0.2.22

use synth key as playback id

lib/config.js

@@ -9,7 +9,7 @@ const JAMBONES_HTTP_PROXY_PORT = process.env.JAMBONES_HTTP_PROXY_PORT;
 const JAMBONES_TTS_CACHE_DURATION_MINS =
   (parseInt(process.env.JAMBONES_TTS_CACHE_DURATION_MINS) || 4 * 60) * 60; // cache tts for 4 hours
 
-const TMP_FOLDER = '/tmp';
+const TMP_FOLDER = process.env.JAMBONES_TMP_FOLDER || '/tmp';
 
 const HTTP_TIMEOUT = 5000;
 

lib/synth-audio.js

@@ -53,7 +53,6 @@ const EXPIRES = JAMBONES_TTS_CACHE_DURATION_MINS;
 const OpenAI = require('openai');
 const getAwsAuthToken = require('./get-aws-sts-token');
 
-
 const trimTrailingSilence = (buffer) => {
   assert.ok(buffer instanceof Buffer, 'trimTrailingSilence - argument is not a Buffer');
 
@@ -98,7 +97,7 @@ async function synthAudio(client, createHash, retrieveHash, logger, stats, { acc
   logger = logger || noopLogger;
 
   assert.ok(['google', 'aws', 'polly', 'microsoft', 'wellsaid', 'nuance', 'nvidia', 'ibm', 'elevenlabs',
-    'whisper', 'deepgram', 'playht', 'rimelabs', 'verbio', 'cartesia', 'inworld'].includes(vendor) ||
+    'whisper', 'deepgram', 'playht', 'rimelabs', 'verbio', 'cartesia', 'inworld', 'resemble'].includes(vendor) ||
   vendor.startsWith('custom'),
   `synthAudio supported vendors are google, aws, microsoft, nuance, nvidia and wellsaid ..etc, not ${vendor}`);
   if ('google' === vendor) {
@@ -166,6 +165,9 @@ async function synthAudio(client, createHash, retrieveHash, logger, stats, { acc
   } else if ('cartesia' === vendor) {
     assert.ok(credentials.api_key, 'synthAudio requires api_key when cartesia is used');
     assert.ok(credentials.model_id, 'synthAudio requires model_id when cartesia is used');
+  } else if (vendor === 'resemble') {
+    assert.ok(voice, 'synthAudio requires voice when resemble is used');
+    assert.ok(credentials.api_key, 'synthAudio requires api_key when resemble is used');
   }
 
