C++ ps_start_utt函数代码示例

int main(int argc, char *argv[]) {
    ps_decoder_t *ps;
    cmd_ln_t *config;
    FILE *fh;
    char const *hyp, *uttid;
    int16 buf[512];
    int rv;
    int32 score;

    config = cmd_ln_init(NULL, ps_args(), TRUE,
            "-hmm",    MODELDIR "/en-us/en-us",
            "-lm",     MODELDIR "/en-us/en-us.lm.bin",
            "-dict",   MODELDIR "/en-us/cmudict-en-us.dict",
            NULL);

    if (config == NULL) {
        fprintf(stderr, "Failed to create config object, see log for details\n");
        return -1;
    }
    
    // Initialize pocketsphinx
    ps = ps_init(config);
    if (ps == NULL) {
        fprintf(stderr, "Failed to create recognizer, see log for details\n");
        return -1;
    }

    // Open the wav file passed from argument
    printf("file: %s\n", argv[1]);
    fh = fopen(argv[1], "rb");
    if (fh == NULL) {
        fprintf(stderr, "Unable to open input file %s\n", argv[1]);
        return -1;
    }

    // Start utterance
    rv = ps_start_utt(ps);
    
    // Process buffer, 512 samples at a time
    while (!feof(fh)) {
        size_t nsamp;
        nsamp = fread(buf, 2, 512, fh);
        rv = ps_process_raw(ps, buf, nsamp, FALSE, FALSE);
    }
    
    // Recieve the recognized string
    rv = ps_end_utt(ps);
    hyp = ps_get_hyp(ps, &score);
    printf("Recognized: |%s|\n", hyp);

    // free memory
    fclose(fh);
    ps_free(ps);
    cmd_ln_free_r(config);
    
    return 0;
}

static void
test_no_search()
{
    cmd_ln_t *config = default_config();
    ps_decoder_t *ps = ps_init(config);
    TEST_ASSERT(ps_start_utt(ps) < 0);
    ps_free(ps);
    cmd_ln_free_r(config);
}

static int config_input(AVFilterLink *inlink)
{
    AVFilterContext *ctx = inlink->dst;
    ASRContext *s = ctx->priv;

    ps_start_utt(s->ps);

    return 0;
}

 ReturnType Recognizer::start() {
   if ((decoder == NULL) || (is_recording)) return BAD_STATE;
   if (ps_start_utt(decoder, NULL) < 0) {
     return RUNTIME_ERROR;
   }
   current_hyp = "";
   is_recording = true;
   return SUCCESS;
 }

int
main(int argc, char *argv[])
{
	ps_decoder_t *ps;
	cmd_ln_t *config;
	FILE *fh;
	char const *hyp, *uttid;
        int16 buf[512];
	int rv;
	int32 score;

	config = cmd_ln_init(NULL, ps_args(), TRUE,
			     "-hmm", MODELDIR "/hmm/en_US/hub4wsj_sc_8k",
			     "-lm", MODELDIR "/lm/en/turtle.DMP",
			     "-dict", MODELDIR "/lm/en/turtle.dic",
			     NULL);
	if (config == NULL)
		return 1;
	ps = ps_init(config);
	if (ps == NULL)
		return 1;

	fh = fopen("goforward.raw", "rb");
	if (fh == NULL) {
		perror("Failed to open goforward.raw");
		return 1;
	}

	rv = ps_decode_raw(ps, fh, "goforward", -1);
	if (rv < 0)
		return 1;
	hyp = ps_get_hyp(ps, &score, &uttid);
	if (hyp == NULL)
		return 1;
	printf("Recognized: %s\n", hyp);

        fseek(fh, 0, SEEK_SET);
        rv = ps_start_utt(ps, "goforward");
	if (rv < 0)
		return 1;
        while (!feof(fh)) {
            size_t nsamp;
            nsamp = fread(buf, 2, 512, fh);
            rv = ps_process_raw(ps, buf, nsamp, FALSE, FALSE);
        }
        rv = ps_end_utt(ps);
	if (rv < 0)
		return 1;
	hyp = ps_get_hyp(ps, &score, &uttid);
	if (hyp == NULL)
		return 1;
	printf("Recognized: %s\n", hyp);

