Towards Robust Transcription: Exploring Noise Injection Strategies for Training Data Augmentation
This is incremental work for researchers in audio processing, aiming to improve transcription robustness in noisy environments.
The study tackled the problem of noise sensitivity in Automatic Piano Transcription by investigating the impact of white noise at various SNR levels on state-of-the-art models and evaluating performance when trained on noise-augmented data, but no concrete results or numbers were provided.
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.