   const key = makeSynthKey({
@@ -202,70 +204,74 @@ async function synthAudio(client, createHash, retrieveHash, logger, stats, { acc
     const startAt = process.hrtime();
     switch (vendor) {
       case 'google':
-        audioData = await synthGoogle(logger, {credentials, stats, language, voice, gender, text});
+        audioData = await synthGoogle(logger, {credentials, stats, language, voice, gender, key, text});
         break;
       case 'aws':
       case 'polly':
         vendorLabel = 'aws';
         audioData = await synthPolly(createHash, retrieveHash, logger,
-          {credentials, stats, language, voice, text, engine});
+          {credentials, stats, language, voice, key, text, engine, renderForCaching, disableTtsStreaming});
         break;
       case 'azure':
       case 'microsoft':
         vendorLabel = 'microsoft';
-        audioData = await synthMicrosoft(logger, {credentials, stats, language, voice, text, deploymentId,
+        audioData = await synthMicrosoft(logger, {credentials, stats, language, voice, key, text, deploymentId,
           renderForCaching, disableTtsStreaming});
         break;
       case 'nuance':
         model = model || 'enhanced';
-        audioData = await synthNuance(client, logger, {credentials, stats, voice, model, text});
+        audioData = await synthNuance(client, logger, {credentials, stats, voice, model, key, text});
         break;
       case 'nvidia':
-        audioData = await synthNvidia(client, logger, {credentials, stats, language, voice, model, text,
+        audioData = await synthNvidia(client, logger, {credentials, stats, language, voice, model, key, text,
           renderForCaching, disableTtsStreaming});
         break;
       case 'ibm':
-        audioData = await synthIbm(logger, {credentials, stats, voice, text});
+        audioData = await synthIbm(logger, {credentials, stats, voice, key, text});
         break;
       case 'wellsaid':
-        audioData = await synthWellSaid(logger, {credentials, stats, language, voice, text});
+        audioData = await synthWellSaid(logger, {credentials, stats, language, voice, key, text});
         break;
       case 'elevenlabs':
         audioData = await synthElevenlabs(logger, {
-          credentials, options, stats, language, voice, text, renderForCaching, disableTtsStreaming});
+          credentials, options, stats, language, voice, key, text, renderForCaching, disableTtsStreaming});
         break;
       case 'playht':
         audioData = await synthPlayHT(client, logger, {
-          credentials, options, stats, language, voice, text, renderForCaching, disableTtsStreaming});
+          credentials, options, stats, language, voice, key, text, renderForCaching, disableTtsStreaming});
         break;
       case 'cartesia':
         audioData = await synthCartesia(logger, {
-          credentials, options, stats, language, voice, text, renderForCaching, disableTtsStreaming});
+          credentials, options, stats, language, voice, key, text, renderForCaching, disableTtsStreaming});
         break;
       case 'inworld':
         audioData = await synthInworld(logger, {
-          credentials, options, stats, language, voice, text, renderForCaching, disableTtsStreaming});
+          credentials, options, stats, language, voice, key, text, renderForCaching, disableTtsStreaming});
         break;
       case 'rimelabs':
         audioData = await synthRimelabs(logger, {
-          credentials, options, stats, language, voice, text, renderForCaching, disableTtsStreaming});
+          credentials, options, stats, language, voice, key, text, renderForCaching, disableTtsStreaming});
         break;
       case 'whisper':
         audioData = await synthWhisper(logger, {
-          credentials, stats, voice, text, instructions, renderForCaching, disableTtsStreaming});
+          credentials, stats, voice, key, text, instructions, renderForCaching, disableTtsStreaming});
         break;
       case 'verbio':
         audioData = await synthVerbio(client, logger, {
-          credentials, stats, voice, text, renderForCaching, disableTtsStreaming});
+          credentials, stats, voice, key, text, renderForCaching, disableTtsStreaming});
         if (audioData?.filePath) return audioData;
         break;
       case 'deepgram':
-        audioData = await synthDeepgram(logger, {credentials, stats, model, text,
+        audioData = await synthDeepgram(logger, {credentials, stats, model, key, text,
           renderForCaching, disableTtsStreaming});
         break;
+      case 'resemble':
+        audioData = await synthResemble(logger, {
+          credentials, stats, voice, key, text, options, renderForCaching, disableTtsStreaming});
+        break;
       case vendor.startsWith('custom') ? vendor : 'cant_match_value':
         audioData = await synthCustomVendor(logger,
-          {credentials, stats, language, voice, text});
+          {credentials, stats, language, voice, key, text});
         break;
       default:
         assert(`synthAudio: unsupported speech vendor ${vendor}`);
@@ -300,9 +306,43 @@ async function synthAudio(client, createHash, retrieveHash, logger, stats, { acc
 }
 
