Let the Music Find You
When using a loudspeaker indoors, the walls and ceilings create a sustained, reverberant sound field that distributes sound evenly and creates an encompassing listening experience. Outdoors, the lack of this sound field causes a significant decline in the loudspeaker’s tonal quality. For this project, we will explore how accurately we can estimate the position of a listener or group of listener(s) relative to an outdoor loudspeaker to improve the listening experience for these listeners. To do this, we will interface a circular, six-microphone array with a Raspberry Pi that pipelines recorded acoustic data as a multi-channel .wav file into MATLAB. From there, the Pi will run our position tracking algorithms that process the audio data to calculate the listener’s position. These algorithms take advantage of the fact that sound’s relatively slow speed causes the sound to pass by the various microphones at different times. The theoretical and measured time delays between microphone recordings provide the basis for our position estimation. This project also builds off the existing SONOS ONE smart-speaker hardware, which utilizes a similarly dimensioned circular microphone array to carry out voice commands. Our system will estimate position of the listener(s) relative to the loudspeaker to hopefully enhance the listening experience for nearby listeners. To account for the outdoor environment, we can estimate the position of a single listener or a single group of listeners to potentially direct the sound in the direction of these listeners and eliminate issues caused by the lack of outdoor reverberation.