gmm.c

julius version 4.12.about sound recognition.

  for(d=gmm->start;d;d=d->next) {
    if (d->state_num > 3) {
      jlog("ERROR: gmm_init: more than three states (one output state) defined in GMM [%s]\n", d->name);
      return FALSE;
    }
  }

  /* check if CMN needed */

  /* allocate work area */
  if (recog->gc == NULL) {
    gc = (GMMCalc *)mymalloc(sizeof(GMMCalc));
    recog->gc = gc;
  } else {
    gc = recog->gc;
  }
  
  /* allocate buffers */
  gc->gmm_score = (LOGPROB *)mymalloc(sizeof(LOGPROB) * gmm->totalhmmnum);

#ifdef GMM_VAD
  gc->nframe = recog->jconf->detect.gmm_margin;
  gc->rates = (LOGPROB *)mymalloc(sizeof(LOGPROB) * gc->nframe);
#endif

  gc->is_voice = (boolean *)mymalloc(sizeof(boolean) * gmm->totalhmmnum);
  i = 0;
  if (recog->jconf->reject.gmm_reject_cmn_string) {
    for(d=recog->gmm->start;d;d=d->next) {
      if (strstr(recog->jconf->reject.gmm_reject_cmn_string, d->name)) {
	gc->is_voice[i] = FALSE;
      } else {
	gc->is_voice[i] = TRUE;
      }
      i++;
    }
  } else {
    for(d=recog->gmm->start;d;d=d->next) {
      gc->is_voice[i] = TRUE;
      i++;
    }
  }

  /* initialize work area */
  gc->OP_nstream = gmm->opt.stream_info.num;
  for(i=0;i<gc->OP_nstream;i++) {
    gc->OP_veclen_stream[i] = gmm->opt.stream_info.vsize[i];
  }
  gmm_gprune_safe_init(gc, gmm, recog->jconf->reject.gmm_gprune_num);

  /* check if variances are inversed */
  if (!gmm->variance_inversed) {
    /* here, inverse all variance values for faster computation */
    htk_hmm_inverse_variances(gmm);
    gmm->variance_inversed = TRUE;
  }

  return TRUE;
}

/** 
 * <JA>
 * GMM纷换のための洁洒を乖なう. １掐蜗倡幌ごとに钙ばれる. 
 * 
 * @param recog [i/o] エンジンインスタンス
 * </JA>
 * <EN>
 * Prepare for the next GMM computation.  This will be called just before
 * an input begins.
 * 
 * @param recog [i/o] engine instance
 * </EN>
 * 
 * @callgraph
 * @callergraph
 */
void
gmm_prepare(Recog *recog)
{
  HTK_HMM_Data *d;
  int i;

  /* initialize score buffer and frame count */
  i = 0;
  for(d=recog->gmm->start;d;d=d->next) {
    recog->gc->gmm_score[i] = 0.0;
    i++;
  }
#ifdef GMM_VAD
  for(i=0;i<recog->gc->nframe;i++) recog->gc->rates[i] = 0.0;
  recog->gc->framep = 0;
  recog->gc->filled = FALSE;
  recog->gc->in_voice = FALSE;
#endif

  recog->gc->framecount = 0;

#ifdef GMM_VAD_DEBUG
  printf("GMM_VAD: init\n");
#endif
}

/** 
 * <JA>
 * 涂えられた掐蜗ベクトル误惧のあるフレ〖ムについて·链GMMのスコアを纷换し·
 * 纷换冯蔡を gmm_score に姥换する. 
 *
 * GMM_VAD 年盗箕は·稿で VAD 冉年するために·册殿 jconf->detect.gmm_margin
 * フレ〖ム尸の VAD スコア ∈不兰GMMの呵络スコア - 花不GMMの呵络スコア∷が
 * 瘦赂される. 
 * 
 * @param recog [i/o] エンジンインスタンス
 * </JA>
 * <EN>
 * Compute output probabilities of all GMM for a given input vector, and
 * accumulate the results to the gmm_score buffer.
 *
 * When GMM_VAD is defined, VAD scores,
 * "(maximum score of speech GMMs) - (maximum score of noise GMMs)" of
 * last frames (jconf->detect.gmm_margin) will be stored for later VAD
 * decision.
 * 
 * @param recog [i/o] engine instance
 * </EN>
 *
 * @callgraph
 * @callergraph
 */
void
gmm_proceed(Recog *recog)
{
  HTK_HMM_Data *d;
  GMMCalc *gc;
  int i;
  MFCCCalc *mfcc;
  LOGPROB score;
#ifdef GMM_VAD
  LOGPROB max_n;
  LOGPROB max_v;
#endif

  mfcc = recog->gmmmfcc;
  gc = recog->gc;

  if (!mfcc->valid) return;

  gc->framecount++;

