static espeak_ERROR set_punctuation(struct synth_t *s, int punct,
enum adjust_t adj)
{
espeak_ERROR rc;
if (adj == ADJ_DEC)
punct = -punct;
if (adj != ADJ_SET)
punct += s->punct;
rc = espeak_SetParameter(espeakPUNCTUATION, punct, 0);
if (rc == EE_OK)
s->punct = punct;
return rc;
}
void tts_capitalize(int capitalize)
{
/* indicate capital by pitch */
espeak_ERROR erc;
//debug_log(logfd, "Called tts_capitalize: %d\n", capitalize);
tts_state.capitalize = capitalize;
if (capitalize) {
tts_allcaps_beep(0);
tts_split_caps(0);
}
erc = espeak_SetParameter(espeakCAPITALS, (capitalize ? 3 : 0), 0);
//debug_log(logfd, "In tts_capitalize espeak_SetParameter returned: %d\n", erc);
return;
} /* end tts_capitalize */
static espeak_ERROR set_pitch(struct synth_t *s, int pitch,
enum adjust_t adj)
{
espeak_ERROR rc;
if (adj == ADJ_DEC)
pitch = -pitch;
if (adj != ADJ_SET)
pitch += s->pitch;
rc = espeak_SetParameter(espeakPITCH, pitch * pitchMultiplier, 0);
if (rc == EE_OK)
s->pitch = pitch;
return rc;
}
static espeak_ERROR set_frequency(struct synth_t *s, int freq,
enum adjust_t adj)
{
espeak_ERROR rc;
if (adj == ADJ_DEC)
freq = -freq;
if (adj != ADJ_SET)
freq += s->frequency;
rc = espeak_SetParameter(espeakRANGE, freq * frequencyMultiplier, 0);
if (rc == EE_OK)
s->frequency = freq;
return rc;
}
void PlatformSpeechSynthesisProviderEfl::speak(PassRefPtr<PlatformSpeechSynthesisUtterance> utterance)
{
if (!engineInit() || !utterance) {
fireSpeechEvent(SpeechError);
return;
}
m_utterance = utterance;
String voice = voiceName(m_utterance);
espeak_SetVoiceByName(voice.utf8().data());
espeak_SetParameter(espeakRATE, convertRateToEspeakValue(m_utterance->rate()), 0);
espeak_SetParameter(espeakVOLUME, convertVolumeToEspeakValue(m_utterance->volume()), 0);
espeak_SetParameter(espeakPITCH, convertPitchToEspeakValue(m_utterance->pitch()), 0);
String textToRead = m_utterance->text();
espeak_ERROR err = espeak_Synth(textToRead.utf8().data(), textToRead.length(), 0, POS_CHARACTER, 0, espeakCHARS_AUTO, 0, nullptr);
if (err == EE_INTERNAL_ERROR) {
fireSpeechEvent(SpeechError);
m_utterance = nullptr;
return;
}
fireSpeechEvent(SpeechStart);
}
void tts_split_caps(int split_caps)
{
/* speak camel-case, IOW say 'capital' for every capital letter */
espeak_ERROR erc;
//debug_log(logfd, "Called tts_split_caps: %d\n", split_caps);
tts_state.split_caps = split_caps;
if (split_caps) {
tts_capitalize(0);
tts_allcaps_beep(0);
}
erc = espeak_SetParameter(espeakCAPITALS, (split_caps ? 2 : 0), 0);
//debug_log(logfd, "In tts_split_caps espeak_SetParameter returned: %d\n", erc);
return;
} /* end tts_split_caps */
int setRange(int range) {
espeak_ERROR rc;
if ( 100 < range ) {
range = 100;
} else if ( 0 > range ) {
range = 0;
}
if ( 1 == initFlag ) {
rc = espeak_SetParameter(espeakRANGE, range, 0);
if ( EE_BUFFER_FULL == rc ) {
sleep(3);
rc = espeak_SetParameter(espeakRANGE, range, 0);
if ( EE_OK != rc ) {
return -1;
}
}
if ( EE_INTERNAL_ERROR == rc ) {
return -1;
}
}
return 0;
}
static espeak_ERROR set_rate(struct synth_t *s, int rate,
enum adjust_t adj)
{
espeak_ERROR rc;
if (adj == ADJ_DEC)
rate = -rate;
if (adj != ADJ_SET)
rate += s->rate;
rc = espeak_SetParameter(espeakRATE,
rate * rateMultiplier + rateOffset, 0);
if (rc == EE_OK)
s->rate = rate;
return rc;
}
int setPunctuation(int punct) {
espeak_ERROR rc;
if ( 0 != punct
|| 1 != punct
|| 2 != punct ) {
punct = 2;
}
if ( 1 == initFlag ) {
rc = espeak_SetParameter(espeakPUNCTUATION, punct, 0);
if ( EE_BUFFER_FULL == rc ) {
sleep(3);
rc = espeak_SetParameter(espeakPUNCTUATION, punct, 0);
if ( EE_OK != rc ) {
return -1;
}
}
if ( EE_INTERNAL_ERROR == rc ) {
return -1;
}
}
return 0;
}
int setPitch(int pitch) {
espeak_ERROR rc;
if ( 100 < pitch ) {
pitch = 100;
} else if ( 0 > pitch ) {
pitch = 0;
}
if ( 1 == initFlag ) {
rc = espeak_SetParameter(espeakPITCH, pitch, 0);
if ( EE_BUFFER_FULL == rc ) {
sleep(3);
rc = espeak_SetParameter(espeakPITCH, pitch, 0);
if ( EE_OK != rc ) {
return -1;
}
}
if ( EE_INTERNAL_ERROR == rc ) {
return -1;
}
}
return 0;
}
static espeak_ERROR set_volume(struct synth_t *s, int vol,
enum adjust_t adj)
{
espeak_ERROR rc;
if (adj == ADJ_DEC)
vol = -vol;
if (adj != ADJ_SET)
vol += s->volume;
rc = espeak_SetParameter(espeakVOLUME, (vol + 1) * volumeMultiplier,
0);
if (rc == EE_OK)
s->volume = vol;