 const synthPolly = async(createHash, retrieveHash, logger,
-  {credentials, stats, language, voice, engine, text}) => {
+  {credentials, stats, language, voice, engine, key, text, renderForCaching, disableTtsStreaming}) => {
+  const {region, accessKeyId, secretAccessKey, roleArn} = credentials;
+  if (!JAMBONES_DISABLE_TTS_STREAMING && !renderForCaching && !disableTtsStreaming) {
+
+    let params = '{';
+    params += `language=${language}`;
+    params += `,playback_id=${key}`;
+    params += ',write_cache_file=1';
+    params += ',vendor=aws';
+    if (accessKeyId && secretAccessKey) {
+      if (accessKeyId) params += `,accessKeyId=${accessKeyId}`;
+      if (secretAccessKey) params += `,secretAccessKey=${secretAccessKey}`;
+    } else if (roleArn) {
+      const cred = await getAwsAuthToken(
+        logger, createHash, retrieveHash,
+        {
+          region,
+          roleArn
+        });
+
+      if (cred) {
+        params += `,accessKeyId=${cred.accessKeyId}`;
+        params += `,secretAccessKey=${cred.secretAccessKey}`;
+        params += `,sessionToken=${cred.sessionToken}`;
+      }
+    }
+    if (region) params += `,region=${region}`;
+    if (engine) params += `,engine=${engine}`;
+    params += '}';
+
+    return {
+      filePath: `say:${params}${text.replace(/\n/g, ' ').replace(/\r/g, ' ')}`,
+      servedFromCache: false,
+      rtt: 0
+    };
+  }
   try {
-    const {region, accessKeyId, secretAccessKey, roleArn} = credentials;
     let polly;
     if (accessKeyId && secretAccessKey) {
       polly = new PollyClient({
@@ -520,6 +560,7 @@ const synthMicrosoft = async(logger, {
   stats,
   language,
   voice,
+  key,
   text,
   renderForCaching,
   disableTtsStreaming
@@ -549,8 +590,9 @@ const synthMicrosoft = async(logger, {
     }
     if (!JAMBONES_DISABLE_TTS_STREAMING && !JAMBONES_DISABLE_AZURE_TTS_STREAMING &&
       !renderForCaching && !disableTtsStreaming) {
-      let params = '';
-      params += `{api_key=${apiKey}`;
+      let params = '{';
+      params += `api_key=${apiKey}`;
+      params += `,playback_id=${key}`;
       params += `,language=${language}`;
       params += ',vendor=microsoft';
       params += `,voice=${voice}`;
@@ -727,12 +769,13 @@ const synthNuance = async(client, logger, {credentials, stats, voice, model, tex
 };
 
 const synthNvidia = async(client, logger, {
-  credentials, stats, language,  voice, model, text, renderForCaching, disableTtsStreaming
+  credentials, stats, language,  voice, model, key, text, renderForCaching, disableTtsStreaming
 }) => {
   const {riva_server_uri} = credentials;
   if (!JAMBONES_DISABLE_TTS_STREAMING && !renderForCaching && !disableTtsStreaming) {
     let params = '';
     params += `{riva_server_uri=${riva_server_uri}`;
+    params += `,playback_id=${key}`;
     params += `,voice=${voice}`;
     params += `,language=${language}`;
     params += ',write_cache_file=1';
@@ -806,15 +849,16 @@ const synthCustomVendor = async(logger, {credentials, stats, language, voice, te
 };
 
 const synthElevenlabs = async(logger, {
-  credentials, options, stats, voice, text, renderForCaching, disableTtsStreaming
+  credentials, options, stats, voice, key, text, renderForCaching, disableTtsStreaming
 }) => {
   const {api_key, model_id, options: credOpts} = credentials;
   const opts = !!options && Object.keys(options).length !== 0 ? options : JSON.parse(credOpts || '{}');
 
   /* default to using the streaming interface, unless disabled by env var OR we want just a cache file */
   if (!JAMBONES_DISABLE_TTS_STREAMING && !renderForCaching && !disableTtsStreaming) {
-    let params = '';
-    params += `{api_key=${api_key}`;
+    let params = '{';
+    params += `api_key=${api_key}`;
+    params += `,playback_id=${key}`;
     params += ',vendor=elevenlabs';
     params += `,voice=${voice}`;
     params += `,model_id=${model_id}`;
@@ -869,7 +913,7 @@ const synthElevenlabs = async(logger, {
 };
 
 const synthPlayHT = async(client, logger, {
-  credentials, options, stats, voice, language, text, renderForCaching, disableTtsStreaming
+  credentials, options, stats, voice, language, key, text, renderForCaching, disableTtsStreaming
 }) => {
   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 || '{}');
@@ -903,8 +947,9 @@ const synthPlayHT = async(client, logger, {
 
