ben.carrasco/speech-utils
1
0
Fork 0
mirror of https://github.com/jambonz/speech-utils.git synced 2025-12-19 03:37:49 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat/893 support default azure ssml config

Browse Source
This commit is contained in:
vasudevanubrolu
2025-05-12 11:30:54 +05:30
parent 59bca302b9
commit 08b55b8d79
1 changed files with 13 additions and 16 deletions
Download Patch File Download Diff File Expand all files Collapse all files

29
lib/synth-audio.js
View File

@@ -516,28 +516,25 @@ const synthMicrosoft = async(logger, {
const {api_key: apiKey, region, use_custom_tts, custom_tts_endpoint, custom_tts_endpoint_url} = credentials;
// let clean up the text
let content = text;
if (!JAMBONES_AZURE_ENABLE_SSML) {
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"
*/
content = `<speak>${text}</speak>`;
}
if (use_custom_tts && !content.startsWith('<speak')) {
/**
* Note: it seems that to use custom voice ssml is required with the voice attribute
* Otherwise sending plain text we get "Voice does not match"
*/
content = `<speak>${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 */
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}">${words}</voice></speak>`;
logger.info({content}, 'synthMicrosoft');
}
} else {
if (content.startsWith('<speak>')) {
/* microsoft enforces some properties and uses voice xml element so if the user did not supply do it for them */
const words = content.slice(7, -8).trim().replace(/(\r\n|\n|\r)/gm, ' ');
// eslint-disable-next-line max-len
content = `<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xmlns:mstts="https://www.w3.org/2001/mstts" xml:lang="${language}"><voice name="${voice}">${words}</voice></speak>`;
logger.info({content}, 'synthMicrosoft');
}
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}">${text}</voice></speak>`;
}
if (!JAMBONES_DISABLE_TTS_STREAMING && !JAMBONES_DISABLE_AZURE_TTS_STREAMING &&
!renderForCaching && !disableTtsStreaming) {
let params = '';