	fclose(fh);
        ps_free(ps);
	return 0;
}

/**
 * You must sucessfully call spInitListener
 * once before using this function.
 *
 * Reads the next block of audio from the microphone
 * up to SPLBUFSIZE number of samples
 * (defined in splistener.h).
 * If an utterance was completed in that block,
 * the transcription is stored in the string words.
 *
 * When calling this function in a realtime loop, delay
 * by some amount of time between calls keeping in mind
 * your recording's sample rate and maximum samples read
 * per call (ex. sleep the thread for 100 milliseconds)
 * so that some audio can be recorded for the next call.
 *
 * @return true if a speech session was completed and
 *         transcribed this block, otherwise false.
 */
static bool spDecode() {
    static bool uttered = false;

    // lock pocketsphinx resources to make sure they 
    // don't get freed by main thread while in use
    std::lock_guard<std::mutex> ps_lock(ps_mtx);
    if(!mic || !ps)
        return false;

    int samples_read = ad_read(mic, buf, SPLBUFSIZE);
    if (samples_read <= 0) {
        spError("failed to read audio :(");
        return false;
    }

    ps_process_raw(ps, buf, samples_read, FALSE, FALSE);
    bool talking = ps_get_in_speech(ps);

    // Just started talking
    if (talking && !uttered) {
        uttered = true;
        return false;
    }

    // Stopped talking, so transcribe what was said
    // and begin the next utterance
    if (!talking && uttered) {
        ps_end_utt(ps);
        const char *trans = ps_get_hyp(ps, NULL);

        if (ps_start_utt(ps) < 0) {
            spError("failed to start utterance :(");
        }
        uttered = false;

        int l = strlen(trans);
        if (trans && l > 0) {
            std::lock_guard<std::mutex> lock(words_mtx);
            if (words && l + 1 > words_buf_size) {
                delete words;
                words = NULL;
            }
            if (!words) {
                words = new char[l + 1];
                words_buf_size = l + 1;
            }

            std::copy(trans, trans + l, words);
            words[l] = '\0';
            
            return true;
        }
    }

    return false;
}

uint32 FSpeechRecognitionWorker::Run() {

	char const *hyp;
	// attempt to open the default recording device
	if ((ad = ad_open_dev(cmd_ln_str_r(config, "-adcdev"),
		(int)cmd_ln_float32_r(config,
		"-samprate"))) == NULL) {
		ClientMessage(FString(TEXT("Failed to open audio device")));
		return 1;
	}
	if (ad_start_rec(ad) < 0) {
		ClientMessage(FString(TEXT("Failed to start recording")));
		return 2;
	}
	if (ps_start_utt(ps) < 0) {
		ClientMessage(FString(TEXT("Failed to start utterance")));
		return 3;
	}

	while (StopTaskCounter.GetValue() == 0) {
		if ((k = ad_read(ad, adbuf, 1024)) < 0)
			ClientMessage(FString(TEXT("Failed to read audio")));
		ps_process_raw(ps, adbuf, k, 0, 0);
		in_speech = ps_get_in_speech(ps);
		if (in_speech && !utt_started) {
			utt_started = 1;
		}
		if (!in_speech && utt_started) {
			/* speech -> silence transition, time to start new utterance  */
			ps_end_utt(ps);
			hyp = ps_get_hyp(ps, NULL);
			if (hyp != NULL)
				Manager->WordSpoken_method(FString(hyp));

			if (ps_start_utt(ps) < 0)
				ClientMessage(FString(TEXT("Failed to start")));
			utt_started = 0;
		}
	}

	ad_close(ad);
	return 0;
}

int decode()
{
    int ret;
    int16 buf[BUFFER_SIZE];
    
    printf("Listening for input...\n");
    
    if (ad_start_rec(ad) < 0) {
        printf("Error starting recording.\n");
        return 0;
    }
    