return rc;
}
static void
espeak_set_rate(signed int rate)
{
assert(rate >= -100 && rate <= +100);
int speed;
int normal_rate = 170, max_rate = 390, min_rate = 80;
if (rate < 0)
speed = normal_rate + (normal_rate - min_rate) * rate / 100;
else
speed = normal_rate + (max_rate - normal_rate) * rate / 100;
espeak_ERROR ret = espeak_SetParameter(espeakRATE, speed, 0);
if (ret != EE_OK) {
DBG("Espeak: Error setting rate %i.", speed);
} else {
DBG("Espeak: Rate set to %i.", speed);
}
}
static void
espeak_set_pitch(signed int pitch)
{
assert(pitch >= -100 && pitch <= +100);
int pitchBaseline;
/* Possible range 0 to 100. */
if (pitch < 0) {
pitchBaseline = ((float)(pitch + 100) * espeak_voice_pitch_baseline) / (float)100;
} else {
pitchBaseline = (((float)pitch * (100 - espeak_voice_pitch_baseline))
/ (float)100) + espeak_voice_pitch_baseline;
}
assert (pitchBaseline >= 0 && pitchBaseline <= 100);
espeak_ERROR ret = espeak_SetParameter(espeakPITCH, pitchBaseline, 0);
if (ret != EE_OK) {
DBG("Espeak: Error setting pitch %i.", pitchBaseline);
} else {
DBG("Espeak: Pitch set to %i.", pitchBaseline);
}
}
static void espeak_set_volume(signed int volume)
{
int vol;
assert(volume >= OTTS_VOICE_VOLUME_MIN
&& volume <= OTTS_VOICE_VOLUME_MAX);
/* Possible range 0 to 100. */
if (volume < OTTS_VOICE_VOLUME_DEFAULT)
vol = (volume - OTTS_VOICE_VOLUME_MIN) * 100 /
(OTTS_VOICE_VOLUME_MAX - OTTS_VOICE_VOLUME_MIN);
else
vol = 50 + ((volume - OTTS_VOICE_VOLUME_DEFAULT) * 100 /
(OTTS_VOICE_VOLUME_MAX - OTTS_VOICE_VOLUME_MIN));
espeak_ERROR ret = espeak_SetParameter(espeakVOLUME, vol, 0);
if (ret != EE_OK) {
log_msg(OTTS_LOG_WARN, "Espeak: Error setting volume %i.", vol);
} else {
log_msg(OTTS_LOG_DEBUG, "Espeak: Volume set to %i.", vol);
}
}
static void
espeak_set_cap_let_recogn(ECapLetRecogn cap_mode)
{
int espeak_cap_mode = 0;
switch (cap_mode) {
case RECOGN_NONE:
espeak_cap_mode = EspeakCapitalPitchRise;
break;
case RECOGN_SPELL:
espeak_cap_mode = 2;
break;
case RECOGN_ICON:
espeak_cap_mode = 1;
break;
}
espeak_ERROR ret = espeak_SetParameter(espeakCAPITALS, espeak_cap_mode, 1);
if (ret != EE_OK) {
DBG("Espeak: Failed to set capitals mode.");
} else {
DBG("Set capitals mode.");
}
}
static void
espeak_set_punctuation_mode(EPunctMode punct_mode)
{
espeak_PUNCT_TYPE espeak_punct_mode = espeakPUNCT_SOME;
switch (punct_mode) {
case PUNCT_ALL:
espeak_punct_mode = espeakPUNCT_ALL;
break;
case PUNCT_SOME:
espeak_punct_mode = espeakPUNCT_SOME;
break;
case PUNCT_NONE:
espeak_punct_mode = espeakPUNCT_NONE;
break;
}
espeak_ERROR ret = espeak_SetParameter(espeakPUNCTUATION, espeak_punct_mode, 0);
if (ret != EE_OK) {
DBG("Espeak: Failed to set punctuation mode.");
} else {
DBG("Set punctuation mode.");
}
}
static void espeak_set_rate(signed int rate)
{
int speed;
int normal_rate = 170, max_rate = 390, min_rate = 80;
assert(rate >= OTTS_VOICE_RATE_MIN && rate <= OTTS_VOICE_RATE_MAX);
if (rate < OTTS_VOICE_RATE_DEFAULT)
speed = min_rate + ((normal_rate - min_rate)
* (rate - OTTS_VOICE_RATE_MIN)
/ (OTTS_VOICE_RATE_DEFAULT -
OTTS_VOICE_RATE_MIN));
else
speed = normal_rate + ((max_rate - normal_rate)
* (rate - OTTS_VOICE_RATE_DEFAULT)
/ (OTTS_VOICE_RATE_MAX -
OTTS_VOICE_RATE_DEFAULT));
espeak_ERROR ret = espeak_SetParameter(espeakRATE, speed, 0);
if (ret != EE_OK) {
log_msg(OTTS_LOG_WARN, "Espeak: Error setting rate %i.", speed);
} else {
log_msg(OTTS_LOG_DEBUG, "Espeak: Rate set to %i.", speed);
}
}
ESpeak::ESpeak(QObject *parent) :
QObject(parent),
queueLength(0)
{
if (0 == refcount)
{
espeak_Initialize(AUDIO_OUTPUT_PLAYBACK, 100, 0, 0);
espeak_VOICE voice;
QString language = QLocale::system().languageToString(QLocale::system().language());
QByteArray languageData = language.toAscii();
languageData.prepend((char)0);
languageData.append((char)0);
QByteArray languages = QByteArray(1, (char)0);
languageData += languages;
voice.languages = languageData.data();
voice.name = 0;
voice.gender = 2;
voice.age = 20;
espeak_SetVoiceByProperties(&voice);
espeak_SetParameter(espeakRATE, 175, false);
}
++refcount;
}
// Sets the property with the specified value
//
// TODO: add pitch property here
tts_result TtsEngine::setProperty(const char *property, const char *value, const size_t size)
{
int rate;
int pitch;
int volume;
/* Set a specific property for the engine.
