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;
}
std::string PocketSphinxServer::decode(const PocketSphinxIce::sample& signal, const Ice::Current& c)
{
std::cout << "Decode\n";
int rv;
const char* hyp;
const char* uttid;
int32 score;
rv = ps_start_utt(ps);
if (rv < 0)
throw PocketSphinxIce::Error("Error in ps_start_utt");
ps_process_raw(ps, signal.data(), signal.size(), FALSE, FALSE);
rv = ps_end_utt(ps);
if (rv < 0)
throw PocketSphinxIce::Error("Error in ps_end_utt");
hyp = ps_get_hyp(ps, &score);
if (!hyp)
throw PocketSphinxIce::Error("ps_get_hyp returned NULL");
std::cout << "return:" << hyp << '\n';
return hyp;
}
int processRaw(const char *rawFile) {
const char *hyp, *uttid;
int16 buf[512];
int rv;
int32 score;
// Open the wav file passed from argument
printf("file: %s\n", rawFile);
fh = fopen(rawFile, "rb");
if (fh == NULL) {
fprintf(stderr, "Unable to open input file %s\n", rawFile);
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);
fflush(stdout);
// Close file
fclose(fh);
return 0;
}
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;
//int i;
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;
}
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;
}
ReturnType Recognizer::process(const std::vector<int16_t>& buffer) {
if ((decoder == NULL) || (!is_recording)) return BAD_STATE;
if (buffer.size() == 0)
return RUNTIME_ERROR;
ps_process_raw(decoder, (short int *) &buffer[0], buffer.size(), 0, 0);
const char* h = ps_get_hyp(decoder, &score, &sentence_id);
current_hyp = (h == NULL) ? "" : h;
return SUCCESS;
}
/**
* 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;
}
long
ps_decode_raw(ps_decoder_t *ps, FILE *rawfh,
long maxsamps)
{
int16 *data;
long total, pos, endpos;
ps_start_stream(ps);
ps_start_utt(ps);
/* If this file is seekable or maxsamps is specified, then decode
* the whole thing at once. */
if (maxsamps != -1) {
data = ckd_calloc(maxsamps, sizeof(*data));
total = fread(data, sizeof(*data), maxsamps, rawfh);
ps_process_raw(ps, data, total, FALSE, TRUE);
ckd_free(data);
} else if ((pos = ftell(rawfh)) >= 0) {
fseek(rawfh, 0, SEEK_END);
endpos = ftell(rawfh);
fseek(rawfh, pos, SEEK_SET);
maxsamps = endpos - pos;
data = ckd_calloc(maxsamps, sizeof(*data));
total = fread(data, sizeof(*data), maxsamps, rawfh);
ps_process_raw(ps, data, total, FALSE, TRUE);
ckd_free(data);
} else {
/* Otherwise decode it in a stream. */
total = 0;
while (!feof(rawfh)) {
int16 data[256];
size_t nread;
nread = fread(data, sizeof(*data), sizeof(data)/sizeof(*data), rawfh);
ps_process_raw(ps, data, nread, FALSE, FALSE);
total += nread;
}
}
ps_end_utt(ps);
return total;
}
void GqAndroidSphinx::received_buf_from_recorder(short *record_buf,
unsigned long buf_size_in_byte) {
m_fout.write((char *) record_buf, buf_size_in_byte);
pthread_mutex_lock(&m_pt_mutex);
LOGD("received_buf_from_recorder before");
ps_process_raw(m_pdecoder, record_buf, buf_size_in_byte / sizeof(int16),
TRUE, FALSE);
LOGD("received_buf_from_recorder end");
pthread_mutex_unlock(&m_pt_mutex);
}
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;
}
void Recognizer::run()
{
// Create audio converter:
auto converter = ci::audio::dsp::Converter::create( mMonitorNode->getSampleRate(), 16000, mMonitorNode->getNumChannels(), 1, mMonitorNode->getFramesPerBlock() );
