Towards Improving Speaker Distance Estimation through Generative Impulse Response Augmentation
Anton Ratnarajah, Mehmet Ergezer, Arun Nair, Mrudula Athi
TLDR
This paper significantly improves speaker distance estimation by augmenting sparse datasets with generated room impulse responses, reducing MAE.
Key contributions
- Augments sparse datasets with generated Room Impulse Responses (RIRs) for SDE.
- Utilizes FastRIR, conditioned on speaker/listener locations, for RIR generation.
- Implements a quality filter for generated RIRs and optimizes model hyperparameters.
- Reduces SDE MAE from 1.66m to 0.6m (GWA) and 2.18m to 0.69m (Treble).
Why it matters
This paper tackles the challenge of sparse datasets in speaker distance estimation by leveraging generative data augmentation. It significantly improves estimation accuracy, particularly at medium to long distances, making SDE models more robust and reliable in real-world scenarios.
Original Abstract
The Room Acoustics and Speaker Distance Estimation (SDE) Challenge at ICASSP 2025 explores the effectiveness of augmented room impulse response (RIR) data for improving SDE model performance. This challenge at GenDARA involves generating RIRs to supplement sparse datasets and fine-tuning SDE models with the augmented data. We employ the open-source fast diffuse room impulse response generator (FastRIR) conditioned only on speaker and listener locations. We design a quality filter to ensure generated RIR alignment with challenge RIRs, and hyperparameter optimization is employed for model fine-tuning. Our approach reduces the mean absolute error (MAE) of the five positions from 1.66m to 0.6m for GWA rooms and from 2.18m to 0.69m for Treble rooms, with results demonstrating that the augmentation approach significantly improves estimation accuracy, particularly at medium to long distances.
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.