SOURCE-LINKED INTELLIGENCE

TUTTI: Toward generalizable audio-to-score transcription via fully synthesized data

arXiv · AI, language, vision and robotics · article · Sep 1, 2026 · UTC

Generalizable Audio-to-Score (A2S) transcription is fundamentally constrained by the severe scarcity of high-quality, real-world paired data. Relying solely on existing human-annotated datasets often restricts the generalization of A2S models, limiting their efficacy primarily to single-instrumentation domains. To break this dependency on scarce real-world data, we introduce TUTTI (Transformer for Unified audio-To-score Transcription trained on Synthetic multi-Instrumentation Data), a pre-training paradigm driven by a purely synthetic, large-scale dataset. Rather than using human-composed scor

Read original source ↗ Open in workspace

recordType

paper

region

Global