ben.carrasco/freeswitch-modules
1
0
Fork 0
mirror of https://github.com/jambonz/freeswitch-modules.git synced 2025-12-19 09:07:44 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'main' of https://github.com/jambonz/freeswitch-modules into feat/mod_custom_tts

Browse Source
This commit is contained in:
Quan HL
2024-04-23 20:56:18 +07:00
parent cfe9cad816 33fee93ec7
commit 9874bf9ae0
7 changed files with 16 additions and 51 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
mod_audio_fork/audio_pipe.cpp
View File

@@ -480,7 +480,7 @@ AudioPipe::AudioPipe(const char* uuid, const char* host, unsigned int port, cons
}
AudioPipe::~AudioPipe() {
if (m_audio_buffer) delete [] m_audio_buffer;
if (m_recv_buf) delete [] m_recv_buf;
if (m_recv_buf) free(m_recv_buf);
}
void AudioPipe::connect(void) {

25
mod_azure_tts/azure_glue.cpp
View File

@@ -135,23 +135,12 @@ extern "C" {
return SWITCH_STATUS_FALSE;
}
if (a->session_id) {
if (a->rate != 8000 /*Hz*/) {
int err;
switch_codec_implementation_t read_impl;
/* lock and unlock session */
switch_core_session_t *psession = switch_core_session_locate(a->session_id);
switch_core_session_get_read_impl(psession, &read_impl);
switch_core_session_rwunlock(psession);
uint32_t samples_per_second = !strcasecmp(read_impl.iananame, "g722") ? read_impl.actual_samples_per_second : read_impl.samples_per_second;
a->samples_rate = samples_per_second;
if (samples_per_second != 8000 /*Hz*/) {
a->resampler = speex_resampler_init(1, 8000, samples_per_second, SWITCH_RESAMPLE_QUALITY, &err);
if (0 != err) {
switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_ERROR, "Error initializing resampler: %s.\n", speex_resampler_strerror(err));
return SWITCH_STATUS_FALSE;
}
a->resampler = speex_resampler_init(1, 8000, a->rate, SWITCH_RESAMPLE_QUALITY, &err);
if (0 != err) {
switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_ERROR, "Error initializing resampler: %s.\n", speex_resampler_strerror(err));
return SWITCH_STATUS_FALSE;
}
}
@@ -276,9 +265,7 @@ extern "C" {
return SWITCH_STATUS_SUCCESS;
}
// azure returned 8000hz 16 bit data, we have to take enough data based on call sample rate.
size_t size = a->samples_rate ?
std::min((*datalen/(2 * a->samples_rate / 8000)), bufSize) :
std::min((*datalen/2), bufSize);
size_t size = std::min((*datalen/(2 * a->rate / 8000)), bufSize);
pcm_data.insert(pcm_data.end(), cBuffer->begin(), cBuffer->begin() + size);
cBuffer->erase(cBuffer->begin(), cBuffer->begin() + size);
switch_mutex_unlock(a->mutex);

1
mod_azure_tts/mod_azure_tts.c
View File

@@ -82,7 +82,6 @@ static switch_status_t a_speech_feed_tts(switch_speech_handle_t *sh, char *text,
a->draining = 0;
a->reads = 0;
a->flushed = 0;
a->samples_rate = 0;
return azure_speech_feed_tts(a, text, flags);
}

1
mod_azure_tts/mod_azure_tts.h
View File

@@ -25,7 +25,6 @@ typedef struct azure_data {
int reads;
int cache_audio;
int flushed;
uint32_t samples_rate;
void *startTime;

23
mod_elevenlabs_tts/elevenlabs_glue.cpp
View File

@@ -841,7 +841,6 @@ extern "C" {
CURL* easy = createEasyHandle();
el->conn = (void *) conn ;
el->sample_rate = 0;
conn->elevenlabs = el;
conn->easy = easy;
conn->global = &global;
@@ -852,20 +851,12 @@ extern "C" {
el->circularBuffer = (void *) new CircularBuffer_t(8192);
if (el->session_id) {
if (el->rate != 8000 /*Hz*/) {
int err;
switch_codec_implementation_t read_impl;
switch_core_session_t *psession = switch_core_session_locate(el->session_id);
switch_core_session_get_read_impl(psession, &read_impl);
uint32_t samples_per_second = !strcasecmp(read_impl.iananame, "g722") ? read_impl.actual_samples_per_second : read_impl.samples_per_second;
el->sample_rate = samples_per_second;
// elevenlabs output is PCMU 8000
if (samples_per_second != 8000 /*Hz*/) {
el->resampler = speex_resampler_init(1, 8000, samples_per_second, SWITCH_RESAMPLE_QUALITY, &err);
if (0 != err) {
switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_ERROR, "Error initializing resampler: %s.\n", speex_resampler_strerror(err));
return SWITCH_STATUS_FALSE;
}
el->resampler = speex_resampler_init(1, 8000, el->rate, SWITCH_RESAMPLE_QUALITY, &err);
if (0 != err) {
switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_ERROR, "Error initializing resampler: %s.\n", speex_resampler_strerror(err));
return SWITCH_STATUS_FALSE;
}
}
@@ -939,9 +930,7 @@ extern "C" {
switch_mutex_unlock(el->mutex);
return SWITCH_STATUS_SUCCESS;
}
size_t size = el->sample_rate ?
std::min((*datalen/(2 * el->sample_rate / 8000)), cBuffer->size()) :
std::min((*datalen/2), cBuffer->size());
size_t size = std::min((*datalen/(2 * el->rate / 8000)), cBuffer->size());
pcm_data.insert(pcm_data.end(), cBuffer->begin(), cBuffer->begin() + size);
cBuffer->erase(cBuffer->begin(), cBuffer->begin() + size);
switch_mutex_unlock(el->mutex);

1
mod_elevenlabs_tts/mod_elevenlabs_tts.h
View File

@@ -31,7 +31,6 @@ struct elevenlabs_data {
char *cache_filename;
int rate;
uint32_t sample_rate;
void *conn;
FILE *file;

14
mod_playht_tts/playht_glue.cpp
View File

@@ -838,17 +838,9 @@ extern "C" {
p->circularBuffer = (void *) new CircularBuffer_t(8192);
if (p->session_id) {
int err;
switch_codec_implementation_t read_impl;
switch_core_session_t *psession = switch_core_session_locate(p->session_id);
switch_core_session_get_read_impl(psession, &read_impl);
switch_core_session_rwunlock(psession);
uint32_t samples_per_second = !strcasecmp(read_impl.iananame, "g722") ? read_impl.actual_samples_per_second : read_impl.samples_per_second;
if (mpg123_param(mh, MPG123_FORCE_RATE, samples_per_second /*Hz*/, 0) != MPG123_OK) {
switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_ERROR, "Error mpg123_param!\n");
return SWITCH_STATUS_FALSE;
}
if (mpg123_param(mh, MPG123_FORCE_RATE, p->rate /*Hz*/, 0) != MPG123_OK) {
switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_ERROR, "Error mpg123_param!\n");
return SWITCH_STATUS_FALSE;
}
std::ostringstream api_key_stream;