	//check if not silent
	while ((ret = cont_ad_read(c_ad, buf, BUFFER_SIZE)) == 0)
        usleep(1000);
	
    if (ps_start_utt(ps, NULL) < 0) {
        printf("Failed to start utterance.\n");
        return 0;
    }
	
	ret = ps_process_raw(ps, buf, BUFFER_SIZE, 0, 0);
    if (ret < 0) {
        printf("Error decoding.\n");
        return 0;
    }
    
    do
    {
        ret = cont_ad_read(c_ad, buf, BUFFER_SIZE);

        if (ret < 0) {
            printf("Failed to record audio.\n");
            return 0;
        } else if(ret > 0) {
            // Valid speech data read.
            ret = ps_process_raw(ps, buf, 4096, 0, 0);
            if (ret < 0) {
                printf("Error decoding.\n");
                return 0;
            }
        } else {
            //no data
            usleep(1000);
        }
    } while(getRun());
    
    ad_stop_rec(ad);
    while (ad_read(ad, buf, BUFFER_SIZE) >= 0);
    cont_ad_reset(c_ad);

    ps_end_utt(ps);
    return 1;
}

int
main(int argc, char *argv[])
{
    ps_decoder_t *ps;
    cmd_ln_t *config;
    FILE *fh;
    char const *hyp, *uttid;
    int16 buf[512];
    int rv;
    int32 score;

    config = cmd_ln_init(NULL, ps_args(), TRUE,
                 "-hmm", MODELDIR "/en-us/en-us",
                 "-keyphrase", "marieta",
                 "-dict", MODELDIR "/en-us/cmudict-en-us.dict",
                 "-kws_threshold", "1e-30",
                 NULL);
    if (config == NULL) {
        fprintf(stderr, "Failed to create config object, see log for details\n");
        return -1;
    }
    
    ps = ps_init(config);
    if (ps == NULL) {
        fprintf(stderr, "Failed to create recognizer, see log for details\n");
        return -1;
    }

    fh = fopen("data/marieta.raw", "rb");
    if (fh == NULL) {
        fprintf(stderr, "Unable to open input file goforward.raw\n");
        return -1;
    }

    rv = ps_start_utt(ps);
    
    while (!feof(fh)) {
        size_t nsamp;
        nsamp = fread(buf, 2, 512, fh);
        rv = ps_process_raw(ps, buf, nsamp, FALSE, FALSE);
    }
    
    rv = ps_end_utt(ps);
    hyp = ps_get_hyp(ps, &score);
    printf("Recognized: %s\n", hyp);

    fclose(fh);
    ps_free(ps);
    cmd_ln_free_r(config);
    
    return 0;
}

static void gst_sphinx_sink_process_chunk (GstSphinxSink *sphinxsink)
{
	int32 k;
	int16 adbuf[REQUIRED_FRAME_SAMPLES];
	
	k = cont_ad_read (sphinxsink->cont, adbuf, REQUIRED_FRAME_SAMPLES);
	
	if (k == 0 && sphinxsink->last_ts == 0) {
	      return;
	} else if (k == 0 && sphinxsink->cont->read_ts - sphinxsink->last_ts > 
			    DEFAULT_SAMPLES_PER_SEC) {
	      int32 score;
	      const char *hyp;
    	      char *stripped_hyp;
    	      int i, j;
  
	      ps_end_utt (sphinxsink->decoder);
   	      if ((hyp = ps_get_hyp (sphinxsink->decoder, &score, NULL)) == NULL) {
	    	
	    	      gst_sphinx_sink_send_message (sphinxsink, "message", "");
                      g_message("Not Recognized");
	      
	      } else {
	    	      stripped_hyp = 
		           g_malloc (strlen (hyp) + 1);
	    	      for (i=0, j=0; hyp[i] != 0; i++) {
	    	    	    if (hyp[i] != '(' && hyp[i] != ')' && (hyp[i] < '0' || hyp[i] > '9')) {
	    	    		    stripped_hyp[j++] = hyp[i];
	    	    	    }
	    	      }
	    	      stripped_hyp [j] = 0;
	    	      
	    	      gst_sphinx_sink_send_message (sphinxsink, "message", stripped_hyp);
                      g_message("Recognized: %s", stripped_hyp);
	      }

	      sphinxsink->last_ts = 0;
	      sphinxsink->ad.listening = 0;