Supported properties include: language (locale), rate, pitch, volume. */
/* Sanity check */
if (property == NULL) {
LOGE("setProperty called with property NULL");
return TTS_PROPERTY_UNSUPPORTED;
}
if (value == NULL) {
LOGE("setProperty called with value NULL");
return TTS_VALUE_INVALID;
}
if (strncmp(property, "language", 8) == 0) {
// TODO: Fix this
return TTS_SUCCESS;
} else if (strncmp(property, "rate", 4) == 0) {
rate = atoi(value);
espeak_SetParameter(espeakRATE, rate, 0);
// TODO: Fix this - use the return value here, don't just automatically return success!
return TTS_SUCCESS;
} else if (strncmp(property, "pitch", 5) == 0) {
// TODO: Fix this
return TTS_SUCCESS;
} else if (strncmp(property, "volume", 6) == 0) {
// TODO: Fix this
return TTS_SUCCESS;
}
return TTS_PROPERTY_UNSUPPORTED;
}
static void
spk_setVolume(volatile SpeechSynthesizer *spk, unsigned char setting)
{
int volume = getIntegerSpeechVolume(setting, 50);
espeak_SetParameter(espeakVOLUME, volume, 0);
}
int main( int argc, char** argv ) {
espeak_Initialize(AUDIO_OUTPUT_PLAYBACK, 0, NULL, 0);
ros::init(argc, argv, "espeak_node");
ros::NodeHandle n;
ros::NodeHandle priv_n("~");
// for list of voices, run "espeak --voices"
// voices: "brazil", "default", "english", "lancashire",
// "english-rp", "english-wmids", "english-us", "en-scottish"
std::string voice_name;
// dialects: "pt-br", "en", "en-uk", "en-uk-north",
// "en-uk-rp", "en-uk-wmids", "en-us", "en-sc"
std::string dialect;
espeak_VOICE voice_select;
// See speak_lib.h for details and range of the parameters below
int rate, volume, pitch, range, punctuation;
int voice_num, dialect_num, capitals, wordgap, age, gender;
// Fetch default parameters
rate = espeak_GetParameter(espeakRATE, 0);
volume = espeak_GetParameter(espeakVOLUME, 0);
pitch = espeak_GetParameter(espeakPITCH, 0);
range = espeak_GetParameter(espeakRANGE, 0);
punctuation = espeak_GetParameter(espeakPUNCTUATION, 0);
capitals = espeak_GetParameter(espeakCAPITALS, 0);
wordgap = espeak_GetParameter(espeakWORDGAP, 0);
voice_num = 1;
dialect_num = 1;
// Fetch (and set if param could not be retrieved) ROS parameters
if (priv_n.getParam("voice", voice_num) == false) {
priv_n.setParam("voice", voice_num);
}
if (priv_n.getParam("dialect", dialect_num) == false) {
priv_n.setParam("dialect", dialect_num);
}
if (priv_n.getParam("rate", rate) == false) {
priv_n.setParam("rate", rate);
}
if (priv_n.getParam("volume", volume) == false) {
priv_n.setParam("volume", volume);
}
if (priv_n.getParam("pitch", pitch) == false) {
priv_n.setParam("pitch", pitch);
}
if (priv_n.getParam("punctuation", punctuation) == false) {
priv_n.setParam("punctuation", punctuation);
}
if (priv_n.getParam("capitals", capitals) == false) {
priv_n.setParam("capitals", capitals);
}
if (priv_n.getParam("wordgap", wordgap) == false) {
priv_n.setParam("wordgap", wordgap);
}
priv_n.param("age", age, int(0));
priv_n.param("gender", gender, int(2));
if (age < 0 || age > 100) {
age = 0;
}
if (gender < 0 || gender > 2) {
gender = 2;
}
voice_name = getVoiceName(voice_num);
dialect = getDialectName(dialect_num);
std::memset(&voice_select,0,sizeof(voice_select));
voice_select.name = voice_name.c_str();
voice_select.languages = dialect.c_str();
voice_select.age = age; // 0=not specified, or age in years
voice_select.gender = gender; // 0=none, 1=male, 2=female
// set espeak parameters
//if (espeak_SetVoiceByName(voice_name.c_str()) != EE_OK) {
// ROS_ERROR("Could not set espeak voice. Aborting.");
// return 1;
//}
if (espeak_SetVoiceByProperties(&voice_select) != EE_OK) {
ROS_ERROR("Could not set espeak voice properties. Aborting.");
return 1;
}
//ROS_INFO("Using voice %s", voice_name.c_str());
//ROS_INFO("Using dialect %s", dialect.c_str());
if (rate < 80 || rate > 450) {
ROS_INFO("Parameter rate = %d out of range, using default value", rate);
}
else {
if (espeak_SetParameter(espeakRATE, rate, 0) != EE_OK) {
ROS_ERROR("Could not set espeak rate. Aborting.");
return 1;
}
}
if (volume < 0 || volume > 200) {
ROS_INFO("Parameter volume = %d out of range, using default value", volume);
}
else {
if (espeak_SetParameter(espeakVOLUME, volume, 0) != EE_OK) {
ROS_ERROR("Could not set espeak volume. Aborting.");
return 1;
}
}
if (pitch < 0 || pitch > 100) {
ROS_INFO("Parameter pitch = %d out of range, using default value", pitch);
}
else {
if (espeak_SetParameter(espeakPITCH, pitch, 0) != EE_OK) {
ROS_ERROR("Could not set espeak pitch. Aborting.");
return 1;
}
}
if (range < 0 || range > 100) {
ROS_INFO("Parameter range = %d out of range, using default value", range);
}
else {
