Implement text-to-speech support on Android, iOS, HTML5, Linux, macOS and Windows.

Implement TextServer word break method.

6 files changed, 407 insertions, 0 deletions

diff --git a/platform/windows/SCsub b/platform/windows/SCsub
index 76234c3065..7e412b140f 100644
--- a/platform/windows/SCsub
+++ b/platform/windows/SCsub
@@ -13,6 +13,7 @@ common_win = [
     "display_server_windows.cpp",
     "key_mapping_windows.cpp",
     "joypad_windows.cpp",
+    "tts_windows.cpp",
     "windows_terminal_logger.cpp",
     "vulkan_context_win.cpp",
     "gl_manager_windows.cpp",
diff --git a/platform/windows/detect.py b/platform/windows/detect.py
index 249a0d2e79..0b18fb74fb 100644
--- a/platform/windows/detect.py
+++ b/platform/windows/detect.py
@@ -252,6 +252,7 @@ def configure_msvc(env, manual_msvc_config):
         "kernel32",
         "ole32",
         "oleaut32",
+        "sapi",
         "user32",
         "gdi32",
         "IPHLPAPI",
@@ -426,6 +427,7 @@ def configure_mingw(env):
             "ws2_32",
             "kernel32",
             "oleaut32",
+            "sapi",
             "dinput8",
             "dxguid",
             "ksuser",
diff --git a/platform/windows/display_server_windows.cpp b/platform/windows/display_server_windows.cpp
index 31bad0f053..0412eb2d9c 100644
--- a/platform/windows/display_server_windows.cpp
+++ b/platform/windows/display_server_windows.cpp
@@ -84,6 +84,7 @@ bool DisplayServerWindows::has_feature(Feature p_feature) const {
 		case FEATURE_NATIVE_ICON:
 		case FEATURE_SWAP_BUFFERS:
 		case FEATURE_KEEP_SCREEN_ON:
+		case FEATURE_TEXT_TO_SPEECH:
 			return true;
 		default:
 			return false;
@@ -133,6 +134,41 @@ void DisplayServerWindows::_set_mouse_mode_impl(MouseMode p_mode) {
 	}
 }
 
+bool DisplayServerWindows::tts_is_speaking() const {
+	ERR_FAIL_COND_V(!tts, false);
+	return tts->is_speaking();
+}
+
+bool DisplayServerWindows::tts_is_paused() const {
+	ERR_FAIL_COND_V(!tts, false);
+	return tts->is_paused();
+}
+
+Array DisplayServerWindows::tts_get_voices() const {
+	ERR_FAIL_COND_V(!tts, Array());
+	return tts->get_voices();
+}
+
+void DisplayServerWindows::tts_speak(const String &p_text, const String &p_voice, int p_volume, float p_pitch, float p_rate, int p_utterance_id, bool p_interrupt) {
+	ERR_FAIL_COND(!tts);
+	tts->speak(p_text, p_voice, p_volume, p_pitch, p_rate, p_utterance_id, p_interrupt);
+}
+
+void DisplayServerWindows::tts_pause() {
+	ERR_FAIL_COND(!tts);
+	tts->pause();
+}
+
+void DisplayServerWindows::tts_resume() {
+	ERR_FAIL_COND(!tts);
+	tts->resume();
+}
+
+void DisplayServerWindows::tts_stop() {
+	ERR_FAIL_COND(!tts);
+	tts->stop();
+}
+
 void DisplayServerWindows::mouse_set_mode(MouseMode p_mode) {
 	_THREAD_SAFE_METHOD_
 
@@ -3497,6 +3533,9 @@ DisplayServerWindows::DisplayServerWindows(const String &p_rendering_driver, Win
 
 	rendering_driver = p_rendering_driver;
 
+	// Init TTS
+	tts = memnew(TTS_Windows);
+
 	// Note: Wacom WinTab driver API for pen input, for devices incompatible with Windows Ink.
 	HMODULE wintab_lib = LoadLibraryW(L"wintab32.dll");
 	if (wintab_lib) {
@@ -3739,4 +3778,8 @@ DisplayServerWindows::~DisplayServerWindows() {
 		gl_manager = nullptr;
 	}
 #endif
+	if (tts) {
+		memdelete(tts);
+	}
+	CoUninitialize();
 }
diff --git a/platform/windows/display_server_windows.h b/platform/windows/display_server_windows.h
index fcf4b5a728..80faf71bd4 100644
--- a/platform/windows/display_server_windows.h
+++ b/platform/windows/display_server_windows.h
@@ -46,6 +46,7 @@
 #include "servers/rendering/renderer_compositor.h"
 #include "servers/rendering/renderer_rd/renderer_compositor_rd.h"
 #include "servers/rendering_server.h"
+#include "tts_windows.h"
 
