0.0.42

support whisper streaming

lib/synth-audio.js

@@ -145,7 +145,7 @@ async function synthAudio(client, logger, stats, { account_sid,
       ['microsoft', 'azure'].includes(vendor)
     ) ||
     (
-      process.env.JAMBONES_ELEVENLABS_STREAMING &&
+      !process.env.JAMBONES_DISABLE_TTS_STREAMING &&
       vendor === 'elevenlabs'
     )
   ) {
@@ -210,6 +210,10 @@ async function synthAudio(client, logger, stats, { account_sid,
         }
         break;
       case 'whisper':
+        audioBuffer = await synthWhisper(logger, {credentials, stats, voice, text, renderForCaching});
+        if (typeof audioBuffer === 'object' && audioBuffer.filePath) {
+          return audioBuffer;
+        }
         audioBuffer = await synthWhisper(logger, {credentials, stats, voice, text});
         break;
       case 'deepgram':
@@ -611,8 +615,8 @@ const synthElevenlabs = async(logger, {credentials, options, stats, language, vo
   const {api_key, model_id, options: credOpts} = credentials;
   const opts = !!options && Object.keys(options).length !== 0 ? options : JSON.parse(credOpts || '{}');
 
-  /* if the env is set to stream then bag out, unless we are specifically rendering to generate a cache file */
-  if (process.env.JAMBONES_ELEVENLABS_STREAMING && !renderForCaching) {
+  /* default to using the streaming interface, unless disabled by env var OR we want just a cache file */
+  if (!process.env.JAMBONES_DISABLE_TTS_STREAMING && !renderForCaching) {
     let params = '';
     params += `{api_key=${api_key}`;
     params += `,model_id=${model_id}`;
@@ -625,7 +629,7 @@ const synthElevenlabs = async(logger, {credentials, options, stats, language, vo
     params += '}';
 
     return {
-      filePath: `say:${params}${text.replace(/\n/g, ' ')}`,
+      filePath: `say:${params}${text.replace(/\n/g, ' ').replace(/\r/g, ' ')}`,
       servedFromCache: false,
       rtt: 0
     };
@@ -656,8 +660,24 @@ const synthElevenlabs = async(logger, {credentials, options, stats, language, vo
   }
 };
 
-const synthWhisper = async(logger, {credentials, stats, voice, text}) => {
-  const {api_key, model_id, baseURL, timeout} = credentials;
+const synthWhisper = async(logger, {credentials, stats, voice, text, renderForCaching}) => {
+  const {api_key, model_id, baseURL, timeout, speed} = credentials;
+  /* if the env is set to stream then bag out, unless we are specifically rendering to generate a cache file */
+  if (!process.env.JAMBONES_DISABLE_TTS_STREAMING && !renderForCaching) {
+    let params = '';
+    params += `{api_key=${api_key}`;
+    params += `,model_id=${model_id}`;
+    params += `,voice=${voice}`;
+    params += ',write_cache_file=1';
+    if (speed) params += `,speed=${speed}`;
+    params += '}';
+
+    return {
+      filePath: `say:${params}${text.replace(/\n/g, ' ')}`,
+      servedFromCache: false,
+      rtt: 0
+    };
+  }
   try {
     const openai = new OpenAI.OpenAI({
       apiKey: api_key,

package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "@jambonz/speech-utils",
-  "version": "0.0.38",
+  "version": "0.0.42",
   "lockfileVersion": 2,
   "requires": true,
   "packages": {
     "": {
       "name": "@jambonz/speech-utils",
-      "version": "0.0.38",
+      "version": "0.0.42",
       "license": "MIT",
       "dependencies": {
         "@aws-sdk/client-polly": "^3.496.0",

package.json

@@ -1,6 +1,6 @@
 {
   "name": "@jambonz/speech-utils",
-  "version": "0.0.38",
+  "version": "0.0.42",
   "description": "TTS-related speech utilities for jambonz",
   "main": "index.js",
   "author": "Dave Horton",