// Create buffer for converted audio:
ci::audio::Buffer destBuffer( converter->getDestMaxFramesPerBlock(), converter->getDestNumChannels() );
bool utt_started, in_speech;
if( ps_start_utt( mDecoder ) < 0 )
throw std::runtime_error( "Could not start utterance" );
utt_started = false;
while( ! mStop ) {
// Convert buffer:
std::pair<size_t,size_t> convertResult = converter->convert( &( mMonitorNode->getBuffer() ), &destBuffer );
// Convert buffer data:
int16_t* data = new int16_t[ convertResult.second ];
convertFloatToInt16( destBuffer.getData(), data, convertResult.second );
// Process buffer:
ps_process_raw( mDecoder, data, convertResult.second, false, false );
// Cleanup buffer data:
delete[] data;
in_speech = static_cast<bool>( ps_get_in_speech( mDecoder ) );
if( in_speech && ! utt_started ) {
utt_started = true;
}
if( ! in_speech && utt_started ) {
// Start new utterance on speech to silence transition:
ps_end_utt( mDecoder );
// Pass to handler:
if( mHandler )
mHandler->event( mDecoder );
// Prepare for next utterance:
if( ps_start_utt( mDecoder ) < 0 )
throw std::runtime_error( "Could not start utterance" );
utt_started = false;
}
std::this_thread::sleep_for( std::chrono::milliseconds( 10 ) );
}
}
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;
}
}
/*
* Continuous recognition from mic
*/
int
recognize_from_mic()
{
ad_rec_t *ad;
int16 adbuf[2048];
const char *fname;
const char* seg;
int32 k;
char str[1000]="";
uint8 utt_started, in_speech;
if ((ad = ad_open_dev(cmd_ln_str_r(config, "-adcdev"),16000)) == NULL)
perror("Failed to open audio device\n");
if (ad_start_rec(ad) < 0)
perror("Failed to start recording\n");
ps_start_utt(ps);
utt_started = FALSE;
ps_seg_t *psegt;
while (!finished) {
if ((k = ad_read(ad, adbuf, 2048)) < 0)
perror("Failed to read audio\n");
ps_process_raw(ps, adbuf, k, FALSE, FALSE);
in_speech = ps_get_in_speech(ps);
if (in_speech && !utt_started) {
utt_started = TRUE;
}
if (!in_speech && utt_started) {
ps_end_utt(ps);
psegt = ps_seg_iter(ps, NULL);
while (psegt!=NULL){
seg = ps_seg_word(psegt);
strncpy_s( str, seg, strlen(seg));
listenCallback(str);
printf("%s\n", seg);
int prob = ps_seg_prob(psegt,NULL,NULL,NULL);
printf("%d\n", prob);
psegt = ps_seg_next(psegt);
}
ps_start_utt(ps);
utt_started = FALSE;
}
Sleep(100);
}
ps_end_utt(ps);
fclose(rawfd);
return 0;
}
/*
* Main utterance processing loop:
* for (;;) {
* start utterance and wait for speech to process
* decoding till end-of-utterance silence will be detected
* print utterance result;
* }
*/
static void
recognize_from_microphone()
{
ad_rec_t *ad;
int16 adbuf[2048];
uint8 utt_started, in_speech;
int32 k;
char const *hyp;
if ((ad = ad_open_dev(cmd_ln_str_r(config, "-adcdev"),
(int) cmd_ln_float32_r(config,
"-samprate"))) == NULL)
E_FATAL("Failed to open audio device\n");
if (ad_start_rec(ad) < 0)
E_FATAL("Failed to start recording\n");
if (ps_start_utt(ps) < 0)
E_FATAL("Failed to start utterance\n");
utt_started = FALSE;
E_INFO("Ready....\n");
for (;;) {
if ((k = ad_read(ad, adbuf, 2048)) < 0)
E_FATAL("Failed to read audio\n");
ps_process_raw(ps, adbuf, k, FALSE, FALSE);
in_speech = ps_get_in_speech(ps);
if (in_speech && !utt_started) {
utt_started = TRUE;
E_INFO("Listening...\n");
}
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) {
