Mosa Lingua Audio File Generator

Preparing Configuration files

Repository contains two example config files, to use in testing.

Voice Map JSON struct

Mapping of voices into speech voice templates, example en_voices_mosa_playht.json

{
	"man1" 		: "en-US-Wavenet-D",
	" " 		: "en-US-AriaNeural",
	"David" 	: "en-US-Wavenet-D",
	"Fabrice"	: "en-US-Wavenet-D",
	"Nadia"		: "en-US-AriaNeural"
}

Left column(json keys) contains aliases used in CSV. Right column(json values) contains appropriate Azure Speech vocals to use in audio rendering

Conversion CSV File Struct

CSV file with, example en_dialogues_to_generate.csv

# This is a comment and this line should be ignored
We rented an apartment in Queens. Once situated, we have the rest of the afternoon for a day of exploration in New York!;Nadia;1640575
Hi Alex, nice to meet you! (5) I'm thirty [30 need to be removed][20](10).;David;1639844
Thursday is my brother's birthday, do you want to come? (need to be removed too).;Nadia;23120

# we can have comments in the middle too
OK, let's do it;UnknownNameRaiseError;1234567

Running the script

Clone this repository

git clone https://github.com/seeker1983/mosa-lingua-speech-generator

Install dependencies

npm install

Setup API key and Region in azure-speech.js

var key = "<API_KEY>";
var region = "<REGION_CODE>";

Execute script using provided files

Run script using nodejs with specified parameters:

node generate.js en_voices_mosa_playht.json en_dialogues_to_generate.csv

Upon successful execution files

• en_1639844.wav
• en_1640575.wav
• en_23120.wav

should appear in execution folder.