	} else if (k != 0) {
	     if (sphinxsink->ad.listening == 0) {
	    	    ps_start_utt (sphinxsink->decoder, NULL);
	    	    sphinxsink->ad.listening = 1;
		    gst_sphinx_sink_send_message (sphinxsink, "listening", NULL);
	    }
	
	     ps_process_raw (sphinxsink->decoder, adbuf, k, 0, 0);
	     sphinxsink->last_ts = sphinxsink->cont->read_ts;
	}
}

void utterance_start(context_t *ctx)
{
    decoder_set_t *decset;
    decoder_t *dec;
    char utid[256];

    if (ctx && (decset = ctx->decset) && (dec = decset->curdec)) {
        if (!dec->utter) {
            snprintf(utid, sizeof(utid), "%07u-%s", dec->utid++, dec->name);
            ps_start_utt(dec->ps, mrp_strdup(utid));
            dec->utter = true;
        }
    }
}

/*! function to read results from the ASR */
static switch_status_t pocketsphinx_asr_get_results(switch_asr_handle_t *ah, char **xmlstr, switch_asr_flag_t *flags)
{
	pocketsphinx_t *ps = (pocketsphinx_t *) ah->private_info;
	switch_status_t status = SWITCH_STATUS_SUCCESS;
    
    //switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_ERROR, ">>>>>>>>pocketsphinx_asr_get_results<<<<<<<<<\n");


	if (switch_test_flag(ps, PSFLAG_BARGE)) {
		switch_clear_flag_locked(ps, PSFLAG_BARGE);
		status = SWITCH_STATUS_BREAK;
	}

	if (switch_test_flag(ps, PSFLAG_HAS_TEXT)) {
		switch_mutex_lock(ps->flag_mutex);
		switch_clear_flag(ps, PSFLAG_HAS_TEXT);

		switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_DEBUG, "Recognized: %s, Confidence: %d, Confidence-Threshold: %d\n", ps->hyp, ps->confidence, ps->confidence_threshold);
		switch_mutex_unlock(ps->flag_mutex);

		*xmlstr = switch_mprintf("%s", ps->hyp);

		if (!switch_test_flag(ps, PSFLAG_INPUT_TIMERS) && switch_test_flag(ah, SWITCH_ASR_FLAG_AUTO_RESUME)) {
			switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_DEBUG, "Auto Resuming\n");
			switch_set_flag(ps, PSFLAG_READY);

			ps_start_utt(ps->ps, NULL);
		}

		status = SWITCH_STATUS_SUCCESS;
	} else if (switch_test_flag(ps, PSFLAG_NOINPUT)) {
		switch_clear_flag_locked(ps, PSFLAG_NOINPUT);
		switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_DEBUG, "NO INPUT\n");

		*xmlstr = switch_mprintf("");

		status = SWITCH_STATUS_SUCCESS;
	} else if (switch_test_flag(ps, PSFLAG_NOMATCH)) {
		switch_clear_flag_locked(ps, PSFLAG_NOMATCH);
		switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_DEBUG, "NO MATCH\n");

		*xmlstr = switch_mprintf("");

		status = SWITCH_STATUS_SUCCESS;
	}

	return status;
}

/*! function to feed audio to the ASR */
static switch_status_t pocketsphinx_asr_feed(switch_asr_handle_t *ah, void *data, unsigned int len, switch_asr_flag_t *flags)
{
	pocketsphinx_t *ps = (pocketsphinx_t *) ah->private_info;
	int rv = 0;

	if (switch_test_flag(ah, SWITCH_ASR_FLAG_CLOSED))
		return SWITCH_STATUS_BREAK;

	if (!switch_test_flag(ps, PSFLAG_HAS_TEXT) && switch_test_flag(ps, PSFLAG_READY)) {
		if (stop_detect(ps, (int16_t *) data, len / 2)) {
			char const *hyp;

			switch_mutex_lock(ps->flag_mutex);
			if ((hyp = ps_get_hyp(ps->ps, &ps->score, &ps->uttid))) {
				if (!zstr(hyp)) {
					ps_end_utt(ps->ps);
					switch_clear_flag(ps, PSFLAG_READY);
					if ((hyp = ps_get_hyp(ps->ps, &ps->score, &ps->uttid))) {
						if (zstr(hyp)) {
							switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_WARNING, "Lost the text, never mind....\n");
							ps_start_utt(ps->ps, NULL);
							switch_set_flag(ps, PSFLAG_READY);
						} else {
							ps->hyp = switch_core_strdup(ah->memory_pool, hyp);
							switch_set_flag(ps, PSFLAG_HAS_TEXT);
						}
					}
				}
			}
			switch_mutex_unlock(ps->flag_mutex);
		}

