sanskrit_tts

A simple python library for converting Sanskrit Text-to-Speech (TTS). The supported TTS engines are:

Both of these API options require authentication, in the form of Google cloud credentials, or Bhashini API key. The developers of Bhashini have generously provided an API key for non-commercial, limited usage of the API for creating audio of Sanskrit texts. This can be used via the bhashini proxy (see Usage below). Please note that the Bhashini proxy (the default option) should not be used for other purposes.

Installation

This package uses pydub for managing audio data, which in turn requires ffmpeg or libav. Please check the details (here)[https://github.com/jiaaro/pydub#dependencies].

This package should work with any version of python >= 3.8.

pip install sanskrit_tts

To install from the master branch of the git repo:

pip install git+https://github.com/avinashvarna/sanskrit_tts.git

For an editable installation (to modify the code and experiment)

git clone https://github.com/avinashvarna/sanskrit_tts.git
cd sanskrit_tts
pip install -e .

Usage

All TTS classes expose the same interface, so that switching should be fairly easy.

Default - Bhashini proxy with embedded API key

from sanskrit_tts import default_tts

text = "तैत्तिरीयोपनिषत् प्रसिद्धासु दशसु उपनिषत्सु अन्यतमा ।"
TTS = default_tts()
audio = TTS.synthesize(text)
# Export the audio as an MP3
audio.export("sanskrit_speech.mp3")

Bhashini API with key

from sanskrit_tts.bhashini_tts import BhashiniTTS

text = "तैत्तिरीयोपनिषत् प्रसिद्धासु दशसु उपनिषत्सु अन्यतमा ।"
api_key = ...
TTS = BhashiniTTS(api_key=api_key)
audio = TTS.synthesize(text)
# Export the audio as an MP3
audio.export("sanskrit_speech.mp3")

Google Cloud

Requires credentials, e.g. from a (service account)[https://cloud.google.com/iam/docs/creating-managing-service-accounts].

import os
from sanskrit_tts.gcloud_tts import GCloudTTS

# Setup credentials
os.environ['GOOGLE_APPLICATION_CREDENTIALS'] = './credentials.json'

text = "तैत्तिरीयोपनिषत् प्रसिद्धासु दशसु उपनिषत्सु अन्यतमा ।"
TTS = GCloudTTS()
audio = TTS.synthesize(text)
# Export the audio as an MP3
audio.export("sanskrit_speech.mp3")

How it works

Both Google Cloud TTS and Bhashini Text-to-Speech do not support Sanskrit yet. As a workaround, this library uses other languages for speech to text conversion. Kannada is used by default for this workaround. Any other language/voice supported by the corresponding TTS API can be used by changing the appropriate parameters while instantiating the TTS class, and the results will vary. A complete list of voices supported by Google Cloud TTS is available here. For Bhashini, please check the (demo)[https://tts.bhashini.ai/demo/].