AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Voicebox: Text-Guided Multilingual Universal Speech Generation at Scale

Matthew Le, Apoorv Vyas, Bowen Shi, Brian Karrer, Leda Sari, Rashel Moritz, Mary Williamson, Vimal Manohar, Yossi Adi, Jay Mahadeokar, Wei-Ning Hsu

NeurIPS 2023 · Image generation › Autoregressive and latent generation · approach: Neural text-to-speech synthesis · 45 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use optimal-transport path instead of diffusion path (flow matching) (ablation)Voicebox audio model · Zero-shot TTS 50K updatesWER: 76.0 -> 2.5Improved
use flow matching OT rather than diffusion path at low NFE (ablation)Voicebox audio model · Zero-shot TTS NFE=4WER: 11.5 -> 2.4Improved
use flow matching instead of score matching with diffusion path (ablation)Voicebox audio model · Zero-shot TTS 150K updatesWER: 5.1 -> 2.1Improved
use flow matching instead of regression objective (ablation)Voicebox audio model · Zero-shot TTS cross-sentenceSIM-r: 0.520 -> 0.597Improved
use flow matching instead of regression objective (ablation)Voicebox audio model · Diverse samplingFSD: 278.8 -> 242.5Improved
use flow matching instead of regression objective (ablation)Voicebox audio model · Zero-shot TTS cross-sentenceWER: 2.0 -> 2.1No clear effect
compute loss only on masked frames (ablation)Voicebox audio model · Zero-shot TTS cross-sentenceSIM-r: 0.528 -> 0.597Improved
compute loss only on masked frames (ablation)Voicebox audio model · Zero-shot TTS cross-sentenceWER: 2.0 -> 2.1No clear effect
scale training data from 60 to 60,000 hours (ablation)Voicebox · Zero-shot TTSSIM-r: 0.151 -> 0.645Improved
scale training data from 6,000 to 60,000 hours (ablation)Voicebox · Diverse samplingFSD: 195.52 -> 214.38Hurt
use flow-matching duration model instead of regression duration model (ablation)Voicebox VB-En · ASR trained on synthetic data, test-cleanWER: 7.1 -> 3.1Improved
use flow-matching duration model instead of regression (ablation)Voicebox VB-En · Diverse speech generationWER: 3.1 -> 5.6Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy