Vāgdhenu: A New Sanskrit Chanting TTS System for Accurate Recitation
Vāgdhenu: A Sanskrit Chanting TTS System
I built Vāgdhenu, a specialized text-to-speech system that transforms Sanskrit verses into authentic chants by detecting meter and preserving complex phonetics. The model, retrained on a dedicated corpus, powers the Bhāgavata-VāNi app for the Śrīmad Bhāgavatam and Vāgbodhinī, a tutor that scores your pronunciation syllable by syllable. This tool overcomes previous technical hurdles to render dense conjuncts and retroflex sounds with expert-level accuracy.
The retrained model reaches an expert MOS of about 4.6, and dense conjuncts — including retroflex aspirates — render correctly, the class earlier architectures could not crack.
- rramadass
People seem to be making all sorts of assumptions/conjectures about the project and confusing themselves without reading the paper explaining the project (linked to in the site itself);
From the introduction;
Classical Sanskrit recitation, parāyaṇa, is a chanted rather than a read register. A faithful synthesizer must hold long vowels, sustain a terminal visarga, articulate retroflex and aspirated consonants, render dense consonant conjuncts cleanly, and respect the metrical structure of the verse.
None of these is well served by general-purpose text-to-speech, and there is essentially no chant-domain training data available off the shelf. The problem is therefore doubly hard: it is low-resource, and the target prosody is a specialized melodic contour rather than ordinary read speech.
This report describes a system, Vāgdhenu, that solves the practical version of this problem well enough to ship two large deployments, and it documents the design decisions, the dead ends, and the one negative result that turned out to be the most useful finding. We do not claim a new model. We claim an honest account of what it takes to build a faithful Sanskrit chant pipeline on top of current open backbones, what works, and what is architecturally out of reach.
Our framing is that of an experience report. The evidence we offer is the comparative lineage across architecture families, a reproducible production system, two shipped artifacts at real scale, and a public release of code, weights, d […]
- lordleft
Forgive my ignorance; I'm not a Sanskrit speaker (Malayalam is my first language) -- but I love the sound of Sanskrit.
I have lots of questions: what's the use case for this? Primarily religious / liturgical? It also seems that the fact that this tool works means that any arbitrary Sanskrit sentence can be translated into a chant by some sort of procedure (dare I say algorithm)? I'm terribly curious and fascinated by this!
- sebmellen
This must be insanely difficult to get right. I was not expecting such an impressive result from a site that looks so vibe coded! The look is underselling how good this is.
- ks2048
It was based on an existing TTS, IndicT5. I wonder how different is “Sanskrit Chanting” to languages it could already do, like Hindi. Is it largely the glyph—to-phoneme that needs relearning? Or pitch control? Or more?
- ultrasounder
Very nice implementation. I tried developing my own to practice Santhai(repeat thrice) to learn Upnishads and this tool would be at the center of my workflow. A locally installable version would be even great! Kudos. Dhanyosmi :-)