Back to results
US 12,413,904 B2Granted
Multi-source audio processing systems and methods
AssigneeQsc, LLC
InventorsChristopher Charles Nighman, Gerrit Eimbertus Rosenboom, Alfredo Martin Aguilar, Matthew George Skogmo
Life of this patent
Estimating the term…
Abstract
A conferencing system includes a plurality of microphones and an audio processing system that performs blind source separation operations on audio signals to identify different audio sources. The system processes the separated audio sources to identify or classify the sources and generates an output stream including the source separated content.
Claims20
A conferencing system comprising:
- a plurality of microphones, the plurality of microphones including at least first and second microphones positioned in a conferencing environment, the first microphone configured, responsive to detected sound, to generate at least one first audio signal, the second microphone configured, responsive to detected sound, to generate at least one second audio signal;
- and
- one or more processors executing an audio processing system and operatively coupled to the plurality of microphones, the audio processing system configured to:
- perform one or more blind source separation operations on the first audio signal to generate a first plurality of source-separated audio signals each corresponding to a different audio source;
- perform one or more blind source separation operations on the second audio signal to generate a second plurality of source-separated audio signals each corresponding to a different audio source;
- process a first source-separated signal of the first plurality of source-separated audio signals to identify a first type of audio source corresponding to the first source-separated signal;
- process a second source-separated signal of the second plurality of source-separated audio signals to identify a second type of audio source corresponding to the second source-separated signal,
- wherein at least one of the first or second types is a non-speech source;
- classifying the non-speech source into a semantic category;
- identify, using an event activity detection technique, the non-speech source as an event, wherein the event corresponds to the semantic category;
- in response to the identification of the semantic category and corresponding event, implement a response action by adjusting audio output characteristics or video characteristics of a peripheral communicably coupled to the audio processing system;
- and
- generate an output audio stream including audio content from one or both of the first source-separated signal and the second source-separated signal, the output stream generated at least in part based on one or more of direction of arrival information, signal strength information and the adjusted audio or video characteristics.
10 more elements
A conferencing method comprising:
- with one or more processors operatively coupled to a plurality of microphones positioned in a conferencing environment:
- performing one or more blind source separation operations on at least one first audio signal generated by a first microphone of the plurality of microphones to generate a first plurality of source-separated audio signals each corresponding to a different audio source;
- performing one or more blind source separation operations on at least one second audio signal generated by a second microphone of the plurality of microphones to generate a second plurality of source-separated audio signals each corresponding to a different audio source;
- processing a first source-separated signal of the first plurality of source-separated audio signals to identify a first type of audio source corresponding to the first source-separated signal;
- processing a second source-separated signal of the second plurality of source-separated audio signals to identify a second type of audio source corresponding to the second source-separated signal,
- wherein at least one of the first or second types is a non-speech source;
- classifying the non-speech source into a semantic category;
- identifying, using an event activity detection technique, the non-speech source as an event, wherein the event corresponds to the semantic category;
- in response to the identification of the semantic category and corresponding event, implementing a response action by adjusting audio output characteristics or video characteristics of a peripheral communicably coupled to the audio processing system;
- and
- generating an output audio stream for playback, the output audio stream including audio content from one or both of the first source-separated signal and the second source-separated signal, the generating based at least in part on one or more of direction of arrival information, signal strength information and the adjusted audio or video characteristics.
8 more elements
A non-transitory computer readable storage medium storing instructions that, when executed by a computing system, cause the computing system to perform operations comprising:
- performing one or more blind source separation operations on at least one first audio signal generated by a first microphone of a plurality of microphones to generate a first plurality of source-separated audio signals each corresponding to a different audio source;
- performing one or more blind source separation operations on at least one second audio signal generated by a second microphone of the plurality of microphones to generate a second plurality of source-separated audio signals each corresponding to a different audio source;
- processing a first source-separated signal of the first plurality of source-separated audio signals to identify a first type of audio source corresponding to the first source-separated signal;
- processing a second source-separated signal of the second plurality of source-separated audio signals to identify a second type of audio source corresponding to the second source-separated signal,
- wherein at least one of the first or second types is a non-speech source;
- classifying the non-speech source into a semantic category;
- identifying, using an event activity detection technique, the non-speech source as an event, wherein the event corresponds to the semantic category;
- in response to the identification of the semantic category and corresponding event, implementing a response action by adjusting audio output characteristics or video characteristics of a peripheral communicably coupled to the audio processing system;
- and
- generating an output audio stream for playback, the output audio stream including audio content from one or both of the first source-separated signal and the second source-separated signal, the generating based at least in part on one or more of direction of arrival information, signal strength information and the adjusted audio or video characteristics.
7 more elements
Dependencies are inferred from the claim text.