Paper ID | AUD-16.6 | ||
Paper Title | EFFICIENT TRAINING DATA GENERATION FOR PHASE-BASED DOA ESTIMATION | ||
Authors | Fabian Hübner, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Germany; Wolfgang Mack, Emanuël Habets, AudioLabs Erlangen, Germany | ||
Session | AUD-16: Modeling, Analysis and Synthesis of Acoustic Environments 2: Spatial Audio | ||
Location | Gather.Town | ||
Session Time: | Wednesday, 09 June, 16:30 - 17:15 | ||
Presentation Time: | Wednesday, 09 June, 16:30 - 17:15 | ||
Presentation | Poster | ||
Topic | Audio and Acoustic Signal Processing: [AUD-ASAP] Acoustic Sensor Array Processing | ||
IEEE Xplore Open Preview | Click here to view in IEEE Xplore | ||
Abstract | Deep learning (DL) based direction of arrival (DOA) estimation is an active research topic and currently represents the state-of-the-art. Usually, DL-based DOA estimators are trained with recorded data or computationally expensive generated data. Both data types require significant storage and excessive time to, respectively, record or generate. We propose a low complexity online data generation method to train DL models with a phase-based feature input. The data generation method models the phases of the microphone signals in the frequency domain by employing a deterministic model for the direct path and a statistical model for the late reverberation of the room transfer function. By an evaluation using data from measured room impulse responses, we demonstrate that a model trained with the proposed training data generation method performs comparably to models trained with data generated based on the source-image method. |