if (espeak_SetParameter(espeakRANGE, range, 0) != EE_OK) {
ROS_ERROR("Could not set espeak range. Aborting.");
return 1;
}
}
if (punctuation < 0 || punctuation > 2) {
ROS_INFO("Parameter punctuation out of range, using default value");
}
else {
if (espeak_SetParameter(espeakPUNCTUATION,
espeak_PUNCT_TYPE(punctuation), 0) != EE_OK) {
ROS_ERROR("Could not set espeak punctuation. Aborting.");
return 1;
}
}
if (capitals < 0 || capitals > 3) {
ROS_INFO("Parameter capitals out of range, using default value");
}
else {
if (espeak_SetParameter(espeakCAPITALS, capitals, 0) != EE_OK) {
ROS_ERROR("Could not set espeak capitals. Aborting.");
return 1;
}
}
if (wordgap < 0 || wordgap > 1000 || wordgap % 10 != 0) {
ROS_INFO("Parameter wordgap out of range, using default value");
}
else {
if (espeak_SetParameter(espeakWORDGAP, wordgap, 0) != EE_OK) {
ROS_ERROR("Could not set espeak wordgap. Aborting.");
return 1;
}
}
dynamic_reconfigure::Server<espeak_ros::EspeakConfig> server;
dynamic_reconfigure::Server<espeak_ros::EspeakConfig>::CallbackType f;
f = boost::bind(&dyn_cfg_callback, _1, _2);
server.setCallback(f);
ros::Subscriber sub = n.subscribe("/espeak_node/speak_line", 20, espeak_callback);
ros::spin();
while (n.ok());
return 0;
}
int main (int argc, char **argv)
//==============================
{
static struct option long_options[] =
{
/* These options set a flag. */
// {"verbose", no_argument, &verbose_flag, 1},
// {"brief", no_argument, &verbose_flag, 0},
/* These options don't set a flag.
We distinguish them by their indices. */
{"help", no_argument, 0, 'h'},
{"stdin", no_argument, 0, 0x100},
{"compile-debug", optional_argument, 0, 0x101},
{"compile", optional_argument, 0, 0x102},
{"punct", optional_argument, 0, 0x103},
{"voices", optional_argument, 0, 0x104},
{"stdout", no_argument, 0, 0x105},
{"split", optional_argument, 0, 0x106},
{0, 0, 0, 0}
};
static const char* err_load = "Failed to read ";
FILE *f_text=NULL;
char *p_text=NULL;
int option_index = 0;
int c;
int ix;
int flag_stdin = 0;
int flag_compile = 0;
int filesize = 0;
int synth_flags = espeakCHARS_AUTO | espeakPHONEMES | espeakENDPAUSE;
int volume = -1;
int speed = -1;
int pitch = -1;
int wordgap = -1;
int option_capitals = -1;
int option_punctuation = -1;
int option_phonemes = -1;
int option_linelength = 0;
int option_waveout = 0;
char filename[120];
char voicename[40];
char voice_mbrola[20];
char dictname[40];
#define N_PUNCTLIST 100
wchar_t option_punctlist[N_PUNCTLIST];
voicename[0] = 0;
voice_mbrola[0] = 0;
dictname[0] = 0;
wavefile[0] = 0;
filename[0] = 0;
option_punctlist[0] = 0;
while(true)
{
c = getopt_long (argc, argv, "a:bf:g:hk:l:mp:qs:v:w:xXz",
long_options, &option_index);
/* Detect the end of the options. */
if (c == -1)
break;
switch (c)
{
case 'b':
synth_flags |= espeakCHARS_8BIT;
break;
case 'h':
printf("\n");
printf("eSpeak text-to-speech: %s\n%s",espeak_Info(NULL),help_text);
exit(0);
break;
case 'k':
option_capitals = atoi(optarg);
break;
case 'x':
option_phonemes = 1;
break;
case 'X':
option_phonemes = 2;
break;
case 'm':
synth_flags |= espeakSSML;
break;
case 'p':
pitch = atoi(optarg);
break;
case 'q':
quiet = 1;
break;
case 'f':
strncpy0(filename,optarg,sizeof(filename));
break;
case 'l':
option_linelength = atoi(optarg);
break;
case 'a':
volume = atoi(optarg);
break;
case 's':
speed = atoi(optarg);
break;
case 'g':
wordgap = atoi(optarg);
break;
case 'v':
strncpy0(voicename,optarg,sizeof(voicename));
break;
case 'w':
option_waveout = 1;
strncpy0(wavefile,optarg,sizeof(filename));
break;
case 'z': // remove pause from the end of a sentence
synth_flags &= ~espeakENDPAUSE;
break;
case 0x100: // --stdin
flag_stdin = 1;
break;
case 0x105: // --stdout
option_waveout = 1;
strcpy(wavefile,"stdout");
break;
case 0x101: // --compile-debug
case 0x102: // --compile
strncpy0(voicename,optarg,sizeof(voicename));
flag_compile = c;
quiet = 1;
break;
case 0x103: // --punct
option_punctuation = 1;
if(optarg != NULL)
{
ix = 0;
while((ix < N_PUNCTLIST) && ((option_punctlist[ix] = optarg[ix]) != 0)) ix++;
option_punctlist[N_PUNCTLIST-1] = 0;
option_punctuation = 2;
}
break;
case 0x104: // --voices
espeak_Initialize(AUDIO_OUTPUT_SYNCHRONOUS,0,NULL,0);
DisplayVoices(stdout,optarg);
exit(0);
case 0x106: // -- split
if(optarg == NULL)
samples_split = 30; // default 30 minutes
else
samples_split = atoi(optarg);
break;
default:
exit(0);
}
}
if(option_waveout || quiet)
{
// writing to a file (or no output), we can use synchronous mode
samplerate = espeak_Initialize(AUDIO_OUTPUT_SYNCHRONOUS,0,NULL,0);
samples_split = (samplerate * samples_split) * 60;