   /* default to using the streaming interface, unless disabled by env var OR we want just a cache file */
   if (!JAMBONES_DISABLE_TTS_STREAMING && !renderForCaching && !disableTtsStreaming) {
-    let params = '';
-    params += `{api_key=${api_key}`;
+    let params = '{';
+    params += `api_key=${api_key}`;
+    params += `,playback_id=${key}`;
     params += `,user_id=${user_id}`;
     params += ',vendor=playht';
     params += `,voice=${voice}`;
@@ -963,15 +1008,16 @@ const synthPlayHT = async(client, logger, {
 };
 
 const synthInworld = async(logger, {
-  credentials, options, stats, voice, text, renderForCaching, disableTtsStreaming
+  credentials, options, stats, voice, key, text, renderForCaching, disableTtsStreaming
 }) => {
   const {api_key, model_id, options: credOpts} = credentials;
   const opts = !!options && Object.keys(options).length !== 0 ? options : JSON.parse(credOpts || '{}');
 
   /* default to using the streaming interface, unless disabled by env var OR we want just a cache file */
   if (!JAMBONES_DISABLE_TTS_STREAMING && !renderForCaching && !disableTtsStreaming) {
-    let params = '';
-    params += `{api_key=${api_key}`;
+    let params = '{';
+    params += `api_key=${api_key}`;
+    params += `,playback_id=${key}`;
     params += `,model_id=${model_id}`;
     params += ',vendor=inworld';
     params += `,voice=${voice}`;
@@ -1027,15 +1073,16 @@ const synthInworld = async(logger, {
 };
 
 const synthRimelabs = async(logger, {
-  credentials, options, stats, language, voice, text, renderForCaching, disableTtsStreaming
+  credentials, options, stats, language, voice, key, text, renderForCaching, disableTtsStreaming
 }) => {
   const {api_key, model_id, options: credOpts} = credentials;
   const opts = !!options && Object.keys(options).length !== 0 ? options : JSON.parse(credOpts || '{}');
 
   /* default to using the streaming interface, unless disabled by env var OR we want just a cache file */
   if (!JAMBONES_DISABLE_TTS_STREAMING && !renderForCaching && !disableTtsStreaming) {
-    let params = '';
-    params += `{api_key=${api_key}`;
+    let params = '{';
+    params += `api_key=${api_key}`;
+    params += `,playback_id=${key}`;
     params += `,model_id=${model_id}`;
     params += ',vendor=rimelabs';
     params += `,language=${language}`;
@@ -1083,15 +1130,18 @@ const synthRimelabs = async(logger, {
     throw err;
   }
 };
-const synthVerbio = async(client, logger, {credentials, stats, voice, text, renderForCaching, disableTtsStreaming}) => {
+const synthVerbio = async(client, logger, {
+  credentials, stats, voice, key, text, renderForCaching, disableTtsStreaming
+}) => {
   //https://doc.speechcenter.verbio.com/#tag/Text-To-Speech-REST-API
   if (text.length > 2000) {
     throw new Error('Verbio cannot synthesize for the text length larger than 2000 characters');
   }
   const token = await getVerbioAccessToken(client, logger, credentials);
   if (!process.env.JAMBONES_DISABLE_TTS_STREAMING && !renderForCaching && !disableTtsStreaming) {
-    let params = '';
-    params += `{access_token=${token.access_token}`;
+    let params = '{';
+    params += `access_token=${token.access_token}`;
+    params += `,playback_id=${key}`;
     params += ',vendor=verbio';
     params += `,voice=${voice}`;
     params += ',write_cache_file=1';
@@ -1128,13 +1178,14 @@ const synthVerbio = async(client, logger, {credentials, stats, voice, text, rend
   }
 };
 