		/* only feed ps_process_raw when we are listening */
		if (ps->listening) {
			switch_mutex_lock(ps->flag_mutex);
			rv = ps_process_raw(ps->ps, (int16 *) data, len / 2, FALSE, FALSE);
			switch_mutex_unlock(ps->flag_mutex);
		}

		if (rv < 0) {
			return SWITCH_STATUS_FALSE;
		}
	}

	return SWITCH_STATUS_SUCCESS;
}

int
main(int argc, char *argv[])
{
  ps_decoder_t *ps;
  cmd_ln_t *config;
  FILE *fh;
  char const *hyp, *uttid;
  int16 buf[512];
  int rv;
  int32 score;

  config = cmd_ln_init(NULL, ps_args(), TRUE,
		       "-hmm", MODELDIR "/en-us/en-us",
		       "-lm", MODELDIR "/en-us/en-us.lm.dmp",
		       "-dict", MODELDIR "/en-us/cmudict-en-us.dict",
		       NULL);
  if (config == NULL)
    return 1;
  ps = ps_init(config);
  if (ps == NULL)
    return 1;

  fh = fopen("goforward.raw", "rb");
  if (fh == NULL)
    return -1;
  rv = ps_start_utt(ps);
  if (rv < 0)
    return 1;
  while (!feof(fh)) {
    size_t nsamp;
    nsamp = fread(buf, 2, 512, fh);
    rv = ps_process_raw(ps, buf, nsamp, FALSE, FALSE);
  }
  rv = ps_end_utt(ps);
  if (rv < 0)
    return 1;
  hyp = ps_get_hyp(ps, &score);
  if (hyp == NULL)
    return 1;
  printf("Recognized: %s\n", hyp);

  fclose(fh);
  ps_free(ps);
  cmd_ln_free_r(config);
  return 0;
}

/*! function to read results from the ASR*/
static switch_status_t pocketsphinx_asr_get_results(switch_asr_handle_t *ah, char **xmlstr, switch_asr_flag_t *flags)
{
	pocketsphinx_t *ps = (pocketsphinx_t *) ah->private_info;
	switch_status_t status = SWITCH_STATUS_SUCCESS;
	int32_t conf;

	if (switch_test_flag(ps, PSFLAG_BARGE)) {
		switch_clear_flag_locked(ps, PSFLAG_BARGE);
		status = SWITCH_STATUS_BREAK;
	}

	if (switch_test_flag(ps, PSFLAG_HAS_TEXT)) {
		switch_mutex_lock(ps->flag_mutex);
		switch_clear_flag(ps, PSFLAG_HAS_TEXT);
		conf = ps_get_prob(ps->ps, &ps->uttid);

		ps->confidence = (conf + 20000) / 200;

		if (ps->confidence < 0) {
			ps->confidence = 0;
		}

		switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_DEBUG, "Recognized: %s, Confidence: %d\n", ps->hyp, ps->confidence);
		switch_mutex_unlock(ps->flag_mutex);

		*xmlstr = switch_mprintf("<?xml version=\"1.0\"?>\n"
								 "<result grammar=\"%s\">\n"
								 "  <interpretation grammar=\"%s\" confidence=\"%d\">\n"
								 "    <input mode=\"speech\">%s</input>\n"
								 "  </interpretation>\n" "</result>\n", ps->grammar, ps->grammar, ps->confidence, ps->hyp);

		if (switch_test_flag(ps, SWITCH_ASR_FLAG_AUTO_RESUME)) {
			switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_DEBUG, "Auto Resuming\n");
			switch_set_flag(ps, PSFLAG_READY);

			ps_start_utt(ps->ps, NULL);
		}

		status = SWITCH_STATUS_SUCCESS;
	}

	return status;
}