0.2.11

Feat/893 azure ssml

lib/config.js

@@ -2,6 +2,7 @@ const JAMBONES_TTS_TRIM_SILENCE = process.env.JAMBONES_TTS_TRIM_SILENCE;
 const JAMBONES_DISABLE_TTS_STREAMING = process.env.JAMBONES_DISABLE_TTS_STREAMING;
 const JAMBONES_DISABLE_AZURE_TTS_STREAMING = process.env.JAMBONES_DISABLE_AZURE_TTS_STREAMING;
 const JAMBONES_EAGERLY_PRE_CACHE_AUDIO = process.env.JAMBONES_EAGERLY_PRE_CACHE_AUDIO;
+const JAMBONES_AZURE_ENABLE_SSML = process.env.JAMBONES_AZURE_ENABLE_SSML;
 
 const JAMBONES_HTTP_PROXY_IP = process.env.JAMBONES_HTTP_PROXY_IP;
 const JAMBONES_HTTP_PROXY_PORT = process.env.JAMBONES_HTTP_PROXY_PORT;
@@ -21,5 +22,6 @@ module.exports = {
   JAMBONES_TTS_CACHE_DURATION_MINS,
   JAMBONES_EAGERLY_PRE_CACHE_AUDIO,
   TMP_FOLDER,
-  HTTP_TIMEOUT
+  HTTP_TIMEOUT,
+  JAMBONES_AZURE_ENABLE_SSML
 };

lib/synth-audio.js

@@ -47,6 +47,7 @@ const {
   JAMBONES_HTTP_PROXY_PORT,
   JAMBONES_TTS_CACHE_DURATION_MINS,
   JAMBONES_TTS_TRIM_SILENCE,
+  JAMBONES_AZURE_ENABLE_SSML
 } = require('./config');
 const EXPIRES = JAMBONES_TTS_CACHE_DURATION_MINS;
 const OpenAI = require('openai');
@@ -466,7 +467,6 @@ async function _synthOnPremMicrosoft(logger, {
 }) {
   const {use_custom_tts, custom_tts_endpoint_url, api_key} = credentials;
   let content = text;
-
   if (use_custom_tts && !content.startsWith('<speak')) {
     /**
      * Note: it seems that to use custom voice ssml is required with the voice attribute
@@ -479,9 +479,13 @@ async function _synthOnPremMicrosoft(logger, {
     /* microsoft enforces some properties and uses voice xml element so if the user did not supply do it for them */
     const words = content.slice(7, -8).trim().replace(/(\r\n|\n|\r)/gm, ' ');
     // eslint-disable-next-line max-len
-    content = `<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xmlns:mstts="https://www.w3.org/2001/mstts" xml:lang="${language}"><voice name="${voice}">${words}</voice></speak>`;
+    content = `<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xmlns:mstts="https://www.w3.org/2001/mstts" xml:lang="${language}"><lang xml:lang="${language}"><voice name="${voice}">${words}</voice></lang></speak>`;
     logger.info({content}, 'synthMicrosoft');
   }
+  else if (JAMBONES_AZURE_ENABLE_SSML) {
+    // eslint-disable-next-line max-len
+    content = `<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xmlns:mstts="https://www.w3.org/2001/mstts" xml:lang="${language}"><voice name="${voice}"><lang xml:lang="${language}">${text}</lang></voice></speak>`;
+  }
 
   try {
     const trimSilence = JAMBONES_TTS_TRIM_SILENCE;
@@ -518,19 +522,23 @@ const synthMicrosoft = async(logger, {
     let content = text;
     if (use_custom_tts && !content.startsWith('<speak')) {
       /**
-       * Note: it seems that to use custom voice ssml is required with the voice attribute
-       * Otherwise sending plain text we get "Voice does not match"
-       */
+     * Note: it seems that to use custom voice ssml is required with the voice attribute
+     * Otherwise sending plain text we get "Voice does not match"
+     */
       content = `<speak>${text}</speak>`;
     }
 
     if (content.startsWith('<speak>')) {
-      /* microsoft enforces some properties and uses voice xml element so if the user did not supply do it for them */
+    /* microsoft enforces some properties and uses voice xml element so if the user did not supply do it for them */
       const words = content.slice(7, -8).trim().replace(/(\r\n|\n|\r)/gm, ' ');
       // eslint-disable-next-line max-len
-      content = `<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xmlns:mstts="https://www.w3.org/2001/mstts" xml:lang="${language}"><voice name="${voice}">${words}</voice></speak>`;
+      content = `<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xmlns:mstts="https://www.w3.org/2001/mstts" xml:lang="${language}"><lang xml:lang="${language}"><voice name="${voice}">${words}</voice></lang></speak>`;
       logger.info({content}, 'synthMicrosoft');
     }
+    else if (JAMBONES_AZURE_ENABLE_SSML) {
+      // eslint-disable-next-line max-len
+      content = `<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xmlns:mstts="https://www.w3.org/2001/mstts" xml:lang="${language}"><voice name="${voice}"><lang xml:lang="${language}">${text}</lang></voice></speak>`;
+    }
     if (!JAMBONES_DISABLE_TTS_STREAMING && !JAMBONES_DISABLE_AZURE_TTS_STREAMING &&
       !renderForCaching && !disableTtsStreaming) {
       let params = '';
@@ -855,11 +863,10 @@ const synthElevenlabs = async(logger, {
 const synthPlayHT = async(client, logger, {
   credentials, options, stats, voice, language, text, renderForCaching, disableTtsStreaming
 }) => {
-  const {api_key, user_id, voice_engine, options: credOpts} = credentials;
+  const {api_key, user_id, voice_engine, playht_tts_uri, options: credOpts} = credentials;
   const opts = !!options && Object.keys(options).length !== 0 ? options : JSON.parse(credOpts || '{}');
 
-  let synthesizeUrl = 'https://api.play.ht/api/v2/tts/stream';
-
+  let synthesizeUrl = playht_tts_uri ? `${playht_tts_uri}/api/v2/tts/stream` : 'https://api.play.ht/api/v2/tts/stream';
   // If model is play3.0, the synthesizeUrl is got from authentication endpoint
   if (voice_engine === 'Play3.0') {
     try {

package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "@jambonz/speech-utils",
-  "version": "0.2.9",
+  "version": "0.2.11",
   "lockfileVersion": 2,
   "requires": true,
   "packages": {
     "": {
       "name": "@jambonz/speech-utils",
-      "version": "0.2.9",
+      "version": "0.2.11",
       "license": "MIT",
       "dependencies": {
         "23": "^0.0.0",

package.json

@@ -1,6 +1,6 @@
 {
   "name": "@jambonz/speech-utils",
-  "version": "0.2.9",
+  "version": "0.2.11",
   "description": "TTS-related speech utilities for jambonz",
   "main": "index.js",
   "author": "Dave Horton",