Abstract: Methods, systems, and apparatus for normalizing audio transmissions from multiple endpoints within a teleconference. A first audio transmission from a first participant of a teleconference can be received for presentation at the teleconference. The first audio transmission can be analyzed to classify one or more audio signatures of the first audio transmission as speech. A difference can be determined between the audio level of the one or more audio signatures and an audio level of second transmissions. Based on the difference, the first audio transmission can be normalized to adjust a gain of the first transmission. The transmission can be output to the teleconference.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for reducing audio noise are disclosed. In one aspect, a method includes the actions of receiving first audio data of a user utterance. The actions further include determining an energy level of second audio data being outputted by the loudspeaker. The actions further include selecting a model from among (i) a first model that is trained using first audio data samples that each encode speech from one speaker and (ii) a second model that is trained using second audio data samples that each encode speech from either one speaker or two speakers. The actions further include providing the first audio data as an input to the selected model. The actions further include receiving processed first audio data. The actions further include outputting the processed first audio data.

Abstract: Method includes receiving, at a server device, from a plurality of input devices, audio data. The audio data of each input device corresponds to a time-related portion of the audio data. The method determines a speech energy level for each input device by providing the time-related audio portion as input to a trained model. For each input device, a statistical value associated with the speech energy level is determined. A strongest input device is identified based on the statistical value. The statistical value associated with the speech energy level of each input device other than the strongest input device is compared to the statistical value of the strongest input device. Depending on the comparison, the method determines whether to update the gain value of an input device to an estimated target gain value based on the statistical value of the speech energy level of the respective input device.

Abstract: Method includes receiving, through a plurality of channels, audio data corresponding to a plurality of frequency ranges; determining, for each channel's frequency ranges, speech audio and/or noise energy level using a model trained by machine learning; determining a speech signal with removed noise for each channel; determining one or more statistical values associated with an energy level of a channel's speech signal with the removed noise; determining a strongest channel that has highest statistical values associated with an energy level of a speech signal; determining that the one or more statistical values associated with the energy level of the strongest channel's speech signal satisfy a threshold condition; comparing statistical values associated with an energy level of a speech signal of each channel with those of the strongest channel; and determining whether to update a gain value for a channel based on the channel's statistical values associated with the energy level.

Abstract: Method includes receiving, through a plurality of channels, audio data corresponding to a plurality of frequency ranges; determining, for each channel's frequency ranges, speech audio and/or noise energy level using a model trained by machine learning; determining a speech signal with removed noise for each channel; determining one or more statistical values associated with an energy level of a channel's speech signal with the removed noise; determining a strongest channel that has highest statistical values associated with an energy level of a speech signal; determining that the one or more statistical values associated with the energy level of the strongest channel's speech signal satisfy a threshold condition; comparing statistical values associated with an energy level of a speech signal of each channel with those of the strongest channel; and determining whether to update a gain value for a channel based on the channel's statistical values associated with the energy level.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for audio channel mixing are disclosed. In one aspect, a method includes the actions of receiving first audio data for a first audio channel. The actions further include transmitting the first audio data. The actions further include, while receiving and transmitting the first audio data, receiving second audio data for a second audio channel; determining a first speech audio energy level of the first audio data and a first noise energy level of the first audio data; determining a second speech audio energy level of the second audio data and a second noise energy level of the second audio data; and determining whether to switch to transmitting the second audio data or continue transmitting the first audio data. The actions further include transmitting the first audio data or the second audio data.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for reducing audio noise are disclosed. In one aspect, a method includes the actions of receiving first audio data of a user utterance. The actions further include determining an energy level of second audio data being outputted by the loudspeaker. The actions further include selecting a model from among (i) a first model that is trained using first audio data samples that each encode speech from one speaker and (ii) a second model that is trained using second audio data samples that each encode speech from either one speaker or two speakers. The actions further include providing the first audio data as an input to the selected model. The actions further include receiving processed first audio data. The actions further include outputting the processed first audio data.

Abstract: A non-transitory computer-readable storage medium may include instructions stored thereon. When executed by at least one processor, the instructions may be configured to cause a computing system to determine that a video system is aiming at a single speaker of a plurality of people, receive audio signals from a plurality of microphones, the received audio signals including audio signals generated by the single speaker, based on determining that the video system is aiming at the single speaker, transmit a monophonic signal, the monophonic signal being based on the received audio signals, determine that the video system is not aiming at the single speaker, and based on the determining that the video system is not aiming at the single speaker, transmit a stereophonic signal, the stereophonic signal being based on the received audio signals.

Abstract: A non-transitory computer-readable storage medium may include instructions stored thereon. When executed by at least one processor, the instructions may be configured to cause a computing system to determine that a video system is aiming at a single speaker of a plurality of people, receive audio signals from a plurality of microphones, the received audio signals including audio signals generated by the single speaker, based on determining that the video system is aiming at the single speaker, transmit a monophonic signal, the monophonic signal being based on the received audio signals, determine that the video system is not aiming at the single speaker, and based on the determining that the video system is not aiming at the single speaker, transmit a stereophonic signal, the stereophonic signal being based on the received audio signals.

Abstract: A non-transitory computer-readable storage medium may include instructions stored thereon. When executed by at least one processor, the instructions may be configured to cause a computing system to determine that a video system is aiming at a single speaker of a plurality of people, receive audio signals from a plurality of microphones, the received audio signals including audio signals generated by the single speaker, based on determining that the video system is aiming at the single speaker, transmit a monophonic signal, the monophonic signal being based on the received audio signals, determine that the video system is not aiming at the single speaker, and based on the determining that the video system is not aiming at the single speaker, transmit a stereophonic signal, the stereophonic signal being based on the received audio signals.

Abstract: A non-transitory computer-readable storage medium may include instructions stored thereon. When executed by at least one processor, the instructions may be configured to cause a computing system to determine that a video system is aiming at a single speaker of a plurality of people, receive audio signals from a plurality of microphones, the received audio signals including audio signals generated by the single speaker, based on determining that the video system is aiming at the single speaker, transmit a monophonic signal, the monophonic signal being based on the received audio signals, determine that the video system is not aiming at the single speaker, and based on the determining that the video system is not aiming at the single speaker, transmit a stereophonic signal, the stereophonic signal being based on the received audio signals.