#ifdef GMM_VAD
  max_n = max_v = LOG_ZERO;
#endif

  i = 0;
  for(d=recog->gmm->start;d;d=d->next) {
    score = outprob_state_nocache(gc, mfcc->f, d->s[1], mfcc->param);
    gc->gmm_score[i] += score;
#ifdef GMM_VAD
    if (gc->is_voice[i]) {
      if (max_v < score) max_v = score;
    } else {
      if (max_n < score) max_n = score;
    }
#endif
#ifdef MES
    jlog("DEBUG: [%s: total=%f avg=%f]\n", d->name, gc->gmm_score[i], gc->gmm_score[i] / (float)gc->framecount);
#endif
    i++;
  }
#ifdef GMM_VAD
#ifdef GMM_VAD_DEBUG
  //printf("GMM_VAD: max_v = %f, max_n = %f, rate = %f\n", max_v, max_n, max_v - max_n, gc->framep);
#endif
  /* set rate of this frame */
  gc->rates[gc->framep] = max_v - max_n;
#ifdef GMM_VAD_DEBUG
  printf("GMM_VAD: %f\n", max_v - max_n);
#endif
  /* increment current frame pointer */
  gc->framep++;
  /* if reached end, go to start point */
  if (gc->framep >= gc->nframe) {
    gc->filled = TRUE;
    gc->framep = 0;
  }
#endif
}

/** 
 * <JA>
 * @brief  GMMの纷换を姜位し·冯蔡を叫蜗する. 
 *
 * gmm_proceed() によって芜姥された称フレ〖ムごとのスコアから·
 * 呵络スコアのGMMを疯年する. その祸稿澄唯に答づく慨完刨を纷换し
 * 呵姜弄な冯蔡を result_gmm() によって叫蜗する. 
 * 
 * @param recog [i/o] エンジンインスタンス
 * </JA>
 * <EN>
 * @brief  Finish the GMM computation for an input, and output the result.
 *
 * The GMM of the maximum score is finally determined from the accumulated
 * scores computed by gmm_proceed(), and compute the confidence score of the
 * maximum GMM using posterior probability.  Then the result will be output
 * using result_gmm().
 * 
 * @param recog [i/o] engine instance
 * </EN>
 * 
 * @callgraph
 * @callergraph
 */
void
gmm_end(Recog *recog)
{
  HTK_HMM_INFO *gmm;
  LOGPROB *score;
  HTK_HMM_Data *d;
  LOGPROB maxprob;
  HTK_HMM_Data *dmax;
#ifdef CONFIDENCE_MEASURE
  LOGPROB sum;
#endif
  int i;
  int maxid;

  if (recog->gc->framecount == 0) return;

  gmm = recog->gmm;
  score = recog->gc->gmm_score;

  /* get max score */
  i = 0;
  maxprob = LOG_ZERO;
  dmax = NULL;
  maxid = 0;
  for(d=gmm->start;d;d=d->next) {
    if (maxprob < score[i]) {
      dmax = d;
      maxprob = score[i];
      maxid = i;
    }
    i++;
  }
  recog->gc->max_d = dmax;
  recog->gc->max_i = maxid;

#ifdef CONFIDENCE_MEASURE
  /* compute CM */
  sum = 0.0;
  i = 0;
  for(d=gmm->start;d;d=d->next) {
    //sum += pow(10, recog->jconf->annotate.cm_alpha * (score[i] - maxprob));
    sum += pow(10, 0.05 * (score[i] - maxprob));
    i++;
  }
  recog->gc->gmm_max_cm = 1.0 / sum;
#endif
  
  /* output result */
  callback_exec(CALLBACK_RESULT_GMM, recog);

}


/** 
 * <JA>
 * GMMの急侍冯蔡·呵稿の掐蜗が不兰掐蜗として铜跟であったか
 * 痰跟であったかを手す. 
 *
 * @param recog [i/o] エンジンインスタンス
 * 
 * @return 办疤のGMMの叹涟が gmm_reject_cmn_string 柒に痰ければ valid として
 * TRUE, あれば invalid として FALSE を手す. 
 * </JA>
 * <EN>
 * Return whether the last input was valid or invalid, from the result of
 * GMM computation.
 *
 * @param recog [i/o] engine instance
 * 
 * @return TRUE if input is valid, i.e. the name of maximum GMM is not included
 * in gmm_reject_cmn_string, or FALSE if input is invalid, i.e. the name is
 * included in that string.
 * </EN>
 * 
 * @callgraph
 * @callergraph
 */
boolean
gmm_valid_input(Recog *recog)
{
  if (recog->gc->max_d == NULL) return FALSE;
  if (recog->gc->is_voice[recog->gc->max_i]) {
    return TRUE;
  }
  return FALSE;
}