espeak_SetSynthCallback(SynthCallback);
if(samples_split)
{
char *extn;
extn = strrchr(wavefile,'.');
if((extn != NULL) && ((wavefile + strlen(wavefile) - extn) <= 4))
{
strcpy(filetype,extn);
*extn = 0;
}
}
else
if(option_waveout)
{
if(OpenWavFile(wavefile,samplerate) != 0)
exit(4);
}
}
else
{
// play the sound output
samplerate = espeak_Initialize(AUDIO_OUTPUT_PLAYBACK,0,NULL,0);
}
if(voicename[0] == 0)
strcpy(voicename,"default");
if(espeak_SetVoiceByName(voicename) != EE_OK)
{
fprintf(stderr,"%svoice '%s'\n",err_load,voicename);
exit(2);
}
if(flag_compile)
{
// This must be done after the voice is set
espeak_CompileDictionary("", stderr, flag_compile & 0x1);
exit(0);
}
// set any non-default values of parameters. This must be done after espeak_Initialize()
if(speed > 0)
espeak_SetParameter(espeakRATE,speed,0);
if(volume >= 0)
espeak_SetParameter(espeakVOLUME,volume,0);
if(pitch >= 0)
espeak_SetParameter(espeakPITCH,pitch,0);
if(option_capitals >= 0)
espeak_SetParameter(espeakCAPITALS,option_capitals,0);
if(option_punctuation >= 0)
espeak_SetParameter(espeakPUNCTUATION,option_punctuation,0);
if(wordgap >= 0)
espeak_SetParameter(espeakWORDGAP,wordgap,0);
if(option_linelength > 0)
espeak_SetParameter(espeakLINELENGTH,option_linelength,0);
if(option_punctuation == 2)
espeak_SetPunctuationList(option_punctlist);
if(option_phonemes >= 0)
espeak_SetPhonemeTrace(option_phonemes,stderr);
if(filename[0]==0)
{
if((optind < argc) && (flag_stdin == 0))
{
// there's a non-option parameter, and no -f or --stdin
// use it as text
p_text = argv[optind];
}
else
{
f_text = stdin;
if(flag_stdin == 0)
{
flag_stdin = 2;
}
}
}
else
{
filesize = GetFileLength(filename);
f_text = fopen(filename,"r");
}
if((f_text == NULL) && (p_text == NULL))
{
fprintf(stderr,"%sfile '%s'\n",err_load,filename);
exit(1);
}
if(p_text != NULL)
{
int size;
size = strlen(p_text);
espeak_Synth(p_text,size+1,0,POS_CHARACTER,0,synth_flags,NULL,NULL);
}
else
if(flag_stdin)
{
int max = 1000;
p_text = (char *)malloc(max);
if(flag_stdin == 2)
{
// line by line input on stdin
while(fgets(p_text,max,stdin) != NULL)
{
p_text[max-1] = 0;
espeak_Synth(p_text,max,0,POS_CHARACTER,0,synth_flags,NULL,NULL);
}
}
else
{
// bulk input on stdin
ix = 0;
while(!feof(stdin))
{
p_text[ix++] = fgetc(stdin);
if(ix >= (max-1))
{
max += 1000;
p_text = (char *)realloc(p_text,max);
}
}
if(ix > 0)
{
p_text[ix-1] = 0;
espeak_Synth(p_text,ix+1,0,POS_CHARACTER,0,synth_flags,NULL,NULL);
}
}
}
else
if(f_text != NULL)
{
if((p_text = (char *)malloc(filesize+1)) == NULL)
{
fprintf(stderr,"Failed to allocate memory %d bytes",filesize);
exit(3);
}
fread(p_text,1,filesize,f_text);
p_text[filesize]=0;
espeak_Synth(p_text,filesize+1,0,POS_CHARACTER,0,synth_flags,NULL,NULL);
fclose(f_text);
}
espeak_Synchronize();
return(0);
}
int main(int argc, char **argv)
{
static struct option long_options[] = {
{ "help", no_argument, 0, 'h' },
{ "stdin", no_argument, 0, 0x100 },
{ "compile-debug", optional_argument, 0, 0x101 },
{ "compile", optional_argument, 0, 0x102 },
{ "punct", optional_argument, 0, 0x103 },
{ "voices", optional_argument, 0, 0x104 },
{ "stdout", no_argument, 0, 0x105 },
{ "split", optional_argument, 0, 0x106 },
{ "path", required_argument, 0, 0x107 },
{ "phonout", required_argument, 0, 0x108 },
{ "pho", no_argument, 0, 0x109 },
{ "ipa", optional_argument, 0, 0x10a },
{ "version", no_argument, 0, 0x10b },
{ "sep", optional_argument, 0, 0x10c },
{ "tie", optional_argument, 0, 0x10d },
{ "compile-mbrola", optional_argument, 0, 0x10e },
{ "compile-intonations", no_argument, 0, 0x10f },
{ "compile-phonemes", optional_argument, 0, 0x110 },
{ 0, 0, 0, 0 }
};
FILE *f_text = NULL;
char *p_text = NULL;
FILE *f_phonemes_out = stdout;
char *data_path = NULL; // use default path for espeak-ng-data
int option_index = 0;
int c;
int ix;
char *optarg2;
int value;
int flag_stdin = 0;
int flag_compile = 0;
int filesize = 0;
int synth_flags = espeakCHARS_AUTO | espeakPHONEMES | espeakENDPAUSE;
int volume = -1;
int speed = -1;
int pitch = -1;
int wordgap = -1;
int option_capitals = -1;
int option_punctuation = -1;
int phonemes_separator = 0;
int phoneme_options = 0;
int option_linelength = 0;
int option_waveout = 0;
espeak_VOICE voice_select;
char filename[200];
char voicename[40];
char devicename[200];
#define N_PUNCTLIST 100
wchar_t option_punctlist[N_PUNCTLIST];
voicename[0] = 0;
wavefile[0] = 0;
filename[0] = 0;
devicename[0] = 0;
option_punctlist[0] = 0;
while (true) {
c = getopt_long(argc, argv, "a:b:d:f:g:hk:l:mp:qs:v:w:xXz",
long_options, &option_index);
// Detect the end of the options.