-const synthWhisper = async(logger, {credentials, stats, voice, text, instructions,
+const synthWhisper = async(logger, {credentials, stats, voice, key, text, instructions,
   renderForCaching, disableTtsStreaming}) => {
   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 (!JAMBONES_DISABLE_TTS_STREAMING && !renderForCaching && !disableTtsStreaming) {
-    let params = '';
-    params += `{api_key=${api_key}`;
+    let params = '{';
+    params += `api_key=${api_key}`;
+    params += `,playback_id=${key}`;
     params += `,model_id=${model_id}`;
     params += ',vendor=whisper';
     params += `,voice=${voice}`;
@@ -1176,11 +1227,12 @@ const synthWhisper = async(logger, {credentials, stats, voice, text, instruction
   }
 };
 
-const synthDeepgram = async(logger, {credentials, stats, model, text, renderForCaching, disableTtsStreaming}) => {
+const synthDeepgram = async(logger, {credentials, stats, model, key, text, renderForCaching, disableTtsStreaming}) => {
   const {api_key, deepgram_tts_uri} = credentials;
   if (!JAMBONES_DISABLE_TTS_STREAMING && !renderForCaching && !disableTtsStreaming) {
-    let params = '';
-    params += `{api_key=${api_key}`;
+    let params = '{';
+    params += `api_key=${api_key}`;
+    params += `,playback_id=${key}`;
     params += ',vendor=deepgram';
     params += `,voice=${model}`;
     params += ',write_cache_file=1';
@@ -1216,14 +1268,15 @@ const synthDeepgram = async(logger, {credentials, stats, model, text, renderForC
 };
 
 const synthCartesia = async(logger, {
-  credentials, options, stats, voice, language, text, renderForCaching, disableTtsStreaming
+  credentials, options, stats, voice, language, key, text, renderForCaching, disableTtsStreaming
 }) => {
   const {api_key, model_id, embedding, options: credOpts} = credentials;
   const opts = !!options && Object.keys(options).length !== 0 ? options : JSON.parse(credOpts || '{}');
 
   if (!JAMBONES_DISABLE_TTS_STREAMING && !renderForCaching && !disableTtsStreaming) {
-    let params = '';
-    params += `{api_key=${api_key}`;
+    let params = '{';
+    params += `api_key=${api_key}`;
+    params += `,playback_id=${key}`;
     params += `,model_id=${model_id}`;
     params += ',vendor=cartesia';
     params += `,voice=${voice}`;
@@ -1285,6 +1338,63 @@ const synthCartesia = async(logger, {
 
 };
 
