Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Linux] Process TTS callback on the main thread to avoid speech-dispatcher deadlock. #73671

Merged
merged 1 commit into from
Feb 21, 2023
Merged

[Linux] Process TTS callback on the main thread to avoid speech-dispatcher deadlock. #73671

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
135 changes: 74 additions & 61 deletions platform/linuxbsd/tts_linux.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -35,6 +35,11 @@

TTS_Linux *TTS_Linux::singleton = nullptr;

void TTS_Linux::_bind_methods() {
ClassDB::bind_method(D_METHOD("_speech_event", "msg_id", "client_id", "type"), &TTS_Linux::_speech_event);
ClassDB::bind_method(D_METHOD("_speech_index_mark", "msg_id", "client_id", "type", "index_mark"), &TTS_Linux::_speech_index_mark);
}

void TTS_Linux::speech_init_thread_func(void *p_userdata) {
TTS_Linux *tts = (TTS_Linux *)p_userdata;
if (tts) {
Expand Down Expand Up @@ -76,76 +81,84 @@ void TTS_Linux::speech_init_thread_func(void *p_userdata) {

void TTS_Linux::speech_event_index_mark(size_t p_msg_id, size_t p_client_id, SPDNotificationType p_type, char *p_index_mark) {
TTS_Linux *tts = TTS_Linux::get_singleton();
if (tts && tts->ids.has(p_msg_id)) {
MutexLock thread_safe_method(tts->_thread_safe_);
// Get word offset from the index mark injected to the text stream.
String mark = String::utf8(p_index_mark);
DisplayServer::get_singleton()->tts_post_utterance_event(DisplayServer::TTS_UTTERANCE_BOUNDARY, tts->ids[p_msg_id], mark.to_int());
if (tts) {
tts->call_deferred(SNAME("_speech_index_mark"), p_msg_id, p_client_id, (int)p_type, String::utf8(p_index_mark));
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Would this work with callable_mp(...).call_deferred() instead of binding the methods?

Registering this with GDCLASS has made it show up in the GDExtension API: godotengine/godot-cpp@c1ff169#diff-fc336699721277f071b492b7c4006667e3073bb75cef6ddf75721ba8a61d12daR217376-R217382

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes, this will work as well - #73764

}
}

void TTS_Linux::_speech_index_mark(size_t p_msg_id, size_t p_client_id, int p_type, const String &p_index_mark) {
_THREAD_SAFE_METHOD_

if (ids.has(p_msg_id)) {
DisplayServer::get_singleton()->tts_post_utterance_event(DisplayServer::TTS_UTTERANCE_BOUNDARY, ids[p_msg_id], p_index_mark.to_int());
}
}

void TTS_Linux::speech_event_callback(size_t p_msg_id, size_t p_client_id, SPDNotificationType p_type) {
TTS_Linux *tts = TTS_Linux::get_singleton();
if (tts) {
MutexLock thread_safe_method(tts->_thread_safe_);
List<DisplayServer::TTSUtterance> &queue = tts->queue;
if (!tts->paused && tts->ids.has(p_msg_id)) {
if (p_type == SPD_EVENT_END) {
DisplayServer::get_singleton()->tts_post_utterance_event(DisplayServer::TTS_UTTERANCE_ENDED, tts->ids[p_msg_id]);
tts->ids.erase(p_msg_id);
tts->last_msg_id = -1;
tts->speaking = false;
} else if (p_type == SPD_EVENT_CANCEL) {
DisplayServer::get_singleton()->tts_post_utterance_event(DisplayServer::TTS_UTTERANCE_CANCELED, tts->ids[p_msg_id]);
tts->ids.erase(p_msg_id);
tts->last_msg_id = -1;
tts->speaking = false;
}
tts->call_deferred(SNAME("_speech_event"), p_msg_id, p_client_id, (int)p_type);
}
}

void TTS_Linux::_speech_event(size_t p_msg_id, size_t p_client_id, int p_type) {
_THREAD_SAFE_METHOD_

if (!paused && ids.has(p_msg_id)) {
if ((SPDNotificationType)p_type == SPD_EVENT_END) {
DisplayServer::get_singleton()->tts_post_utterance_event(DisplayServer::TTS_UTTERANCE_ENDED, ids[p_msg_id]);
ids.erase(p_msg_id);
last_msg_id = -1;
speaking = false;
} else if ((SPDNotificationType)p_type == SPD_EVENT_CANCEL) {
DisplayServer::get_singleton()->tts_post_utterance_event(DisplayServer::TTS_UTTERANCE_CANCELED, ids[p_msg_id]);
ids.erase(p_msg_id);
last_msg_id = -1;
speaking = false;
}
if (!tts->speaking && queue.size() > 0) {
DisplayServer::TTSUtterance &message = queue.front()->get();

