Apparatus, methods and computer programs for obtaining spatial metadata

A machine learning model is trained to enable high quality spatial audio metadata to be obtained even from sub-optimal or low-quality microphone arrays. Input data for the machine learning model based on two or more microphone signals is determined (eg. by cross-correlation of delay or frequency dat...

Ausführliche Beschreibung

Gespeichert in:

Bibliographische Detailangaben
Hauptverfasser:	Juha Tapio Vilkamo, Mikko Johannes Honkala
Format:	Patent
Sprache:	eng
Schlagworte:	ACOUSTICS ELECTRIC COMMUNICATION TECHNIQUE ELECTRICITY MUSICAL INSTRUMENTS PHYSICS SPEECH ANALYSIS OR SYNTHESIS SPEECH OR AUDIO CODING OR DECODING SPEECH OR VOICE PROCESSING SPEECH RECOGNITION STEREOPHONIC SYSTEMS
Online-Zugang:	Volltext bestellen
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!

Beschreibung
Zusammenfassung:	A machine learning model is trained to enable high quality spatial audio metadata to be obtained even from sub-optimal or low-quality microphone arrays. Input data for the machine learning model based on two or more microphone signals is determined (eg. by cross-correlation of delay or frequency data in the channels) and processed to obtain spatial metadata (eg. source direction or directionality) which is used in turn to render the signal.