if (c == -1)
break;
optarg2 = optarg;
switch (c)
{
case 'b':
// input character encoding, 8bit, 16bit, UTF8
if ((sscanf(optarg2, "%d", &value) == 1) && (value <= 4))
synth_flags |= value;
else
synth_flags |= espeakCHARS_8BIT;
break;
case 'd':
strncpy0(devicename, optarg2, sizeof(devicename));
break;
case 'h':
printf("\n");
PrintVersion();
printf("%s", help_text);
return 0;
case 'k':
option_capitals = atoi(optarg2);
break;
case 'x':
phoneme_options |= espeakPHONEMES_SHOW;
break;
case 'X':
phoneme_options |= espeakPHONEMES_TRACE;
break;
case 'm':
synth_flags |= espeakSSML;
break;
case 'p':
pitch = atoi(optarg2);
break;
case 'q':
quiet = true;
break;
case 'f':
strncpy0(filename, optarg2, sizeof(filename));
break;
case 'l':
option_linelength = atoi(optarg2);
break;
case 'a':
volume = atoi(optarg2);
break;
case 's':
speed = atoi(optarg2);
break;
case 'g':
wordgap = atoi(optarg2);
break;
case 'v':
strncpy0(voicename, optarg2, sizeof(voicename));
break;
case 'w':
option_waveout = 1;
strncpy0(wavefile, optarg2, sizeof(filename));
break;
case 'z': // remove pause from the end of a sentence
synth_flags &= ~espeakENDPAUSE;
break;
case 0x100: // --stdin
flag_stdin = 1;
break;
case 0x105: // --stdout
option_waveout = 1;
strcpy(wavefile, "stdout");
break;
case 0x101: // --compile-debug
case 0x102: // --compile
if (optarg2 != NULL && *optarg2) {
strncpy0(voicename, optarg2, sizeof(voicename));
flag_compile = c;
quiet = true;
break;
} else {
fprintf(stderr, "Voice name to '%s' not specified.\n", c == 0x101 ? "--compile-debug" : "--compile");
exit(EXIT_FAILURE);
}
case 0x103: // --punct
option_punctuation = 1;
if (optarg2 != NULL) {
ix = 0;
while ((ix < N_PUNCTLIST) && ((option_punctlist[ix] = optarg2[ix]) != 0)) ix++;
option_punctlist[N_PUNCTLIST-1] = 0;
option_punctuation = 2;
}
break;
case 0x104: // --voices
espeak_Initialize(AUDIO_OUTPUT_SYNCHRONOUS, 0, data_path, 0);
DisplayVoices(stdout, optarg2);
exit(0);
case 0x106: // -- split
if (optarg2 == NULL)
samples_split_seconds = 30 * 60; // default 30 minutes
else
samples_split_seconds = atoi(optarg2) * 60;
break;
case 0x107: // --path
data_path = optarg2;
break;
case 0x108: // --phonout
if ((f_phonemes_out = fopen(optarg2, "w")) == NULL)
fprintf(stderr, "Can't write to: %s\n", optarg2);
break;
case 0x109: // --pho
phoneme_options |= espeakPHONEMES_MBROLA;
break;
case 0x10a: // --ipa
phoneme_options |= espeakPHONEMES_IPA;
if (optarg2 != NULL) {
// deprecated and obsolete
switch (atoi(optarg2))
{
case 1:
phonemes_separator = '_';
break;
case 2:
phonemes_separator = 0x0361;
phoneme_options |= espeakPHONEMES_TIE;
break;
case 3:
phonemes_separator = 0x200d; // ZWJ
phoneme_options |= espeakPHONEMES_TIE;
break;
}
}
break;
case 0x10b: // --version
PrintVersion();
exit(0);
case 0x10c: // --sep
phoneme_options |= espeakPHONEMES_SHOW;
if (optarg2 == 0)
phonemes_separator = ' ';
else
utf8_in(&phonemes_separator, optarg2);
if (phonemes_separator == 'z')
phonemes_separator = 0x200c; // ZWNJ
break;
case 0x10d: // --tie
phoneme_options |= (espeakPHONEMES_SHOW | espeakPHONEMES_TIE);
if (optarg2 == 0)
phonemes_separator = 0x0361; // default: combining-double-inverted-breve
else
utf8_in(&phonemes_separator, optarg2);
if (phonemes_separator == 'z')
phonemes_separator = 0x200d; // ZWJ
break;
case 0x10e: // --compile-mbrola
{
espeak_ng_InitializePath(data_path);
espeak_ng_ERROR_CONTEXT context = NULL;
espeak_ng_STATUS result = espeak_ng_CompileMbrolaVoice(optarg2, stdout, &context);
if (result != ENS_OK) {
espeak_ng_PrintStatusCodeMessage(result, stderr, context);
espeak_ng_ClearErrorContext(&context);
return EXIT_FAILURE;
}
return EXIT_SUCCESS;
}
case 0x10f: // --compile-intonations
{
espeak_ng_InitializePath(data_path);
espeak_ng_ERROR_CONTEXT context = NULL;
espeak_ng_STATUS result = espeak_ng_CompileIntonation(stdout, &context);
if (result != ENS_OK) {
espeak_ng_PrintStatusCodeMessage(result, stderr, context);
espeak_ng_ClearErrorContext(&context);
return EXIT_FAILURE;
}
return EXIT_SUCCESS;
}
case 0x110: // --compile-phonemes
{
espeak_ng_InitializePath(data_path);
espeak_ng_ERROR_CONTEXT context = NULL;
