Towards Robust Transcription: Exploring Noise Injection Strategies for Training Data Augmentation

TL;DR

This paper investigates how adding white noise at different SNR levels affects the performance of state-of-the-art Automatic Piano Transcription models, aiming to improve robustness across noisy environments.

Contribution

It explores noise injection strategies for data augmentation in training APT models, a relatively underexplored area in improving transcription robustness.

Findings

Noise augmentation improves model robustness in noisy conditions

Performance varies with different SNR levels during training

Training on noise-augmented data maintains accuracy across acoustic environments

Abstract

Recent advancements in Automatic Piano Transcription (APT) have significantly improved system performance, but the impact of noisy environments on the system performance remains largely unexplored. This study investigates the impact of white noise at various Signal-to-Noise Ratio (SNR) levels on state-of-the-art APT models and evaluates the performance of the Onsets and Frames model when trained on noise-augmented data. We hope this research provides valuable insights as preliminary work toward developing transcription models that maintain consistent performance across a range of acoustic conditions.