0.2.18

support mod_aws_tts

lib/config.js

@@ -2,13 +2,14 @@ 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;
 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;
 
@@ -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');
@@ -97,7 +98,7 @@ 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'].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) {
@@ -140,6 +141,10 @@ async function synthAudio(client, createHash, retrieveHash, logger, stats, { acc
     assert.ok(credentials.api_key, 'synthAudio requires api_key when playht is used');
     assert.ok(credentials.user_id, 'synthAudio requires user_id when playht is used');
     assert.ok(credentials.voice_engine, 'synthAudio requires voice_engine when playht is used');
+  } else if ('inworld' === vendor) {
+    assert.ok(voice, 'synthAudio requires voice when inworld is used');
+    assert.ok(credentials.api_key, 'synthAudio requires api_key when inworld is used');
+    assert.ok(credentials.model_id, 'synthAudio requires model_id when inworld is used');
   } else if ('rimelabs' === vendor) {
     assert.ok(voice, 'synthAudio requires voice when rimelabs is used');
     assert.ok(credentials.api_key, 'synthAudio requires api_key when rimelabs is used');
@@ -161,6 +166,9 @@ 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({
@@ -203,7 +211,7 @@ async function synthAudio(client, createHash, retrieveHash, logger, stats, { acc
       case 'polly':
         vendorLabel = 'aws';
         audioData = await synthPolly(createHash, retrieveHash, logger,
-          {credentials, stats, language, voice, text, engine});
+          {credentials, stats, language, voice, text, engine, renderForCaching, disableTtsStreaming});
         break;
       case 'azure':
       case 'microsoft':
@@ -237,6 +245,10 @@ async function synthAudio(client, createHash, retrieveHash, logger, stats, { acc
         audioData = await synthCartesia(logger, {
           credentials, options, stats, language, voice, text, renderForCaching, disableTtsStreaming});
         break;
+      case 'inworld':
+        audioData = await synthInworld(logger, {
+          credentials, options, stats, language, voice, text, renderForCaching, disableTtsStreaming});
+        break;
       case 'rimelabs':
         audioData = await synthRimelabs(logger, {
           credentials, options, stats, language, voice, text, renderForCaching, disableTtsStreaming});
@@ -254,6 +266,10 @@ async function synthAudio(client, createHash, retrieveHash, logger, stats, { acc
         audioData = await synthDeepgram(logger, {credentials, stats, model, text,
           renderForCaching, disableTtsStreaming});
         break;
+      case 'resemble':
+        audioData = await synthResemble(logger, {
+          credentials, stats, voice, text, options, renderForCaching, disableTtsStreaming});
+        break;
       case vendor.startsWith('custom') ? vendor : 'cant_match_value':
         audioData = await synthCustomVendor(logger,
           {credentials, stats, language, voice, text});
@@ -291,9 +307,41 @@ 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, text, renderForCaching, disableTtsStreaming}) => {
+  const {region, accessKeyId, secretAccessKey, roleArn} = credentials;
+  if (!JAMBONES_DISABLE_TTS_STREAMING && !renderForCaching && !disableTtsStreaming) {
+
+    let params = '{';
+    params += `language=${language}`;
+    params += ',write_cache_file=1';
+    params += ',vendor=aws';
+    if (accessKeyId && secretAccessKey) {
+      if (accessKeyId) params += `,accessKeyId=${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=${region}`;
+    params += '}';
+
+    return {
+      filePath: `say:${params}${text.replace(/\n/g, ' ').replace(/\r/g, ' ')}`,
+      servedFromCache: false,
+      rtt: 0
+    };
+  }
   try {
-    const {region, accessKeyId, secretAccessKey, roleArn} = credentials;
     let polly;
     if (accessKeyId && secretAccessKey) {
       polly = new PollyClient({
@@ -466,7 +514,6 @@ async function _synthOnPremMicrosoft(logger, {
 }) {
   const {use_custom_tts, custom_tts_endpoint_url, api_key} = credentials;
   let content = text;
-
   if (use_custom_tts && !content.startsWith('<speak')) {
     /**
      * Note: it seems that to use custom voice ssml is required with the voice attribute
@@ -482,6 +529,10 @@ async function _synthOnPremMicrosoft(logger, {
     content = `<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xmlns:mstts="https://www.w3.org/2001/mstts" xml:lang="${language}"><voice name="${voice}">${words}</voice></speak>`;
     logger.info({content}, 'synthMicrosoft');
   }
+  else if (JAMBONES_AZURE_ENABLE_SSML && !content.startsWith('<speak')) {
+    // 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 +569,23 @@ 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}">${words}</voice></speak>`;
       logger.info({content}, 'synthMicrosoft');
     }
+    else if (JAMBONES_AZURE_ENABLE_SSML && !content.startsWith('<speak')) {
+      // eslint-disable-next-line max-len
+      content = `<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xmlns:mstts="https://www.w3.org/2001/mstts" xml:lang="${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 +910,10 @@ 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 {
@@ -947,6 +1001,70 @@ const synthPlayHT = async(client, logger, {
   }
 };
 
