MeloTTS Server

This is TTS Server use Melo TTS as a base, We do some changes in the code base and created a FAST API server by which we get good performance for 100 chars we get response about 200~230 milliseconds latency using G5 aws instance

Install

python3 -m venv env
. ./env/bin/activate
pip install -r requirements.txt
python3 -m unidic download

Server Start

python3 Server.py

Request

{
    "voice_id": "EN-US",
    "text": "Every morning, Sarah walked her dog, Max, through the old, enchanting forest near her home. One day, thanyou have a nice day",
    "sr": 8000,
    "sdp_ratio" : 0.2,
    "noise_scale" : 0.6,
    "noise_scale_w" :  0.8,
    "speed" : 1.0
}

there are 5 diffrent type of voices ids

• 'EN-US'
• 'EN-BR'
• 'EN-AU'
• 'EN-Default'
• 'EN_INDIA'

Response

{
    "audio": "audio file in base64",
    "sr": "audio sample rate",
    "time": "time required for genrate audio"
}