak now\n");
    fflush(stdout);
    uttno = 0;
    for (;;) {
        /* Wait for beginning of next utterance; for non-silence data */
        while ((k = cont_ad_read(cont, buf, 4096)) == 0);
        if (k < 0)
            E_FATAL("cont_ad_read failed\n");

        /* Non-silence data received; open and write to new logging file */
        uttno++;
        sprintf(file, "%04d.raw", uttno);
        if ((fp = fopen(file, "wb")) == NULL)
            E_FATAL("fopen(%s,wb) failed\n", file);
        fwrite(buf, sizeof(int16), k, fp);
        uttlen = k;
        printf("Utterance %04d, logging to %s\n", uttno, file);

        /* Note current timestamp */
        ts = cont->read_ts;

        /* Read utterance data until a gap of at least 1 sec observed */
        for (;;) {
            if ((k = cont_ad_read(cont, buf, 4096)) < 0)
                E_FATAL("cont_ad_read failed\n");
            if (k == 0) {
                /*
                 * No speech data available; check current timestamp.  End of
                 * utterance if no non-silence data been read for at least 1 sec.
                 */
                if ((cont->read_ts - ts) > endsilsamples)
                    break;
            }
            else {
                /* Note timestamp at the end of most recently read speech data */
                ts = cont->read_ts;
                uttlen += k;
                fwrite(buf, sizeof(int16), k, fp);
            }
        }
        fclose(fp);

        printf("\tUtterance %04d = %d samples (%.1fsec)\n\n",
               uttno, uttlen, (double) uttlen / (double) sps);
    }

    ad_stop_rec(ad);
    cont_ad_close(cont);
    ad_close(ad);
    return 0;
}