E2 TTS: Embarrassingly Easy Fully Non-Autoregressive Zero-Shot TTS

TL;DR

E2 TTS is a simple, fully non-autoregressive zero-shot TTS system that achieves human-level naturalness and state-of-the-art speaker similarity without complex components.

Contribution

It introduces a straightforward flow-matching-based mel spectrogram generator for zero-shot TTS, eliminating the need for additional models or complex alignment techniques.

Findings

Achieves state-of-the-art zero-shot TTS performance.

Comparable or surpasses Voicebox and NaturalSpeech 3.

Flexible input representation and inference variants.

Abstract

This paper introduces Embarrassingly Easy Text-to-Speech (E2 TTS), a fully non-autoregressive zero-shot text-to-speech system that offers human-level naturalness and state-of-the-art speaker similarity and intelligibility. In the E2 TTS framework, the text input is converted into a character sequence with filler tokens. The flow-matching-based mel spectrogram generator is then trained based on the audio infilling task. Unlike many previous works, it does not require additional components (e.g., duration model, grapheme-to-phoneme) or complex techniques (e.g., monotonic alignment search). Despite its simplicity, E2 TTS achieves state-of-the-art zero-shot TTS capabilities that are comparable to or surpass previous works, including Voicebox and NaturalSpeech 3. The simplicity of E2 TTS also allows for flexibility in the input representation. We propose several variants of E2 TTS to improve usability during inference. See https://aka.ms/e2tts/ for demo samples.