printf("%s\n", hyp);
fflush(stdout);
}
if (ps_start_utt(ps) < 0)
E_FATAL("Failed to start utterance\n");
utt_started = FALSE;
E_INFO("Ready....\n");
}
sleep_msec(100);
}
ad_close(ad);
}
static void
recognize_from_microphone()
{
ad_rec_t *ad;
int16 adbuf[2048];
uint8 utt_started, in_speech;
int32 k;
char const *hyp;
if ((ad = ad_open_dev(AUDIO_DEVICE_NAME,
(int) SAMPLE_RATE )) == NULL) {
E_FATAL("Failed to open audio device\n");
}
if (ad_start_rec(ad) < 0) {
E_FATAL("Failed to start recording\n");
}
if (ps_start_utt(ps) < 0) {
E_FATAL("Failed to start utterance\n");
}
utt_started = FALSE;
printf("READY....\n");
for (;;) {
if ((k = ad_read(ad, adbuf, 2048)) < 0)
E_FATAL("Failed to read audio\n");
ps_process_raw(ps, adbuf, k, FALSE, FALSE);
in_speech = ps_get_in_speech(ps);
if (in_speech && !utt_started) {
utt_started = TRUE;
printf("Listening...\n");
}
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)
printf("%s\n", hyp);
if (ps_start_utt(ps) < 0)
E_FATAL("Failed to start utterance\n");
utt_started = FALSE;
printf("READY....\n");
}
sleep_msec(100);
}
ad_close(ad);
}
/*
* Continuous recognition from a file
*/
int
recognize_from_file()
{
int16 adbuf[2048];
const char *fname;
const char *hyp;
int32 k;
char str[1000]="";
uint8 utt_started, in_speech;
fname = "C:/Users/Reza/Documents/GitHub/speech_agent/presentation_samples/italy1_reza.wav";
rawfd = fopen(fname, "rb");
if (strlen(fname) > 4 && strcmp(fname + strlen(fname) - 4, ".wav") == 0) {
char waveheader[44];
fread(waveheader, 1, 44, rawfd);
}
ps_start_utt(ps);
utt_started = FALSE;
while ((k = fread(adbuf, sizeof(int16), 2048, rawfd)) > 0) {
ps_process_raw(ps, adbuf, k, FALSE, FALSE);
in_speech = ps_get_in_speech(ps);
if (in_speech && !utt_started) {
utt_started = TRUE;
}
if (!in_speech && utt_started) {
ps_end_utt(ps);
hyp = ps_get_hyp(ps, NULL);
if (hyp != NULL){
strncpy_s( str, hyp, strlen(hyp));
printf("%s\n", hyp);
listenCallback(str);
}
ps_start_utt(ps);
utt_started = FALSE;
}
}
ps_end_utt(ps);
fclose(rawfd);
return 0;
}
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 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;
}
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;
}
static GstFlowReturn
gst_pocketsphinx_chain(GstPad * pad, GstBuffer * buffer)
{
GstPocketSphinx *ps;
ps = GST_POCKETSPHINX(GST_OBJECT_PARENT(pad));
/* Start an utterance for the first buffer we get (i.e. we assume
* that the VADER is "leaky") */
if (!ps->listening) {
ps->listening = TRUE;
ps_start_utt(ps->ps, NULL);
}
ps_process_raw(ps->ps,
(short *)GST_BUFFER_DATA(buffer),
GST_BUFFER_SIZE(buffer) / sizeof(short),
FALSE, FALSE);
/* Get a partial result every now and then, see if it is different. */
if (ps->last_result_time == 0
/* Check every 100 milliseconds. */
|| (GST_BUFFER_TIMESTAMP(buffer) - ps->last_result_time) > 100*10*1000) {
int32 score;
char const *hyp;
char const *uttid;
hyp = ps_get_hyp(ps->ps, &score, &uttid);
ps->last_result_time = GST_BUFFER_TIMESTAMP(buffer);
if (hyp && strlen(hyp) > 0) {
if (ps->last_result == NULL || 0 != strcmp(ps->last_result, hyp)) {
g_free(ps->last_result);
ps->last_result = g_strdup(hyp);
/* Emit a signal for applications. */
g_signal_emit(ps, gst_pocketsphinx_signals[SIGNAL_PARTIAL_RESULT],
0, hyp, uttid);