+const synthInworld = async(logger, {
+  credentials, options, stats, voice, 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=${api_key}`;
+    params += `,model_id=${model_id}`;
+    params += ',vendor=inworld';
+    params += `,voice=${voice}`;
+    params += ',write_cache_file=1';
+    if (opts.temperature) params += `,temperature=${opts.temperature}`;
+    if (opts.audioConfig?.pitch) params += `,pitch=${opts.pitch}`;
+    if (opts.audioConfig?.speakingRate) params += `,speakingRate=${opts.speakingRate}`;
+    params += '}';
+
+    return {
+      filePath: `say:${params}${text.replace(/\n/g, ' ').replace(/\r/g, ' ')}`,
+      servedFromCache: false,
+      rtt: 0
+    };
+  }
+
+  try {
+    const url = 'https://api.inworld.ai/tts/v1/voice';
+    const sampleRate = 8000;
+    const options = {
+      method: 'POST',
+      headers: {
+        'Authorization': `Basic ${api_key}`,
+        'Content-Type': 'application/json'
+      },
+      body: JSON.stringify({
+        text,
+        voiceId: voice,
+        modelId: model_id,
+        audioConfig: {
+          ...(opts.audioConfig || {}),
+          audioEncoding: 'MP3',
+        }
+      })
+    };
+
+    const response = await fetch(url, options);
+
+    if (!response.ok) {
+      throw new Error(await response.text());
+    }
+    const json = await response.json();
+    return {
+      audioContent: Buffer.from(json.audioContent, 'base64'),
+      extension: 'mp3',
+      sampleRate
+    };
+  } catch (err) {
+    logger.info({err}, 'synth inworld returned error');
+    stats.increment('tts.count', ['vendor:inworld', 'accepted:no']);
+    throw err;
+  }
+};
+
 const synthRimelabs = async(logger, {
   credentials, options, stats, language, voice, text, renderForCaching, disableTtsStreaming
 }) => {
@@ -1206,6 +1324,62 @@ const synthCartesia = async(logger, {
 
 };
 
+const synthResemble = async(logger, {
+  credentials, options, stats, voice, 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=${api_key}`;
+    params += ',vendor=resemble';
+    params += `,voice=${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": "0.2.9",
+  "version": "0.2.18",
   "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) {
@@ -720,6 +722,65 @@ test('Cartesia speech synth tests', async(t) => {
   client.quit();
 });
 
+test('inworld speech synth', async(t) => {
+  const fn = require('..');
+  const {synthAudio, client} = fn(opts, logger);
+
+  if (!process.env.INWORLD_API_KEY) {
+    t.pass('skipping inworld speech synth tests since INWORLD_API_KEY is not provided');
+    return t.end();
+  }
+  const text = 'Hi there and welcome to jambones!';
+  try {
+    const opts = await synthAudio(stats, {
+      vendor: 'inworld',
+      credentials: {
+        api_key: process.env.INWORLD_API_KEY,
+        model_id: 'inworld-tts-1'
+      },
+      language: 'en',
+      voice: 'Ashley',
+      text,
+      renderForCaching: true
+    });
+    t.ok(!opts.servedFromCache, `successfully synthesized inworld audio to ${opts.filePath}`);
+
+  } catch (err) {
+    console.error(JSON.stringify(err));
+    t.end(err);
+  }
+  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);