Methods and systems are provided for detecting artifacts in an electronic signal. In an embodiment, a method is provided comprising: connecting a first input of an electronic device to a first signal line of a signal processing device, such as an amplification device; connecting a second input of the electronic device to a second signal line of the signal processing device, the second signal line being downstream from the first signal line; establishing, based on an observed behavior of a first signal on the first signal line, an expected behavior of a second signal on the second signal line; and determining whether a difference exists between the expected behavior of the second signal and an observed behavior of the second signal. If a difference is detected, the expected behavior of a second signal and the observed behavior of the second signal may be recorded for later analysis.

Methods and systems are provided for detecting artifacts in an electronic signal. In an embodiment, a method is provided comprising: connecting a first input of an electronic device to a first signal line of a signal processing device, such as an amplification device; connecting a second input of the electronic device to a second signal line of the signal processing device, the second signal line being downstream from the first signal line; establishing, based on an observed behavior of a first signal on the first signal line, an expected behavior of a second signal on the second signal line; and determining whether a difference exists between the expected behavior of the second signal and an observed behavior of the second signal. If a difference is detected, the expected behavior of a second signal and the observed behavior of the second signal may be recorded for later analysis.

A system and method for reverberation reduction is disclosed. A first Deep Neural Network (DNN) produces a first estimate of a target direct-path signal from a mixture of acoustic signals that include the target direct-path signal and a reverberation of the target direct-path signal. A filter modeling a room impulse response (RIR) for the first estimate is estimated. The filter when applied to the first estimate of the target direct-path signal generates a result closest to a residual between the mixture of the acoustic signals and the first estimate of the target direct-path signal according to a distance function. A mixture with reduced reverberation of the target direct-path signal is obtained by removing the result of applying the filter to the first estimate of the target direct-path signal from the received mixture. A second DNN produces a second estimate of the target direct-path signal from the mixture with reduced reverberation.

A system and method for reverberation reduction is disclosed. A first Deep Neural Network (DNN) produces a first estimate of a target direct-path signal from a mixture of acoustic signals that include the target direct-path signal and a reverberation of the target direct-path signal. A filter modeling a room impulse response (RIR) for the first estimate is estimated. The filter when applied to the first estimate of the target direct-path signal generates a result closest to a residual between the mixture of the acoustic signals and the first estimate of the target direct-path signal according to a distance function. A mixture with reduced reverberation of the target direct-path signal is obtained by removing the result of applying the filter to the first estimate of the target direct-path signal from the received mixture. A second DNN produces a second estimate of the target direct-path signal from the mixture with reduced reverberation.

An electronic device includes one or more microphones that generate audio signals and a wind noise detection subsystem. The electronic device may also include a wind noise reduction subsystem. The wind noise detection subsystem applies multiple wind noise detection techniques to the set of audio signals to generate corresponding indications of whether wind noise is present. The wind noise detection subsystem determines whether wind noise is present based on the indications generated by each detection technique and generates an overall indication of whether wind noise is present. The wind noise reduction subsystem applies one or more wind noise reduction techniques to the audio signal if wind noise is detected. The wind noise detection and reduction techniques may work in multiple domains (e.g., the time, spatial, and frequency domains).

A computing device with a microphone system is disclosed. The computing device includes a microphone system with an environment microphone and a noise microphone. The environment microphone picks up an environment microphone signal which includes (1) a desired signal component based on desired sound and (2) a noise component based on noise from a noise source. The noise microphone picks up a noise microphone signal based on the noise, and is configured such that contributions to the noise microphone signal from the desired sound, if present, are attenuated relative to the environment microphone. A controller receives and processes time samples from the noise microphone signal to yield a noise estimation of the noise component. The estimation is subtracted from the environment microphone signal to yield and end-user output.

A method of speech detection and speech enhancement in a speech detection and speech enhancement unit of Multipoint Conferencing Node (MCN) and a method of training the same. The method comprising receiving input audio segments, and determining an acoustic environment based on input audio auxiliary information, extracting T-F-domain features from the received input audio segments, determining if each of the received input audio segments is speech by inputting the T-F domain features into a speech detection classifier trained for the determined acoustic environment, determining, when one of the received input audio segments is speech, if the received audio segment is noisy speech by inputting the T-F domain features into a noise classifier using a statistical generative model representing the probability distributions of the T-F domain features of noisy speech trained for the determined acoustic environment, and applying a noise reduction mask on the received input audio segments according to the determination of the received audio segment is noisy speech

A method of speech detection and speech enhancement in a speech detection and speech enhancement unit of Multipoint Conferencing Node (MCN) and a method of training the same. The method comprising receiving input audio segments, and determining an acoustic environment based on input audio auxiliary information, extracting T-F-domain features from the received input audio segments, determining if each of the received input audio segments is speech by inputting the T-F domain features into a speech detection classifier trained for the determined acoustic environment, determining, when one of the received input audio segments is speech, if the received audio segment is noisy speech by inputting the T-F domain features into a noise classifier using a statistical generative model representing the probability distributions of the T-F domain features of noisy speech trained for the determined acoustic environment, and applying a noise reduction mask on the received input audio segments according to the determination of the received audio segment is noisy speech

Systems, methods, and computer program product embodiments are disclosed for removing any fixed frequency interfering signal from an input signal without introducing artifacts that are not part of the original signal of interest. An embodiment operates by using a virtual buffer with a length that matches a length of one cycle of an interfering signal. The embodiment extracts the interfering signal into the virtual buffer. For a sample in the next cycle of the interfering signal that corresponds to a virtual memory location for the virtual buffer, the embodiment can update one or more physical memory locations of the virtual buffer that are in the vicinity of the virtual memory location. This use of virtual buffer can remove any interfering signal without creating the artifacts associated with conventional notch filters.

The present document describes a method (600) for encoding a multi-channel input signal (101) which comprises N different channels. The method (600) comprises, for a current frame of a sequence of frames, determining (601) whether the current frame is an active frame or an inactive frame using a signal and/or a voice activity detector, and determining (602) a downmix signal (103) based on the multi-channel input signal (101), wherein the downmix signal (103) comprises N channels or less. In addition, the method (600) comprises determining (603) upmixing metadata (105) comprising a set of parameters for generating, based on the downmix signal (103), a reconstructed multi-channel signal (111) comprising N channels, wherein the upmixing metadata (105) is determined in dependence of whether the current frame is an active frame or an inactive frame. The method (600) further comprises encoding (604) the upmixing metadata (105) into a bitstream.