Multitalker optimised beamforming system and method

A method of processing a series of microphone inputs of an audio conference, each microphone input being captured by a respective one among an array of microphones, the method including the steps of:

• (a) conducting a spatial analysis and feature extraction of the series of microphone inputs of the audio conference based on current audio activity to obtain information about the approximate relative location of currently active sound objects relative to the array of microphones;
• (b) aggregating the information about the approximate relative location of currently active sound objects relative to the array of microphones over time to obtain information about recent sound objects in terms of their optimal microphone beam characteristics and their degree of direct to reverb ratio or coherence;
• and
• (c) adjusting, based on the information about the recent sound objects, an optimal microphone beam for an active sound object in such a way as to reduce a suppression of sound capture of currently inactive sound objects in case that they become active,
• wherein the optimal microphone beam for the active sound object is adjusted so as to avoid a null or suppression above a given threshold for currently inactive sound objects that are likely to become active.

An apparatus for the selective processing of a series of microphone inputs of an audio conference, the apparatus including:

• an array of microphones, each microphone capturing a respective microphone input;
• a beamformer interconnected to the array of microphones for producing a beamformed audio output of the microphone inputs;
• a location determination unit for estimating the location of sound objects detected by the series of microphone inputs, wherein estimating the location of sound objects involves conducting a spatial analysis and feature extraction of the microphone inputs of the audio conference based on current audio activity to obtain information about the approximate relative location of currently active sound objects relative to the array of microphones;
• and
• a beamformer actuation unit, interconnected to said location determination unit and to said beamformer, for actuating the beamformer,
• wherein the beamformer actuation unit is adapted to:
• aggregate the information about the approximate relative location of currently active sound objects relative to the array of microphones over time to obtain information about the approximate relative location of recent sound objects relative to the array of microphones;
• and
• utilise the approximate relative location of the recent sound objects relative to the array of microphones to determine whether beam forming is to be utilised to enhance the audio reception from recent sound objects when these recent sound objects are detected;
• and
• determining that beam forming is to be utilized to enhance the audio reception from the recent sound objects when recent sound objects of a predetermined estimated distance range from the array of microphones are detected.

An apparatus for the selective processing of a series of microphone inputs of an audio conference, the apparatus including:

• an array of microphones, each microphone capturing a respective microphone input;
• a beamformer interconnected to the array of microphones for producing a beamformed audio output of the microphone inputs;
• a location determination unit for estimating the location of sound objects detected by the series of microphone inputs, wherein estimating the location of sound objects involves conducting a spatial analysis and feature extraction of the microphone inputs of the audio conference based on current audio activity to obtain information about the approximate relative location of currently active sound objects relative to the array of microphones;
• and
• a beamformer actuation unit, interconnected to said location determination unit and to said beamformer, for actuating the beamformer,
• wherein the beamformer actuation unit is adapted to:
• aggregate the information about the approximate relative location of currently active sound objects relative to the array of microphones over time to obtain information about recent sound objects in terms of their optimal microphone beam characteristics and their degree of direct to reverb ratio or coherence;
• adjust, based on the information about the recent sound objects, an optimal microphone beam for an active sound object in such a way as to reduce a suppression of sound capture of currently inactive sound objects in case that they become active;
• and
• adjusting the optimal microphone beam for the active sound object so as to avoid a null or suppression above a given threshold for currently inactive sound objects that are likely to become active.