/** 
 * <EN>
 * Free work area used for GMM calculation.
 * </EN>
 * <JA>
 * GMM纷换に脱いたワ〖クエリアを倡庶する. 
 * </JA>
 * 
 * @param recog [i/o] engine instance
 * 
 * @callgraph
 * @callergraph
 * 
 */
void
gmm_free(Recog *recog)
{
  if (recog->gc) {
    free(recog->gc->OP_calced_score);
    free(recog->gc->OP_calced_id);
    free(recog->gc->is_voice);
#ifdef GMM_VAD
    free(recog->gc->rates);
#endif
    free(recog->gc->gmm_score);
    free(recog->gc);
    recog->gc = NULL;
  }
}

#ifdef GMM_VAD

/** 
 * <EN>
 * Compute score of voice activity from the last (jconf->detect.gmm_margin)
 * frames.  Positive value designates speech, and negative means noise.
 * </EN>
 * <JA>
 * 木涟の (jconf->detect.gmm_margin) フレ〖ム尸のスコアから
 * voice activity のスコアを纷换する. 赖の猛は不兰·砷の猛は花不を山す. 
 * </JA>
 * 
 * @param gc [i/o] work area for GMM calculation
 * @param mean_ret [out] mean value of last (jconf->detect.gmm_margin) frames
 * @param var_ret [out] variance of last (jconf->detect.gmm_margin) frames
 * @param count_ret [out] count of speech frames in last (jconf->detect.gmm_margin) frames
 * 
 */
static void
voice_activity_score(GMMCalc *gc, float *mean_ret, float *var_ret, int *count_ret)
{
  int i, len;
  LOGPROB mean;
  LOGPROB var;
  LOGPROB x;
  int count;

  if (!gc->filled) {
    /* cycle buffer not filled yet */
    *mean_ret = 0.0;
    *var_ret = 0.0;
    *count_ret = 0;
    return;
  }

  if (gc->filled) {
    len = gc->nframe;
  } else {
    len = gc->framep;
  }

  mean = 0;
  count = 0;
  for(i=0;i<len;i++) {
    mean += gc->rates[i];
    if (gc->rates[i] > 0.0) count++;
  }
  mean /= (float)len;
  var = 0.0;
  for(i=0;i<len;i++) {
    x = mean - gc->rates[i];
    var += x * x;
  }
  var /= (float)len;

  *mean_ret = mean;
  *var_ret = var;
  *count_ret = count;
}

/** 
 * <EN>
 * Check if trigger of speech / noise segment.  If we are in noise segment
 * and some speech input begins at this frame, recog->gc->up_trigger will
 * be set to TRUE.  If current is in speech segment and it ended at
 * this frame, recog->gc->down_trigger will be set to FALSE.
 * </EN>
 * <JA>
 * 不兰/润不兰惰粗の惰磊りを浮梦する. これまでが润不兰惰粗でこのフレ〖ムで
 * 不兰トリガを浮梦したとき·recog->gc->up_trigger を TRUE にセットする. 附哼
 * 不兰惰粗で惰粗姜位を浮梦したとき·recog->gc->down_trigger を TRUE に
 * セットする. 
 * </JA>
 * 
 * @param recog [i/o] engine instance
 * 
 * @callgraph
 * @callergraph
 */
void
gmm_check_trigger(Recog *recog)
{
  GMMCalc *gc;
  gc = recog->gc;
  float mean;
  float var;
  int count;

  gc->up_trigger = gc->down_trigger = FALSE;

  voice_activity_score(gc, &mean, &var, &count);

  if (gc->in_voice) {
    if (mean <= recog->jconf->detect.gmm_downtrigger_thres) {
      gc->down_trigger = TRUE;
      gc->in_voice = FALSE;
    }
  } else {
    if (mean >= recog->jconf->detect.gmm_uptrigger_thres) {
      gc->up_trigger = TRUE;
      gc->in_voice = TRUE;
    }
  }

#ifdef GMM_VAD_DEBUG
  printf("GMM_VAD: %s: %f %f %d", gc->in_voice ? "VOICE" : "NOISE", mean, var, count);
  if (gc->up_trigger) printf(": BEGIN");
  if (gc->down_trigger) printf(": END");
  printf("\n");
#endif

}
#endif /* GMM_VAD */

/* end of file */