 #ifdef XAUDIO2_ENABLED
 #include "drivers/xaudio2/audio_driver_xaudio2.h"
@@ -320,6 +321,8 @@ class DisplayServerWindows : public DisplayServer {
 	String rendering_driver;
 	bool app_focused = false;
 
+	TTS_Windows *tts = nullptr;
+
 	struct WindowData {
 		HWND hWnd;
 		//layered window
@@ -454,6 +457,15 @@ public:
 	virtual bool has_feature(Feature p_feature) const override;
 	virtual String get_name() const override;
 
+	virtual bool tts_is_speaking() const override;
+	virtual bool tts_is_paused() const override;
+	virtual Array tts_get_voices() const override;
+
+	virtual void tts_speak(const String &p_text, const String &p_voice, int p_volume = 50, float p_pitch = 1.f, float p_rate = 1.f, int p_utterance_id = 0, bool p_interrupt = false) override;
+	virtual void tts_pause() override;
+	virtual void tts_resume() override;
+	virtual void tts_stop() override;
+
 	virtual void mouse_set_mode(MouseMode p_mode) override;
 	virtual MouseMode mouse_get_mode() const override;
 
diff --git a/platform/windows/tts_windows.cpp b/platform/windows/tts_windows.cpp
new file mode 100644
index 0000000000..05249934ba
--- /dev/null
+++ b/platform/windows/tts_windows.cpp
@@ -0,0 +1,269 @@
+/*************************************************************************/
+/*  tts_windows.cpp                                                      */
+/*************************************************************************/
+/*                       This file is part of:                           */
+/*                           GODOT ENGINE                                */
+/*                      https://godotengine.org                          */
+/*************************************************************************/
+/* Copyright (c) 2007-2022 Juan Linietsky, Ariel Manzur.                 */
+/* Copyright (c) 2014-2022 Godot Engine contributors (cf. AUTHORS.md).   */
+/*                                                                       */
+/* Permission is hereby granted, free of charge, to any person obtaining */
+/* a copy of this software and associated documentation files (the       */
+/* "Software"), to deal in the Software without restriction, including   */
+/* without limitation the rights to use, copy, modify, merge, publish,   */
+/* distribute, sublicense, and/or sell copies of the Software, and to    */
+/* permit persons to whom the Software is furnished to do so, subject to */
+/* the following conditions:                                             */
+/*                                                                       */
+/* The above copyright notice and this permission notice shall be        */
+/* included in all copies or substantial portions of the Software.       */
+/*                                                                       */
+/* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,       */
+/* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF    */
+/* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.*/
+/* IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY  */
+/* CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,  */
+/* TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE     */
+/* SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.                */
+/*************************************************************************/
+
+#include "tts_windows.h"
+
+TTS_Windows *TTS_Windows::singleton = nullptr;
+
+void __stdcall TTS_Windows::speech_event_callback(WPARAM wParam, LPARAM lParam) {
+	TTS_Windows *tts = TTS_Windows::get_singleton();
+	SPEVENT event;
+	while (tts->synth->GetEvents(1, &event, NULL) == S_OK) {
+		if (tts->ids.has(event.ulStreamNum)) {