espeak_ng_STATUS result;
if (optarg2) {
result = espeak_ng_CompilePhonemeDataPath(22050, optarg2, NULL, stdout, &context);
} else {
result = espeak_ng_CompilePhonemeData(22050, stdout, &context);
}
if (result != ENS_OK) {
espeak_ng_PrintStatusCodeMessage(result, stderr, context);
espeak_ng_ClearErrorContext(&context);
return EXIT_FAILURE;
}
return EXIT_SUCCESS;
}
default:
exit(0);
}
}
espeak_ng_InitializePath(data_path);
espeak_ng_ERROR_CONTEXT context = NULL;
espeak_ng_STATUS result = espeak_ng_Initialize(&context);
if (result != ENS_OK) {
espeak_ng_PrintStatusCodeMessage(result, stderr, context);
espeak_ng_ClearErrorContext(&context);
exit(1);
}
if (option_waveout || quiet) {
// writing to a file (or no output), we can use synchronous mode
result = espeak_ng_InitializeOutput(ENOUTPUT_MODE_SYNCHRONOUS, 0, devicename[0] ? devicename : NULL);
samplerate = espeak_ng_GetSampleRate();
samples_split = samplerate * samples_split_seconds;
espeak_SetSynthCallback(SynthCallback);
if (samples_split) {
char *extn;
extn = strrchr(wavefile, '.');
if ((extn != NULL) && ((wavefile + strlen(wavefile) - extn) <= 4)) {
strcpy(filetype, extn);
*extn = 0;
}
}
} else {
// play the sound output
result = espeak_ng_InitializeOutput(PLAYBACK_MODE, 0, devicename[0] ? devicename : NULL);
samplerate = espeak_ng_GetSampleRate();
}
if (result != ENS_OK) {
espeak_ng_PrintStatusCodeMessage(result, stderr, NULL);
exit(EXIT_FAILURE);
}
if (voicename[0] == 0)
strcpy(voicename, ESPEAKNG_DEFAULT_VOICE);
result = espeak_ng_SetVoiceByName(voicename);
if (result != ENS_OK) {
memset(&voice_select, 0, sizeof(voice_select));
voice_select.languages = voicename;
result = espeak_ng_SetVoiceByProperties(&voice_select);
if (result != ENS_OK) {
espeak_ng_PrintStatusCodeMessage(result, stderr, NULL);
exit(EXIT_FAILURE);
}
}
if (flag_compile) {
// This must be done after the voice is set
espeak_ng_ERROR_CONTEXT context = NULL;
espeak_ng_STATUS result = espeak_ng_CompileDictionary("", NULL, stderr, flag_compile & 0x1, &context);
if (result != ENS_OK) {
espeak_ng_PrintStatusCodeMessage(result, stderr, context);
espeak_ng_ClearErrorContext(&context);
return EXIT_FAILURE;
}
return EXIT_SUCCESS;
}
// set any non-default values of parameters. This must be done after espeak_Initialize()
if (speed > 0)
espeak_SetParameter(espeakRATE, speed, 0);
if (volume >= 0)
espeak_SetParameter(espeakVOLUME, volume, 0);
if (pitch >= 0)
espeak_SetParameter(espeakPITCH, pitch, 0);
if (option_capitals >= 0)
espeak_SetParameter(espeakCAPITALS, option_capitals, 0);
if (option_punctuation >= 0)
espeak_SetParameter(espeakPUNCTUATION, option_punctuation, 0);
if (wordgap >= 0)
espeak_SetParameter(espeakWORDGAP, wordgap, 0);
if (option_linelength > 0)
espeak_SetParameter(espeakLINELENGTH, option_linelength, 0);
if (option_punctuation == 2)
espeak_SetPunctuationList(option_punctlist);
espeak_SetPhonemeTrace(phoneme_options | (phonemes_separator << 8), f_phonemes_out);
if (filename[0] == 0) {
if ((optind < argc) && (flag_stdin == 0)) {
// there's a non-option parameter, and no -f or --stdin
// use it as text
p_text = argv[optind];
} else {
f_text = stdin;
if (flag_stdin == 0)
flag_stdin = 2;
}
} else {
struct stat st;
if (stat(filename, &st) != 0) {
fprintf(stderr, "Failed to stat() file '%s'\n", filename);
exit(EXIT_FAILURE);
}
filesize = GetFileLength(filename);
f_text = fopen(filename, "r");
if (f_text == NULL) {
fprintf(stderr, "Failed to read file '%s'\n", filename);
exit(EXIT_FAILURE);
}
if (S_ISFIFO(st.st_mode)) {
flag_stdin = 2;
}
}
if (p_text != NULL) {
int size;
size = strlen(p_text);
espeak_Synth(p_text, size+1, 0, POS_CHARACTER, 0, synth_flags, NULL, NULL);
} else if (flag_stdin) {
size_t max = 1000;
if ((p_text = (char *)malloc(max)) == NULL) {
espeak_ng_PrintStatusCodeMessage(ENOMEM, stderr, NULL);
exit(EXIT_FAILURE);
}
if (flag_stdin == 2) {
// line by line input on stdin or from FIFO
while (fgets(p_text, max, f_text) != NULL) {
p_text[max-1] = 0;
espeak_Synth(p_text, max, 0, POS_CHARACTER, 0, synth_flags, NULL, NULL);
// Allow subprocesses to use the audio data through pipes.