}
}
}
gst_buffer_unref(buffer);
return GST_FLOW_OK;
}
int processCommands()
{
int32 samples;
int16 audioBuf[BUFFER_SIZE];
char const *uttid;
char const *hyp;
while(run)
{
printf("Waiting for utterance...\n");
samples = waitForNextUtterance();
if(samples < 0)
return -1;
if(ps_start_utt(psDecoder, NULL) < 0) {
fprintf(stderr, "Failed to start next utterance\n");
return -1;
}
ps_process_raw(psDecoder, audioBuf, samples, FALSE, FALSE);
printf("Recording...\n");
fflush(stdout);
record();
ad_stop_rec(audioDevice);
while(ad_read(audioDevice, audioBuf, BUFFER_SIZE) >= 0);
cont_ad_reset(continousAudoDevice);
ps_end_utt(psDecoder);
hyp = ps_get_hyp(psDecoder, NULL, &uttid);
printf("Heard: %s\n", hyp);
if (ad_start_rec(audioDevice) < 0) {
fprintf(stderr, "Failed to start audio device.\n");
return -1;
}
}
return 0;
}
int ofApp::engineOpen(string filename)
{
FILE *fh;
char const *uttid;
int16 buf[512];
int rv; int32 score;
fh = fopen((directoryString + filename).c_str(), "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);
}
engineClose();
}
void record()
{
int32 samples, timeStamp, rem;
int16 audioBuf[BUFFER_SIZE];
timeStamp = continousAudoDevice->read_ts;
while(run)
{
samples = cont_ad_read(continousAudoDevice, audioBuf, BUFFER_SIZE);
if (samples == 0) {
if ((continousAudoDevice->read_ts - timeStamp) > DEFAULT_SAMPLES_PER_SEC)
break;
} else {
timeStamp = continousAudoDevice->read_ts;
}
rem = ps_process_raw(psDecoder, audioBuf, samples, FALSE, FALSE);
if ((rem == 0) && (samples == 0))
usleep(20000);
}
}
void filter_buffer_utter(context_t *ctx, bool full_utterance)
{
decoder_set_t *decset;
decoder_t *dec;
filter_buf_t *filtbuf;
int sts, cnt, size;
if (!ctx || !(decset = ctx->decset) || !(dec = decset->curdec) ||
!(filtbuf = ctx->filtbuf))
return;
mrp_debug("utterance length %d samples", filtbuf->len);
if (filtbuf->len > 0) {
if (filtbuf->fdrec >= 0) {
size = filtbuf->len * sizeof(int16);
for (;;) {
cnt = write(filtbuf->fdrec, filtbuf->buf, size);
if (cnt != size) {
if (cnt < 0 && errno == EINTR)
continue;
mrp_log_error("failed to record samples (fd %d): %s",
filtbuf->fdrec, strerror(errno));
}
break;
}
}
sts = ps_process_raw(dec->ps, filtbuf->buf, filtbuf->len,
FALSE, full_utterance);
if (sts < 0)
mrp_log_error("Failed to process %d samples", filtbuf->len);
}
}
static int filter_frame(AVFilterLink *inlink, AVFrame *in)
{
AVFilterContext *ctx = inlink->dst;
AVDictionary **metadata = &in->metadata;
ASRContext *s = ctx->priv;
int have_speech;
const char *speech;
ps_process_raw(s->ps, (const int16_t *)in->data[0], in->nb_samples, 0, 0);
have_speech = ps_get_in_speech(s->ps);
if (have_speech && !s->utt_started)
s->utt_started = 1;
if (!have_speech && s->utt_started) {
ps_end_utt(s->ps);
speech = ps_get_hyp(s->ps, NULL);
if (speech != NULL)
av_dict_set(metadata, "lavfi.asr.text", speech, 0);
ps_start_utt(s->ps);
s->utt_started = 0;
}
return ff_filter_frame(ctx->outputs[0], in);
}
/*
* Continuous recognition from a file
*/
static void
recognize_from_file()
{
int16 adbuf[2048];
const char *fname;
const char *hyp;
int32 k;
uint8 utt_started, in_speech;
int32 print_times = cmd_ln_boolean_r(config, "-time");
fname = cmd_ln_str_r(config, "-infile");
if ((rawfd = fopen(fname, "rb")) == NULL) {
E_FATAL_SYSTEM("Failed to open file '%s' for reading",