+			if (event.eEventId == SPEI_START_INPUT_STREAM) {
+				DisplayServer::get_singleton()->tts_post_utterance_event(DisplayServer::TTS_UTTERANCE_STARTED, tts->ids[event.ulStreamNum].id);
+			} else if (event.eEventId == SPEI_END_INPUT_STREAM) {
+				DisplayServer::get_singleton()->tts_post_utterance_event(DisplayServer::TTS_UTTERANCE_ENDED, tts->ids[event.ulStreamNum].id);
+				tts->ids.erase(event.ulStreamNum);
+				tts->_update_tts();
+			} else if (event.eEventId == SPEI_WORD_BOUNDARY) {
+				const Char16String &string = tts->ids[event.ulStreamNum].string;
+				int pos = 0;
+				for (int i = 0; i < MIN(event.lParam, string.length()); i++) {
+					char16_t c = string[i];
+					if ((c & 0xfffffc00) == 0xd800) {
+						i++;
+					}
+					pos++;
+				}
+				DisplayServer::get_singleton()->tts_post_utterance_event(DisplayServer::TTS_UTTERANCE_BOUNDARY, tts->ids[event.ulStreamNum].id, pos - tts->ids[event.ulStreamNum].offset);
+			}
+		}
+	}
+}
+
+void TTS_Windows::_update_tts() {
+	if (!is_speaking() && !paused && queue.size() > 0) {
+		DisplayServer::TTSUtterance &message = queue.front()->get();
+
+		String text;
+		DWORD flags = SPF_ASYNC | SPF_PURGEBEFORESPEAK | SPF_IS_XML;
+		String pitch_tag = String("<pitch absmiddle=\"") + String::num_int64(message.pitch * 10 - 10, 10) + String("\">");
+		text = pitch_tag + message.text + String("</pitch>");
+
+		IEnumSpObjectTokens *cpEnum;
+		ISpObjectToken *cpVoiceToken;
+		ULONG ulCount = 0;
+		ULONG stream_number = 0;
+		ISpObjectTokenCategory *cpCategory;
+		HRESULT hr = CoCreateInstance(CLSID_SpObjectTokenCategory, nullptr, CLSCTX_INPROC_SERVER, IID_ISpObjectTokenCategory, (void **)&cpCategory);
+		if (SUCCEEDED(hr)) {
+			hr = cpCategory->SetId(SPCAT_VOICES, false);
+			if (SUCCEEDED(hr)) {
+				hr = cpCategory->EnumTokens(nullptr, nullptr, &cpEnum);
+				if (SUCCEEDED(hr)) {
+					hr = cpEnum->GetCount(&ulCount);
+					while (SUCCEEDED(hr) && ulCount--) {
+						wchar_t *w_id = 0L;
+						hr = cpEnum->Next(1, &cpVoiceToken, nullptr);
+						cpVoiceToken->GetId(&w_id);
+						if (String::utf16((const char16_t *)w_id) == message.voice) {
+							synth->SetVoice(cpVoiceToken);
+							cpVoiceToken->Release();
+							break;
+						}
+						cpVoiceToken->Release();
+					}
+					cpEnum->Release();
+				}
+			}
+			cpCategory->Release();
+		}
+
+		UTData ut;
+		ut.string = text.utf16();
+		ut.offset = pitch_tag.length(); // Substract injected <pitch> tag offset.
+		ut.id = message.id;
+
+		synth->SetVolume(message.volume);
+		synth->SetRate(10.f * log10(message.rate) / log10(3.f));
+		synth->Speak((LPCWSTR)ut.string.get_data(), flags, &stream_number);
+
+		ids[stream_number] = ut;
+
+		queue.pop_front();
+	}
+}
+
+bool TTS_Windows::is_speaking() const {
+	ERR_FAIL_COND_V(!synth, false);
+
+	SPVOICESTATUS status;
+	synth->GetStatus(&status, nullptr);
+	return (status.dwRunningState == SPRS_IS_SPEAKING);
+}
+
+bool TTS_Windows::is_paused() const {
+	ERR_FAIL_COND_V(!synth, false);
+	return paused;
+}
+
+Array TTS_Windows::get_voices() const {
+	Array list;
+	IEnumSpObjectTokens *cpEnum;
+	ISpObjectToken *cpVoiceToken;
+	ISpDataKey *cpDataKeyAttribs;
+	ULONG ulCount = 0;
+	ISpObjectTokenCategory *cpCategory;
+	HRESULT hr = CoCreateInstance(CLSID_SpObjectTokenCategory, nullptr, CLSCTX_INPROC_SERVER, IID_ISpObjectTokenCategory, (void **)&cpCategory);
+	if (SUCCEEDED(hr)) {
+		hr = cpCategory->SetId(SPCAT_VOICES, false);
+		if (SUCCEEDED(hr)) {
+			hr = cpCategory->EnumTokens(nullptr, nullptr, &cpEnum);
+			if (SUCCEEDED(hr)) {
+				hr = cpEnum->GetCount(&ulCount);
+				while (SUCCEEDED(hr) && ulCount--) {
+					hr = cpEnum->Next(1, &cpVoiceToken, nullptr);
+					HRESULT hr_attr = cpVoiceToken->OpenKey(SPTOKENKEY_ATTRIBUTES, &cpDataKeyAttribs);
+					if (SUCCEEDED(hr_attr)) {
+						wchar_t *w_id = nullptr;
+						wchar_t *w_lang = nullptr;
+						wchar_t *w_name = nullptr;
+						cpVoiceToken->GetId(&w_id);
+						cpDataKeyAttribs->GetStringValue(L"Language", &w_lang);
+						cpDataKeyAttribs->GetStringValue(nullptr, &w_name);
+						LCID locale = wcstol(w_lang, nullptr, 16);
+
+						int locale_chars = GetLocaleInfoW(locale, LOCALE_SISO639LANGNAME, nullptr, 0);
+						int region_chars = GetLocaleInfoW(locale, LOCALE_SISO3166CTRYNAME, nullptr, 0);
+						wchar_t *w_lang_code = new wchar_t[locale_chars];
+						wchar_t *w_reg_code = new wchar_t[region_chars];
+						GetLocaleInfoW(locale, LOCALE_SISO639LANGNAME, w_lang_code, locale_chars);
+						GetLocaleInfoW(locale, LOCALE_SISO3166CTRYNAME, w_reg_code, region_chars);
+
+						Dictionary voice_d;
+						voice_d["id"] = String::utf16((const char16_t *)w_id);
+						if (w_name) {
+							voice_d["name"] = String::utf16((const char16_t *)w_name);
+						} else {
+							voice_d["name"] = voice_d["id"].operator String().replace("HKEY_LOCAL_MACHINE\\SOFTWARE\\Microsoft\\Speech\\Voices\\Tokens\\", "");
+						}
+						voice_d["language"] = String::utf16((const char16_t *)w_lang_code) + "_" + String::utf16((const char16_t *)w_reg_code);
+						list.push_back(voice_d);
+
+						delete[] w_lang_code;
+						delete[] w_reg_code;
+
+						cpDataKeyAttribs->Release();
+					}
+					cpVoiceToken->Release();
+				}
+				cpEnum->Release();
+			}
+		}
+		cpCategory->Release();
+	}
+	return list;
+}
+
+void TTS_Windows::speak(const String &p_text, const String &p_voice, int p_volume, float p_pitch, float p_rate, int p_utterance_id, bool p_interrupt) {
+	ERR_FAIL_COND(!synth);
+	if (p_interrupt) {
+		stop();
+	}
+
+	if (p_text.is_empty()) {
+		DisplayServer::get_singleton()->tts_post_utterance_event(DisplayServer::TTS_UTTERANCE_CANCELED, p_utterance_id);
+		return;
+	}
+
+	DisplayServer::TTSUtterance message;
+	message.text = p_text;
+	message.voice = p_voice;
+	message.volume = CLAMP(p_volume, 0, 100);
+	message.pitch = CLAMP(p_pitch, 0.f, 2.f);
+	message.rate = CLAMP(p_rate, 0.1f, 10.f);
+	message.id = p_utterance_id;
+	queue.push_back(message);
+
+	if (is_paused()) {
+		resume();
+	} else {
+		_update_tts();
+	}
+}
+
+void TTS_Windows::pause() {
+	ERR_FAIL_COND(!synth);
+	if (!paused) {
+		if (synth->Pause() == S_OK) {
+			paused = true;
+		}
+	}
+}
+
+void TTS_Windows::resume() {
+	ERR_FAIL_COND(!synth);
+	synth->Resume();
+	paused = false;
+}
+
+void TTS_Windows::stop() {
+	ERR_FAIL_COND(!synth);
+
+	SPVOICESTATUS status;
+	synth->GetStatus(&status, nullptr);
+	if (ids.has(status.ulCurrentStream)) {
+		DisplayServer::get_singleton()->tts_post_utterance_event(DisplayServer::TTS_UTTERANCE_CANCELED, ids[status.ulCurrentStream].id);
+		ids.erase(status.ulCurrentStream);
+	}
+	for (DisplayServer::TTSUtterance &message : queue) {
+		DisplayServer::get_singleton()->tts_post_utterance_event(DisplayServer::TTS_UTTERANCE_CANCELED, message.id);
+	}
+	queue.clear();
+	synth->Speak(nullptr, SPF_PURGEBEFORESPEAK, nullptr);
+	synth->Resume();
+	paused = false;
+}
+
+TTS_Windows *TTS_Windows::get_singleton() {
+	return singleton;
+}
+
+TTS_Windows::TTS_Windows() {
+	singleton = this;
+	CoInitialize(nullptr);
+
+	if (SUCCEEDED(CoCreateInstance(CLSID_SpVoice, nullptr, CLSCTX_ALL, IID_ISpVoice, (void **)&synth))) {
+		ULONGLONG event_mask = SPFEI(SPEI_END_INPUT_STREAM) | SPFEI(SPEI_START_INPUT_STREAM) | SPFEI(SPEI_WORD_BOUNDARY);
+		synth->SetInterest(event_mask, event_mask);
+		synth->SetNotifyCallbackFunction(&speech_event_callback, (WPARAM)(this), 0);
+		print_verbose("Text-to-Speech: SAPI initialized.");
