AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

StyleStream 2.0: Fast and Controllable Streaming Voice Style Conversion

Louis Liu, Nicholas Lee, Gopala Anumanchipalli

NeurIPS 2026 · Image generation › Autoregressive and latent generation · approach: Neural text-to-speech synthesis · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Presents StyleStream 2.0, a streaming zero-shot voice style conversion system that aims to cut latency and add a text-based interface for style control, building on StyleStream 1.0.

Problem

Voice style conversion (transforming timbre, accent, and emotion of a source utterance to a target voice while preserving linguistic content) needs to be streamable with low latency and controllable via text; prior StyleStream 1.0 had about 1s end-to-end latency on consumer hardware and no text-base

Approach

Extends StyleStream 1.0 toward lower-latency streaming operation and a text-based interface for designing or controlling voice style; the abstract does not specify the mechanism.

Key results

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Add autoregressive (Self Forcing style) fine-tuning (ablation)StyleStream 2.0 streaming Stylizer · Voice style conversion (streaming)WER (%): 19.54 -> 12.03Improved
Add autoregressive fine-tuning (ablation)StyleStream 2.0 streaming Stylizer · Voice style conversion (streaming)S-SIM / E-SIM: S-SIM 0.866 -> 0.866; E-SIM 0.767 -> 0.765No clear effect
Use perceiver-resampler style encoder instead of StyleStream 1.0 TDNN encoder (ablation)StyleStream 2.0 offline · Voice style conversion (offline)A-SIM: 0.633 -> 0.705Improved
Use perceiver-resampler style encoder instead of TDNN encoder (ablation)StyleStream 2.0 offline · Voice style conversion (offline)WER (%): 8.04 -> 7.00Improved
Replace multi-step diffusion Stylizer with few-step pixel MeanFlow Stylizer (4 NFE vs 16)StyleStream 1.0 offline · Voice style conversion (offline)S-SIM: 0.869 -> 0.879Improved
Replace multi-step diffusion Stylizer with pMF plus AR fine-tuningStyleStream 1.0 streaming · Streaming, 300 ms chunk, RTX 4060per-chunk processing time (ms): 559.3 -> 140.7Improved
Replace multi-step diffusion Stylizer with pMF plus AR fine-tuningStyleStream 1.0 streaming · Voice style conversion (streaming)WER (%): 11.41 -> 12.02No clear effect
Increase streaming chunk size from 160 ms to 600 ms (ablation)StyleStream 2.0 · Streaming, RTX 4060WER (%): 14.02 -> 10.87Improved
Use offline instead of streaming inference for voice design (ablation)StyleStream 2.0 · Voice designWER (%): 10.42 -> 6.89Improved
Use StyleStream 2.0 voice editing versus cascaded VevoVoice editing · Voice editingA-pres.: 0.6303 -> 0.5102Hurt
Use StyleStream 2.0 versus Qwen3-TTS cascade for voice designVoice design · Voice designE-acc (%): 36.33 -> 32.20Hurt

Limitations

Experimental setup

Baselines: StyleStream 1.0

Builds on: StyleStream 1.0

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