fflush(stdout);
}
if (f_text != stdin) {
fclose(f_text);
}
} else {
// bulk input on stdin
ix = 0;
while (true) {
if ((c = fgetc(stdin)) == EOF)
break;
p_text[ix++] = (char)c;
if (ix >= (max-1)) {
char *new_text = NULL;
if (max <= SIZE_MAX - 1000) {
max += 1000;
new_text = (char *)realloc(p_text, max);
}
if (new_text == NULL) {
free(p_text);
espeak_ng_PrintStatusCodeMessage(ENOMEM, stderr, NULL);
exit(EXIT_FAILURE);
}
p_text = new_text;
}
}
if (ix > 0) {
p_text[ix-1] = 0;
espeak_Synth(p_text, ix+1, 0, POS_CHARACTER, 0, synth_flags, NULL, NULL);
}
}
free(p_text);
} else if (f_text != NULL) {
if ((p_text = (char *)malloc(filesize+1)) == NULL) {
espeak_ng_PrintStatusCodeMessage(ENOMEM, stderr, NULL);
exit(EXIT_FAILURE);
}
fread(p_text, 1, filesize, f_text);
p_text[filesize] = 0;
espeak_Synth(p_text, filesize+1, 0, POS_CHARACTER, 0, synth_flags, NULL, NULL);
fclose(f_text);
free(p_text);
}
result = espeak_ng_Synchronize();
if (result != ENS_OK) {
espeak_ng_PrintStatusCodeMessage(result, stderr, NULL);
exit(EXIT_FAILURE);
}
if (f_phonemes_out != stdout)
fclose(f_phonemes_out);
CloseWavFile();
espeak_ng_Terminate();
return 0;
}
/* Functions internal to the eSpeak engine wrapper */
static void setSpeechRate(int speechRate)
{
espeak_ERROR err = espeak_SetParameter(espeakRATE, speechRate, 0);
}
autoSound SpeechSynthesizer_to_Sound (SpeechSynthesizer me, const char32 *text, autoTextGrid *tg, autoTable *events) {
try {
int fsamp = espeak_Initialize (AUDIO_OUTPUT_SYNCHRONOUS, 0, nullptr, // 5000ms
espeakINITIALIZE_PHONEME_EVENTS|espeakINITIALIZE_PHONEME_IPA);
if (fsamp == -1) {
Melder_throw (U"Internal espeak error.");
}
int synth_flags = espeakCHARS_WCHAR;
if (my d_inputTextFormat == SpeechSynthesizer_INPUT_TAGGEDTEXT) {
synth_flags |= espeakSSML;
}
if (my d_inputTextFormat != SpeechSynthesizer_INPUT_TEXTONLY) {
synth_flags |= espeakPHONEMES;
}
option_phoneme_events = espeakINITIALIZE_PHONEME_EVENTS; // extern int option_phoneme_events;
if (my d_outputPhonemeCoding == SpeechSynthesizer_PHONEMECODINGS_IPA) {
option_phoneme_events |= espeakINITIALIZE_PHONEME_IPA;
}
espeak_SetParameter (espeakRATE, my d_wordsPerMinute, 0);
espeak_SetParameter (espeakPITCH, my d_pitchAdjustment, 0);
espeak_SetParameter (espeakRANGE, my d_pitchRange, 0);
const char32 *voiceLanguageCode = SpeechSynthesizer_getVoiceLanguageCodeFromName (me, my d_voiceLanguageName);
const char32 *voiceVariantCode = SpeechSynthesizer_getVoiceVariantCodeFromName (me, my d_voiceVariantName);
espeakdata_SetVoiceByName ((const char *) Melder_peek32to8 (voiceLanguageCode),
(const char *) Melder_peek32to8 (voiceVariantCode));
espeak_SetParameter (espeakWORDGAP, my d_wordgap * 100, 0); // espeak wordgap is in units of 10 ms
espeak_SetParameter (espeakCAPITALS, 0, 0);
espeak_SetParameter (espeakPUNCTUATION, espeakPUNCT_NONE, 0);
espeak_SetSynthCallback (synthCallback);
my d_events = Table_createWithColumnNames (0, U"time type type-t t-pos length a-pos sample id uniq");
#ifdef _WIN32
wchar_t *textW = Melder_peek32toW (text);
espeak_Synth (textW, wcslen (textW) + 1, 0, POS_CHARACTER, 0, synth_flags, nullptr, me);
#else
espeak_Synth (text, str32len (text) + 1, 0, POS_CHARACTER, 0, synth_flags, nullptr, me);
#endif
espeak_Terminate ();
autoSound thee = buffer_to_Sound (my d_wav, my d_numberOfSamples, my d_internalSamplingFrequency);
if (my d_samplingFrequency != my d_internalSamplingFrequency) {
thee = Sound_resample (thee.get(), my d_samplingFrequency, 50);
}
my d_numberOfSamples = 0; // re-use the wav-buffer
if (tg) {
double xmin = Table_getNumericValue_Assert (my d_events.get(), 1, 1);
if (xmin > thy xmin) {
xmin = thy xmin;
}
double xmax = Table_getNumericValue_Assert (my d_events.get(), my d_events -> rows.size, 1);
if (xmax < thy xmax) {
xmax = thy xmax;
}
autoTextGrid tg1 = Table_to_TextGrid (my d_events.get(), text, xmin, xmax);
*tg = TextGrid_extractPart (tg1.get(), thy xmin, thy xmax, 0);
}
if (events) {
Table_setEventTypeString (my d_events.get());
*events = my d_events.move();
}
my d_events.reset();
return thee;
} catch (MelderError) {
espeak_Terminate ();
Melder_throw (U"Text not played.");
}
}
/* Set the rate of TTS.
* Hear in case of espeak the rate is ranging
* from 80 to 450. So we multiply the given
* rate with 3.7 and add 80 to get exact
* rate for espeak */
void T4K_Tts_set_rate(int rate){
espeak_SetParameter(espeakRATE,(3.7*rate)+80,0);
}
void T4K_Tts_set_pitch(int pitch){
espeak_SetParameter(espeakPITCH,pitch,0);
}
//TTS Parameters
void T4K_Tts_set_volume(int volume){
espeak_SetParameter(espeakVOLUME,2*volume,0);
}
static void
spk_setPitch(volatile SpeechSynthesizer *spk, unsigned char setting)
{
int pitch = getIntegerSpeechPitch(setting, 50);
espeak_SetParameter(espeakPITCH, pitch, 0);
}