+	} else {
+		print_verbose("Text-to-Speech: Cannot initialize ISpVoice!");
+	}
+}
+
+TTS_Windows::~TTS_Windows() {
+	if (synth) {
+		synth->Release();
+	}
+	singleton = nullptr;
+}
diff --git a/platform/windows/tts_windows.h b/platform/windows/tts_windows.h
new file mode 100644
index 0000000000..5da404baf9
--- /dev/null
+++ b/platform/windows/tts_windows.h
@@ -0,0 +1,80 @@
+/*************************************************************************/
+/*  tts_windows.h                                                        */
+/*************************************************************************/
+/*                       This file is part of:                           */
+/*                           GODOT ENGINE                                */
+/*                      https://godotengine.org                          */
+/*************************************************************************/
+/* Copyright (c) 2007-2022 Juan Linietsky, Ariel Manzur.                 */
+/* Copyright (c) 2014-2022 Godot Engine contributors (cf. AUTHORS.md).   */
+/*                                                                       */
+/* Permission is hereby granted, free of charge, to any person obtaining */
+/* a copy of this software and associated documentation files (the       */
+/* "Software"), to deal in the Software without restriction, including   */
+/* without limitation the rights to use, copy, modify, merge, publish,   */
+/* distribute, sublicense, and/or sell copies of the Software, and to    */
+/* permit persons to whom the Software is furnished to do so, subject to */
+/* the following conditions:                                             */
+/*                                                                       */
+/* The above copyright notice and this permission notice shall be        */
+/* included in all copies or substantial portions of the Software.       */
+/*                                                                       */
+/* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,       */
+/* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF    */
+/* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.*/
+/* IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY  */
+/* CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,  */
+/* TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE     */
+/* SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.                */
+/*************************************************************************/
+
+#ifndef TTS_WINDOWS_H
+#define TTS_WINDOWS_H
+
+#include "core/string/ustring.h"
+#include "core/templates/list.h"
+#include "core/templates/map.h"
+#include "core/variant/array.h"
+#include "servers/display_server.h"
+
+#include <objbase.h>
+#include <sapi.h>
+#include <wchar.h>
+#include <winnls.h>
+
+#define WIN32_LEAN_AND_MEAN
+#include <windows.h>
+
+class TTS_Windows {
+	List<DisplayServer::TTSUtterance> queue;
+	ISpVoice *synth = nullptr;
+	bool paused = false;
+	struct UTData {
+		Char16String string;
+		int offset;
+		int id;
+	};
+	Map<ULONG, UTData> ids;
+
+	static void __stdcall speech_event_callback(WPARAM wParam, LPARAM lParam);
+	void _update_tts();
+
+	static TTS_Windows *singleton;
+
+public:
+	static TTS_Windows *get_singleton();
+
+	bool is_speaking() const;
+	bool is_paused() const;
+	Array get_voices() const;
+
+	void speak(const String &p_text, const String &p_voice, int p_volume = 50, float p_pitch = 1.f, float p_rate = 1.f, int p_utterance_id = 0, bool p_interrupt = false);
+	void pause();
+	void resume();
+	void stop();
+
+	TTS_Windows();
+	~TTS_Windows();
+};
+
+#endif // TTS_WINDOWS_H