fname);
}
if (strlen(fname) > 4 && strcmp(fname + strlen(fname) - 4, ".wav") == 0) {
char waveheader[44];
fread(waveheader, 1, 44, rawfd);
if (!check_wav_header(waveheader, (int)cmd_ln_float32_r(config, "-samprate")))
E_FATAL("Failed to process file '%s' due to format mismatch.\n", fname);
}
if (strlen(fname) > 4 && strcmp(fname + strlen(fname) - 4, ".mp3") == 0) {
E_FATAL("Can not decode mp3 files, convert input file to WAV 16kHz 16-bit mono before decoding.\n");
}
ps_start_utt(ps);
utt_started = FALSE;
while ((k = fread(adbuf, sizeof(int16), 2048, rawfd)) > 0) {
ps_process_raw(ps, adbuf, k, FALSE, FALSE);
in_speech = ps_get_in_speech(ps);
if (in_speech && !utt_started) {
utt_started = TRUE;
}
if (!in_speech && utt_started) {
ps_end_utt(ps);
hyp = ps_get_hyp(ps, NULL);
if (hyp != NULL)
printf("%s\n", hyp);
if (print_times)
print_word_times();
fflush(stdout);
ps_start_utt(ps);
utt_started = FALSE;
}
}
ps_end_utt(ps);
if (utt_started) {
hyp = ps_get_hyp(ps, NULL);
if (hyp != NULL) {
printf("%s\n", hyp);
if (print_times) {
print_word_times();
}
}
}
fclose(rawfd);
}
/*
* Main utterance processing loop:
* for (;;) {
* wait for start of next utterance;
* decode utterance until silence of at least 1 sec observed;
* print utterance result;
* }
*/
static void
recognize_from_microphone()
{
ad_rec_t *ad;
int16 adbuf[4096];
int32 k, ts, rem;
char const *hyp;
char const *uttid;
cont_ad_t *cont;
char word[256];
char c1[256], c2[256];
int tracking = 0;
int halted = 0;
int LEFT = 0;
int RIGHT = 1;
int MOVE_CENT = 100; //1 meter
int numwords;
setlinebuf(stdout);
if ((ad = ad_open_dev(cmd_ln_str_r(config, "-adcdev"),
(int)cmd_ln_float32_r(config, "-samprate"))) == NULL)
E_FATAL("Failed to open audio device\n");
/* Initialize continuous listening module */
if ((cont = cont_ad_init(ad, ad_read)) == NULL)
E_FATAL("Failed to initialize voice activity detection\n");
if (ad_start_rec(ad) < 0)
E_FATAL("Failed to start recording\n");
if (cont_ad_calib(cont) < 0)
E_FATAL("Failed to calibrate voice activity detection\n");
printf("LEDON BLUE\n");
for (;;) {
/* Indicate listening for next utterance */
fprintf(stderr, "READY....\n");
fflush(stderr);
/* Wait data for next utterance */
while ((k = cont_ad_read(cont, adbuf, 4096)) == 0)
sleep_msec(100);
if (k < 0)
E_FATAL("Failed to read audio\n");
/*
* Non-zero amount of data received; start recognition of new utterance.
* NULL argument to uttproc_begin_utt => automatic generation of utterance-id.
*/
if (ps_start_utt(ps, NULL) < 0)
E_FATAL("Failed to start utterance\n");
ps_process_raw(ps, adbuf, k, FALSE, FALSE);
fprintf(stderr, "Listening...\n");
/* Note timestamp for this first block of data */
ts = cont->read_ts;
/* Decode utterance until end (marked by a "long" silence, >1sec) */
for (;;) {
/* Read non-silence audio data, if any, from continuous listening module */
if ((k = cont_ad_read(cont, adbuf, 4096)) < 0)
E_FATAL("Failed to read audio\n");
if (k == 0) {
/*
* No speech data available; check current timestamp with most recent
* speech to see if more than 1 sec elapsed. If so, end of utterance.
*/
if ((cont->read_ts - ts) > DEFAULT_SAMPLES_PER_SEC)
break;
}
else {
/* New speech data received; note current timestamp */
ts = cont->read_ts;
}
/*
* Decode whatever data was read above.
