---
title: "Sony CSL's LiveBand Jams With Live Musicians in Real Time on One GPU"
subtitle: "LiveBand generates real-time music accompaniments on consumer hardware by replacing teacher forcing with adversarial sequence-level training, eliminating error drift entirely"
url: "https://alphasignal.ai/news/sony-csl-s-liveband-jams-with-live-musicians-in-real-time-on-one-gpu"
author: "AlphaSignal Newsroom"
author_url: "https://alphasignal.ai/editorial-team"
company: "AlphaSignal"
company_url: "https://alphasignal.ai/company/alphasignal"
published: "2026-06-02T00:00:00.000Z"
updated: "2026-06-10T13:40:49.361Z"
type: "Paper"
topics:
  - "Audio"
  - "Infra"
subtopics:
  - "Music Generation"
  - "Realtime Voice"
  - "Inference Optimization"
upvotes: 2338
source_url: "https://arxiv.org/abs/2606.03803"
comments: 0
truncated: true
---

# Sony CSL's LiveBand Jams With Live Musicians in Real Time on One GPU

> LiveBand generates real-time music accompaniments on consumer hardware by replacing teacher forcing with adversarial sequence-level training, eliminating error drift entirely

Teaching an AI to jam with a live musician is harder than it sounds. The model has to listen, understand what's being played, and respond musically -- all in real time, without ever peeking at what comes next. A new system from Sony CSL Paris called [LiveBand](https://sonycslparis.github.io/liveband-companion) takes a fresh approach to this problem and, for the first time, makes it work reliably on a single consumer GPU without the musical drift that plagued earlier attempts.

## The problem with jamming in real time

Most AI music generation systems are offline: they see the whole song, or at least a large chunk of it, before producing output. Live accompaniment is a fundamentally different challenge. Real-time music generation has been demonstrated in unconditioned or weakly conditioned settings, and offline systems have shown strong results for accompaniment generation, but the specific problem of live, mix-conditioned accompaniment remains substantially more difficult.

The core difficulty is something called **exposure bias** \-- a well-known failure mode in autoregressive models. Standard training uses **teacher forcing**: at each step, the model is fed the ground-truth previous output as context, making training easy and stable. But at inference time, the model must feed its own predictions back as context. If an early prediction is slightly wrong, that error becomes part of the next input context; subsequent predictions are then made from an off-distribution state, which increases the chance of further errors. Repeating this feedback loop causes drift and long-horizon error accumulation. In live music, a slightly misaligned beat can quickly cascade into complete musical asynchrony.

There is also a second, more subtle problem: real-time generation requires the model to produce audio slightly _ahead_ of the live input, to give the hardware time to decode and play it back. Recent work based on causal transformers over discrete neural audio tokens suggests that achieving adherence between the generated accompaniment and the input may require some lookahead into the future mix. LiveBand's central claim is that this is not a fundamental limitation -- it is a training mismatch that can be fixed.

![Diagram comparing teacher forcing with error accumulation versus sequence-level adversarial training with stable output](https://arxiv.org/html/2606.03803v2/x1.png) 

## The fix: ditch teacher forcing, use a discriminator instead

LiveBand replaces next-step predictive supervision with adversarial sequence-level supervision, and eliminates teacher forcing entirely by driving each step only with causal mix conditioning and independent noise. Instead of predicting the next frame given the correct previous frame, the generator receives a fresh Gaussian noise vector at each step alongside the current mix context. It never sees its own past outputs as inputs -- temporal dependencies are carried entirely through the transformer's internal attention state.

_Preview only. Full text requires an AlphaSignal Pro account._

## Sources

- [LiveBand: Live Accompaniment Generation in the Audio Domain (arXiv)](https://arxiv.org/html/2606.03803)
- [LiveBand Abstract (arXiv)](https://arxiv.org/abs/2606.03803)
- [LiveBand Demo Page (Sony CSL Paris)](https://sonycslparis.github.io/liveband-companion)