ben.carrasco/speech-utils
1
0
Fork 0
mirror of https://github.com/jambonz/speech-utils.git synced 2026-07-23 12:41:48 +00:00
Code Issues Packages Projects Releases Wiki Activity

support whisper streaming

Browse Source
This commit is contained in:
Quan HL
2024-02-05 11:49:38 +07:00
parent 436b15d648
commit 3cf9894b44
1 changed files with 24 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
lib/synth-audio.js
+24 -4
View File
@@ -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':
@@ -608,11 +612,11 @@ const synthCustomVendor = async(logger, {credentials, stats, language, voice, te
};
const synthElevenlabs = async(logger, {credentials, options, stats, language, voice, text, renderForCaching}) => {
const {api_key, model_id, options: credOpts} = credentials;
const {api_key, model_id, options: credOpts, use_streaming} = 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) {
if (use_streaming && !renderForCaching) {
let params = '';
params += `{api_key=${api_key}`;
params += `,model_id=${model_id}`;
@@ -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, use_streaming, speed} = credentials;
/* if the env is set to stream then bag out, unless we are specifically rendering to generate a cache file */
if (use_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,