CMU Arctic speech data coverter

This repository contains the PowerShell Core script to convert the audio and talking script data of CMU Arctic speech synthesis database (such as this) into Azure Cognitive service - Custom Voice acceptable format.

Invoke the script like the following, if you extract the CMU arctic speech data zip file in data folder:

./convert_name.ps1 -sourceScript ./data/cmu_us_slt_arctic/etc/arctic.data -targetScript ./output/script.txt -wavSource ./data/cmu_us_slt_arctic/wav -wavTarget ./output/wav

The output folder will contain the resulting script.txt & wav.zip files, that are acceptable input training data that you can upload to Azure Custom Voice portal.