4 files changed, 34 insertions, 25 deletions

diff --git a/ext/bg/js/api.js b/ext/bg/js/api.js
index 93e43a7d..4e5d81db 100644
--- a/ext/bg/js/api.js
+++ b/ext/bg/js/api.js
@@ -21,10 +21,6 @@ function apiTemplateRender(template, data) {
     return _apiInvoke('templateRender', {data, template});
 }
 
-function apiAudioGetUrl(definition, source, optionsContext) {
-    return _apiInvoke('audioGetUrl', {definition, source, optionsContext});
-}
-
 function _apiInvoke(action, params={}) {
     const data = {action, params};
     return new Promise((resolve, reject) => {
diff --git a/ext/bg/js/audio.js b/ext/bg/js/audio.js
index 972e2b8b..c94121ae 100644
--- a/ext/bg/js/audio.js
+++ b/ext/bg/js/audio.js
@@ -16,7 +16,7 @@
  * along with this program.  If not, see <https://www.gnu.org/licenses/>.
  */
 
-/*global jpIsStringEntirelyKana, audioGetFromSources*/
+/*global jpIsStringEntirelyKana*/
 
 const audioUrlBuilders = new Map([
     ['jpod101', async (definition) => {
@@ -154,7 +154,7 @@ function audioBuildFilename(definition) {
     return null;
 }
 
-async function audioInject(definition, fields, sources, optionsContext) {
+async function audioInject(definition, fields, sources, optionsContext, audioSystem) {
     let usesAudio = false;
     for (const fieldValue of Object.values(fields)) {
         if (fieldValue.includes('{audio}')) {
@@ -171,12 +171,10 @@ async function audioInject(definition, fields, sources, optionsContext) {
         const expressions = definition.expressions;
         const audioSourceDefinition = Array.isArray(expressions) ? expressions[0] : definition;
 
-        const {url} = await audioGetFromSources(audioSourceDefinition, sources, optionsContext, true);
-        if (url !== null) {
-            const filename = audioBuildFilename(audioSourceDefinition);
-            if (filename !== null) {
-                definition.audio = {url, filename};
-            }
+        const {uri} = await audioSystem.getDefinitionAudio(audioSourceDefinition, sources, {tts: false, optionsContext});
+        const filename = audioBuildFilename(audioSourceDefinition);
+        if (filename !== null) {
+            definition.audio = {url: uri, filename};
         }
 
         return true;
diff --git a/ext/bg/js/backend.js b/ext/bg/js/backend.js
index 04bf240d..60a87916 100644
--- a/ext/bg/js/backend.js
+++ b/ext/bg/js/backend.js
@@ -23,7 +23,7 @@ requestText, requestJson, optionsLoad
 dictConfigured, dictTermsSort, dictEnabledSet, dictNoteFormat
 audioGetUrl, audioInject
 jpConvertReading, jpDistributeFuriganaInflected, jpKatakanaToHiragana
-Translator, AnkiConnect, AnkiNull, Mecab, BackendApiForwarder, JsonSchema, ClipboardMonitor*/
+AudioSystem, Translator, AnkiConnect, AnkiNull, Mecab, BackendApiForwarder, JsonSchema, ClipboardMonitor*/
 
 class Backend {
     constructor() {
@@ -34,6 +34,7 @@ class Backend {
         this.options = null;
         this.optionsSchema = null;
         this.defaultAnkiFieldTemplates = null;
+        this.audioSystem = new AudioSystem({getAudioUri: this._getAudioUri.bind(this)});
         this.optionsContext = {
             depth: 0,
             url: window.location.href
@@ -436,7 +437,8 @@ class Backend {
                 definition,
                 options.anki.terms.fields,
                 options.audio.sources,
-                optionsContext
+                optionsContext,
+                this.audioSystem
             );
         }
 
@@ -762,6 +764,16 @@ class Backend {
 
     // Utilities
 
+    async _getAudioUri(definition, source, details) {
+        let optionsContext = (typeof details === 'object' && details !== null ? details.optionsContext : null);
+        if (!(typeof optionsContext === 'object' && optionsContext !== null)) {
+            optionsContext = this.optionsContext;
+        }
+
+        const options = this.getOptions(optionsContext);
+        return await audioGetUrl(definition, source, options);
+    }
+
     async _injectScreenshot(definition, fields, screenshot) {
         let usesScreenshot = false;
         for (const fieldValue of Object.values(fields)) {
diff --git a/ext/bg/js/settings/audio.js b/ext/bg/js/settings/audio.js
index 6d183a43..6f581d9b 100644
--- a/ext/bg/js/settings/audio.js
+++ b/ext/bg/js/settings/audio.js
@@ -16,12 +16,20 @@
  * along with this program.  If not, see <https://www.gnu.org/licenses/>.
  */
 
-/*global getOptionsContext, getOptionsMutable, settingsSaveOptions
-AudioSourceUI, audioGetTextToSpeechVoice*/
+/*global getOptionsContext, getOptionsMutable, settingsSaveOptions, apiAudioGetUrl
+AudioSystem, AudioSourceUI*/
 
 let audioSourceUI = null;
+let audioSystem = null;
 
 async function audioSettingsInitialize() {
+    audioSystem = new AudioSystem({
+        getAudioUri: async (definition, source) => {
+            const optionsContext = getOptionsContext();
+            return await apiAudioGetUrl(definition, source, optionsContext);
+        }
+    });
+
     const optionsContext = getOptionsContext();
     const options = await getOptionsMutable(optionsContext);
     audioSourceUI = new AudioSourceUI.Container(
@@ -100,16 +108,11 @@ function textToSpeechVoiceCompare(a, b) {
 function textToSpeechTest() {
     try {
         const text = document.querySelector('#text-to-speech-voice-test').dataset.speechText || '';
-        const voiceURI = document.querySelector('#text-to-speech-voice').value;
-        const voice = audioGetTextToSpeechVoice(voiceURI);
-        if (voice === null) { return; }
-
-        const utterance = new SpeechSynthesisUtterance(text);
-        utterance.lang = 'ja-JP';
-        utterance.voice = voice;
-        utterance.volume = 1.0;
+        const voiceUri = document.querySelector('#text-to-speech-voice').value;
 
-        speechSynthesis.speak(utterance);
+        const audio = audioSystem.createTextToSpeechAudio({text, voiceUri});
+        audio.volume = 1.0;
+        audio.play();
     } catch (e) {
         // NOP
     }