DOUBLE-TALK DETECTION USING UPSAMPLING Russian patent published in 2024 - IPC H04M9/08 H04B3/23 

FIELD: physics.

SUBSTANCE: invention relates to processing audio data and, in particular, to detection of double talk. Disclosed method of processing audio data, which enables to detect simultaneous conversation, involves using upsampling. Audio signals received from the far end are upsampled before output by the loudspeaker at the near end. Near-end microphone captures audio at an upsampling rate, and the audio output signal is detected by the loudspeaker due to the absence of energy in the upsampled frequency bands. Double talk detector uses this information to generate a signal for echo cancellation in audio at the far end from the captured audio signal, which is transmitted to the far end. Also disclosed is a persistent machine-readable medium and a device for implementing the method of processing audio data.

EFFECT: high quality of communication and improved user experience.

13 cl, 7 dwg