// Inject index mark after each word.
String text;
String language;
SPDVoice **voices = spd_list_synthesis_voices(tts->synth);
if (voices != nullptr) {
SPDVoice **voices_ptr = voices;
while (*voices_ptr != nullptr) {
if (String::utf8((*voices_ptr)->name) == message.voice) {
language = String::utf8((*voices_ptr)->language);
break;
}
voices_ptr++;
}
free_spd_voices(voices);
}
PackedInt32Array breaks = TS->string_get_word_breaks(message.text, language);
for (int i = 0; i < breaks.size(); i += 2) {
const int start = breaks[i];
const int end = breaks[i + 1];
text += message.text.substr(start, end - start + 1);
text += "<mark name=\"" + String::num_int64(end, 10) + "\"/>";
}
}
if (!speaking && queue.size() > 0) {
DisplayServer::TTSUtterance &message = queue.front()->get();

spd_set_synthesis_voice(tts->synth, message.voice.utf8().get_data());
spd_set_volume(tts->synth, message.volume * 2 - 100);
spd_set_voice_pitch(tts->synth, (message.pitch - 1) * 100);
float rate = 0;
if (message.rate > 1.f) {
rate = log10(MIN(message.rate, 2.5f)) / log10(2.5f) * 100;
} else if (message.rate < 1.f) {
rate = log10(MAX(message.rate, 0.5f)) / log10(0.5f) * -100;
// Inject index mark after each word.
String text;
String language;
SPDVoice **voices = spd_list_synthesis_voices(synth);
if (voices != nullptr) {
SPDVoice **voices_ptr = voices;
while (*voices_ptr != nullptr) {
if (String::utf8((*voices_ptr)->name) == message.voice) {
language = String::utf8((*voices_ptr)->language);
break;
}
voices_ptr++;
}
spd_set_voice_rate(tts->synth, rate);
spd_set_data_mode(tts->synth, SPD_DATA_SSML);
tts->last_msg_id = spd_say(tts->synth, SPD_TEXT, text.utf8().get_data());
tts->ids[tts->last_msg_id] = message.id;
DisplayServer::get_singleton()->tts_post_utterance_event(DisplayServer::TTS_UTTERANCE_STARTED, message.id);

queue.pop_front();
tts->speaking = true;
free_spd_voices(voices);
}
PackedInt32Array breaks = TS->string_get_word_breaks(message.text, language);
for (int i = 0; i < breaks.size(); i += 2) {
const int start = breaks[i];
const int end = breaks[i + 1];
text += message.text.substr(start, end - start + 1);
text += "<mark name=\"" + String::num_int64(end, 10) + "\"/>";
}
spd_set_synthesis_voice(synth, message.voice.utf8().get_data());
spd_set_volume(synth, message.volume * 2 - 100);
spd_set_voice_pitch(synth, (message.pitch - 1) * 100);
float rate = 0;
if (message.rate > 1.f) {
rate = log10(MIN(message.rate, 2.5f)) / log10(2.5f) * 100;
} else if (message.rate < 1.f) {
rate = log10(MAX(message.rate, 0.5f)) / log10(0.5f) * -100;
}
spd_set_voice_rate(synth, rate);
spd_set_data_mode(synth, SPD_DATA_SSML);
last_msg_id = spd_say(synth, SPD_TEXT, text.utf8().get_data());
ids[last_msg_id] = message.id;
DisplayServer::get_singleton()->tts_post_utterance_event(DisplayServer::TTS_UTTERANCE_STARTED, message.id);

queue.pop_front();
speaking = true;
}
}

Expand Down Expand Up @@ -204,7 +217,7 @@ void TTS_Linux::speak(const String &p_text, const String &p_voice, int p_volume,
if (is_paused()) {
resume();
} else {
speech_event_callback(0, 0, SPD_EVENT_BEGIN);
_speech_event(0, 0, (int)SPD_EVENT_BEGIN);
}
}

Expand Down
8 changes: 7 additions & 1 deletion platform/linuxbsd/tts_linux.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -45,7 +45,8 @@
#include <libspeechd.h>
#endif

class TTS_Linux {
class TTS_Linux : public Object {
GDCLASS(TTS_Linux, Object);
_THREAD_SAFE_CLASS_

List<DisplayServer::TTSUtterance> queue;
Expand All @@ -63,6 +64,11 @@ class TTS_Linux {

static TTS_Linux *singleton;

protected:
static void _bind_methods();
void _speech_event(size_t p_msg_id, size_t p_client_id, int p_type);
void _speech_index_mark(size_t p_msg_id, size_t p_client_id, int p_type, const String &p_index_mark);

public:
static TTS_Linux *get_singleton();

Expand Down