Delay estimation for performing echo cancellation for co-located devices

A participant computing device, comprising:

• one or more processors;
• and
• one or more non-transitory computer-readable media that store instructions that, when executed by the one or more processors, cause the participant computing device to perform operations, the operations comprising:
• receiving an audio transmission comprising a transmitted audio signal for playback at an audio output device associated with the participant computing device, wherein the participant computing device is one of a plurality of participant computing devices of a participant cohort that are co-located within a same area;
• based on the transmitted audio signal, generating a plurality of matched filters that are configured to predict at least a portion of an audio signal corresponding to audio caused by playback of the transmitted audio signal, wherein each of the plurality of matched filters comprises a plurality of coefficients;
• capturing audio data comprising a captured audio signal with an audio capture device associated with the participant computing device, wherein the captured audio signal corresponds to audio produced by playback of the transmitted audio signal with audio output devices of two or more participant computing devices of the participant cohort;
• identifying a matched filter of the plurality of matched filters that most accurately predicts the at least the portion of the audio signal;
• and
• generating a delay estimate based on a predictive contribution of one or more of the plurality of coefficients of the matched filter.

One or more non-transitory computer-readable media that store instructions that, when executed by one or more processors of a participant computing device, cause the one or more processors to perform operations, the operations comprising:

• receiving an audio transmission comprising a transmitted audio signal for playback at an audio output device associated with the participant computing device, wherein the participant computing device is one of a plurality of participant computing devices of a participant cohort that are co-located within a same area;
• based on the transmitted audio signal, generating a plurality of matched filters that are each configured to predict at least a portion of an audio signal corresponding to audio caused by playback of the transmitted audio signal, wherein each of the plurality of matched filters comprises a plurality of coefficients;
• capturing audio data comprising a captured audio signal with an audio capture device associated with the participant computing device, wherein the captured audio signal corresponds to audio produced by playback of the transmitted audio signal with audio output devices of two or more participant computing devices of the participant cohort;
• identifying a matched filter of the plurality of matched filters that most accurately predicts the at least the portion of the audio signal;
• iteratively analyzing the one or more coefficients of the plurality of coefficients of the matched filter to select a coefficient from the plurality of coefficients that corresponds to a point at which the matched filter predicts the at least the portion of the audio signal;
• generating a delay estimate based on the coefficient;
• and
• performing, based at least in part on the delay estimate, an AEC process to remove the audio produced by the playback of the transmitted audio signal with the audio output devices of the two or more participant computing devices from the audio data.