*/
rem = ps_process_raw(ps, adbuf, k, FALSE, FALSE);
/* If no work to be done, sleep a bit */
if ((rem == 0) && (k == 0))
sleep_msec(20);
}
/*
* Utterance ended; flush any accumulated, unprocessed A/D data and stop
* listening until current utterance completely decoded
*/
ad_stop_rec(ad);
while (ad_read(ad, adbuf, 4096) >= 0);
cont_ad_reset(cont);
fprintf(stderr, "Stopped listening, please wait...\n");
fflush(stdout);
/* Finish decoding, obtain and print result */
ps_end_utt(ps);
hyp = ps_get_hyp(ps, NULL, &uttid);
fprintf(stderr, "%s: %s\n", uttid, hyp);
/* Exit if the first word spoken was GOODBYE */
if (hyp) {
numwords = sscanf(hyp, "%s %s %s", word, c1, c2);
if(strcmp(word, "GUGGUG") == 0) {
if(strcmp(c1, "HALT") == 0) {
printf("LEDOFF BLUE\n");
halted = 1;
} else if(strcmp(c1, "RESUME") == 0) {
printf("LEDON BLUE\n");
halted = 0;
}
if(strcmp(c1, "BEGIN") == 0 || strcmp(c1, "START") == 0) {
if(strcmp(c2, "TRACKING") == 0 && !tracking) {
printf("START TRACKING\n");
tracking = 1;
halted = 0;
}
} else if(strcmp(c1, "STOP") == 0) {
if(strcmp(c2, "TRACKING") == 0 && tracking) {
printf("STOP TRACKING\n");
tracking = 0;
}
}
if(!tracking && !halted && numwords == 3) {
if(strcmp(c1, "TURN") == 0) {
if(strcmp(c2, "AROUND") == 0) {
printf("TURN %d 180\n", LEFT);
} else if(strcmp(c2, "LEFT") == 0) {
printf("TURN %d 90\n", LEFT);
} else if(strcmp(c2, "RIGHT") == 0) {
printf("TURN %d 90\n", RIGHT);
}
} else if(strcmp(c1, "MOVE") == 0) {
if(strcmp(c2, "FORWARD") == 0) {
printf("MOVE 0 %d\n", MOVE_CENT);
} else if(strcmp(c2, "BACKWARD") == 0) {
printf("MOVE 1 %d\n", MOVE_CENT);
}
}
}
}
}
/* Resume A/D recording for next utterance */
if (ad_start_rec(ad) < 0)
E_FATAL("Failed to start recording\n");
}
cont_ad_close(cont);
ad_close(ad);
}
/*
* Continuous recognition from a file
*/
static void
recognize_from_file() {
cont_ad_t *cont;
ad_rec_t file_ad = {0};
int16 adbuf[4096];
const char* hyp;
const char* uttid;
int32 k, ts, start;
char waveheader[44];
if ((rawfd = fopen(cmd_ln_str_r(config, "-infile"), "rb")) == NULL) {
E_FATAL_SYSTEM("Failed to open file '%s' for reading",
cmd_ln_str_r(config, "-infile"));
}
fread(waveheader, 1, 44, rawfd);
file_ad.sps = (int32)cmd_ln_float32_r(config, "-samprate");
file_ad.bps = sizeof(int16);
if ((cont = cont_ad_init(&file_ad, ad_file_read)) == NULL) {
E_FATAL("Failed to initialize voice activity detection");
}
if (cont_ad_calib(cont) < 0)
E_FATAL("Failed to calibrate voice activity detection\n");
rewind (rawfd);
for (;;) {
while ((k = cont_ad_read(cont, adbuf, 4096)) == 0);
if (k < 0) {
break;
}
if (ps_start_utt(ps, NULL) < 0)
E_FATAL("ps_start_utt() failed\n");
ps_process_raw(ps, adbuf, k, FALSE, FALSE);
ts = cont->read_ts;
start = ((ts - k) * 100.0) / file_ad.sps;
for (;;) {
if ((k = cont_ad_read(cont, adbuf, 4096)) < 0)
break;
if (k == 0) {
/*
* No speech data available; check current timestamp with most recent
* speech to see if more than 1 sec elapsed. If so, end of utterance.