+const synthResemble = async(logger, {
+  credentials, options, stats, voice, key, text, renderForCaching, disableTtsStreaming
+}) => {
+  const {api_key, resemble_tts_uri, resemble_tts_use_tls} = credentials;
+  const {project_uuid, use_hd} = options || {};
+
+  /* default to using the streaming interface, unless disabled by env var OR we want just a cache file */
+  if (!JAMBONES_DISABLE_TTS_STREAMING && !renderForCaching && !disableTtsStreaming) {
+    let params = '{';
+    params += `api_key=${api_key}`;
+    params += `,playback_id=${key}`;
+    params += ',vendor=resemble';
+    params += `,voice=${voice}`;
+    params += ',write_cache_file=1';
+    if (project_uuid) params += `,project_uuid=${project_uuid}`;
+    if (use_hd) params += `,use_hd=${use_hd}`;
+    if (resemble_tts_uri) params += `,endpoint=${resemble_tts_uri}`;
+    if (resemble_tts_use_tls) params += `,use_tls=${resemble_tts_use_tls}`;
+
+    params += '}';
+
+    return {
+      filePath: `say:${params}${text.replace(/\n/g, ' ').replace(/\r/g, ' ')}`,
+      servedFromCache: false,
+      rtt: 0
+    };
+  }
+
+  try {
+    const baseUrl = resemble_tts_uri || 'https://f.cluster.resemble.ai';
+    const post = bent(baseUrl, 'POST', 'buffer', {
+      'Authorization': `Bearer ${api_key}`,
+      'Content-Type': 'application/json'
+    });
+    const response = await post('/synthesize', {
+      voice_uuid: voice,
+      data: text,
+      sample_rate: 8000,
+      output_format: 'mp3',
+      ...(project_uuid && {project_uuid}),
+      ...(use_hd && {use_hd}),
+    });
+
+    const json = JSON.parse(response.toString('utf8'));
+    const audioContent = Buffer.from(json.audio_content, 'base64');
+    return {
+      audioContent,
+      extension: 'mp3',
+      sampleRate: 8000
+    };
+  } catch (err) {
+    logger.info({err}, 'synth Elevenlabs returned error');
+    stats.increment('tts.count', ['vendor:elevenlabs', 'accepted:no']);
+    throw err;
+  }
+};
+
 const getFileExtFromMime = (mime) => {
   switch (mime) {
     case 'audio/wav':

package.json

@@ -1,6 +1,6 @@
 {
   "name": "@jambonz/speech-utils",
-  "version": "0.2.13",
+  "version": "0.2.22",
   "description": "TTS-related speech utilities for jambonz",
   "main": "index.js",
   "author": "Dave Horton",
@@ -37,7 +37,7 @@
     "debug": "^4.3.4",
     "form-urlencoded": "^6.1.4",
     "google-protobuf": "^3.21.2",
-    "ibm-watson": "^8.0.0",
+    "ibm-watson": "^11.0.0",
     "microsoft-cognitiveservices-speech-sdk": "1.38.0",
     "openai": "^4.98.0",
     "undici": "^7.5.0"

test/synth.js

@@ -185,6 +185,7 @@ test('AWS speech synth tests', async(t) => {
       language: 'en-US',
       voice: 'Joey',
       text: 'This is a test.  This is only a test',
+      renderForCaching: true,
     });
     t.ok(!opts.servedFromCache, `successfully synthesized aws audio to ${opts.filePath}`);
 
@@ -198,6 +199,7 @@ test('AWS speech synth tests', async(t) => {
       language: 'en-US',
       voice: 'Joey',
       text: 'This is a test.  This is only a test',
+      renderForCaching: true,
     });
     t.ok(opts.servedFromCache, `successfully retrieved aws audio from cache ${opts.filePath}`);
   } catch (err) {
@@ -750,6 +752,35 @@ test('inworld speech synth', async(t) => {
   client.quit();
 });
 
+test('resemble speech synth', async(t) => {
+  const fn = require('..');
+  const {synthAudio, client} = fn(opts, logger);
+
+  if (!process.env.RESEMBLE_API_KEY) {
+    t.pass('skipping resemble speech synth tests since RESEMBLE_API_KEY is not provided');
+    return t.end();
+  }
+  const text = '<speak prompt="Speak in an excited, upbeat tone">Hello from Resemble!</speak>';
+  try {
+    const opts = await synthAudio(stats, {
+      vendor: 'resemble',
+      credentials: {
+        api_key: process.env.RESEMBLE_API_KEY,
+      },
+      language: 'en',
+      voice: '3f5fb9f1',
+      text,
+      renderForCaching: true
+    });
+    t.ok(!opts.servedFromCache, `successfully synthesized resemble audio to ${opts.filePath}`);
+
+  } catch (err) {
+    console.error(JSON.stringify(err));
+    t.end(err);
+  }
+  client.quit();
+});
+
 test('rimelabs speech synth tests mist', async(t) => {
   const fn = require('..');
   const {synthAudio, client} = fn(opts, logger);