Seminarium wydziałowe, 10.04.2024 o godzinie w 13:00 w sali 128

  
Prelegent: mgr inż. Paweł Antoniuk, Szkoła Doktorska (SDPB)

Publikacje osoby – Paweł Antoniuk – Politechnika Białostocka (pb.edu.pl)

Tytuł wystąpienia: Ensemble Width in Binaural Music Recordings with Gradient- Boosted Decision Trees under Simulated Anechoic Conditions

Abstract: In recent years, there has been an observable increase in the

significance of binaural sound, particularly in the context of VR

technology and the expanding availability of publicly accessible

recording collections. This trend has led to a growing demand for

methods that characterize the spatial properties of such sound. Current

studies primarily focus on the identification of individual sound

sources, often limiting their scope to speech signals. This presentation

will showcase our latest research addressing this gap by providing a

novel method for estimating the ensemble width in binaural music

recordings, leveraging gradient-boosted decision trees. Tested under

simulated anechoic conditions on datasets mimicking realistic music

scenarios, our approach demonstrates significant potential for enhancing

spatial analysis in real-life musical contexts.