Videos from the 2025 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA 2025) are now available on the IEEE Signal Processing Society (SPS) Resource Center. Last year's workshop continued its long-standing tradition of exploring the cutting edge of audio and acoustic research. For those who were unable to attend in person, these resources provide an invaluable opportunity to engage with the latest breakthroughs in the field from the comfort of your home or office.

The highlights of the digital collection include three featured keynote presentations that cover the breadth of modern audio signal processing:

Reframing SELD: Learned Localization, Multichannel Processing, and the Beamforming Gap: This talk examines the evolution of Sound Event Localization and Detection (SELD) from channel-independent models to classical spatial features, while highlighting how adopting advanced speech processing techniques—such as learned spatial modeling and neural beamforming—can drive future progress in open-vocabulary detection and complex scene analysis.

Text-Speech tasks as Delayed Stream Modeling: This talk presents a unified "delayed stream modeling" framework that leverages decoder-only Transformers to provide a systematic, efficient, and controllable approach to diverse speech and audio tasks—such as separation, transcription, and translation—replacing task-specific architectures with a shared pre-training methodology.

On the (Co-)evolution of Universal Written, Spoken, and Signed Language Processing:

This talk traces the evolution toward universal, task-independent language models across written, spoken, and signed modalities, highlighting how recent research aims to bridge these traditionally separate fields to serve diverse global language users.

Visit the SPS Resource Center today to access these recordings and the accompanying technical slides.