*/
if ((cont->read_ts - ts) > DEFAULT_SAMPLES_PER_SEC)
break;
}
else {
/* New speech data received; note current timestamp */
ts = cont->read_ts;
}
ps_process_raw(ps, adbuf, k, FALSE, FALSE);
}
ps_end_utt(ps);
if (cmd_ln_boolean_r(config, "-time")) {
print_word_times(start);
} else {
hyp = ps_get_hyp(ps, NULL, &uttid);
fprintf(stderr, "%s: %s\n", uttid, hyp);
}
fflush(stdout);
}
cont_ad_close(cont);
fclose(rawfd);
}
void listen::recognize_from_microphone(){
ad_rec_t *ad;
int16 adbuf[4096];
int32 k, ts, rem;
char buffer[128];
char const *hyp;
char const *uttid;
cont_ad_t *cont;
state = SLEEPING;
FILE* pipe = popen(c.getValue("[General]", "Hcidump").c_str(), "r");
std::string bufferStr;
std::size_t found;
if((ad = ad_open_dev(cmd_ln_str_r(config, "-adcdev"), (int)cmd_ln_float32_r(config, "-samprate"))) == NULL)
E_FATAL("Failed to open audio device\n");
/* Initialize continuous listening module */
if((cont = cont_ad_init(ad, ad_read)) == NULL)
E_FATAL("Failed to initialize voice activity detection\n");
if(ad_start_rec(ad) < 0)
E_FATAL("Failed to start recording\n");
if(cont_ad_calib(cont) < 0)
E_FATAL("Failed to calibrate voice activity detection\n");
while(!feof(pipe) || state == SLEEPING){
fgets(buffer, 128, pipe);
bufferStr = buffer;
found = bufferStr.find(c.getValue("[General]", "KeyPress"));
if(found!=std::string::npos){
i.pauseIfPlaying();
s.speakThis(c.getValue("[General]", "WakeUpPhrase"));
state = ACTIVE;
while(state != SLEEPING){
/* Indicate listening for next utterance */
printf("READY....\n");
fflush(stdout);
fflush(stderr);
/* Wait data for next utterance */
while ((k = cont_ad_read(cont, adbuf, 4096)) == 0){
sleep_msec(100);
}
if (k < 0)
E_FATAL("Failed to read audio\n");
/*
* Non-zero amount of data received; start recognition of new utterance.
* NULL argument to uttproc_begin_utt => automatic generation of utterance-id.
*/
if (ps_start_utt(ps, NULL) < 0)
E_FATAL("Failed to start utterance\n");
ps_process_raw(ps, adbuf, k, FALSE, FALSE);
printf("Listening...\n");
fflush(stdout);
/* Note timestamp for this first block of data */
ts = cont->read_ts;
/* Decode utterance until end (marked by a "long" silence, >1sec) */
for(;;){
//while(sleep(2)){
/* Read non-silence audio data, if any, from continuous listening module */
if ((k = cont_ad_read(cont, adbuf, 4096)) < 0)
E_FATAL("Failed to read audio\n");
if (k == 0){
/*
* No speech data available; check current timestamp with most recent
* speech to see if more than 1 sec elapsed. If so, end of utterance.
*/
if ((cont->read_ts - ts) > DEFAULT_SAMPLES_PER_SEC)
break;
}
else {
/* New speech data received; note current timestamp */
ts = cont->read_ts;
}
/*
* Decode whatever data was read above.
*/
rem = ps_process_raw(ps, adbuf, k, FALSE, FALSE);
/* If no work to be done, sleep a bit */
if ((rem == 0) && (k == 0))
sleep_msec(20);
}
/*
* Utterance ended; flush any accumulated, unprocessed A/D data and stop
* listening until current utterance completely decoded
*/
ad_stop_rec(ad);
while (ad_read(ad, adbuf, 4096) >= 0);
cont_ad_reset(cont);
printf("Stopped listening, please wait...\n");
fflush(stdout);
/* Finish decoding, obtain and print result */
ps_end_utt(ps);
hyp = ps_get_hyp(ps, NULL, &uttid);
fflush(stdout);
/* Resume A/D recording for next utterance */
if (ad_start_rec(ad) < 0)
E_FATAL("Failed to start recording\n");
if(hyp != NULL){
if(hyp == c.getValue("[General]", "Sleep")){
state = SLEEPING;
s.speakThis(c.getValue("[General]", "SleepPhrase"));
}else{
if(state != SLEEPING){
i.parse(hyp);
}
//Hack for play/pause/select
std::string hypStr = hyp;
if(hypStr == "PLAY ITEM" || hypStr == "PAUSE ITEM" || hypStr == "SELECT ITEM"){
state = SLEEPING;
}
}
}
}
}
}
pclose(pipe);
cont_ad_close(cont);
ad_close(ad);
}
SWIGINTERN int Decoder_processRaw__SWIG_1(Decoder *self,short const shorts[],size_t nshorts,bool no_search,bool full_utt){
return ps_process_raw(self, shorts, nshorts, no_search, full_utt);
}
