First page Back Continue Last page Overview Text


Once I have synchronized audio streams from the people in the study – I look for proximity – a first order check uses AP info – if they have the same AP info and are speaking, I do a waveform correlation. At an instance where one speaker is speaking loudly, I look for that speech’s unique waveform in the audio of the other people nearby. It’s possible to detect the person who is being addressed simply because the energy in the wave form – as well as get a better idea of the physical distance between speakers.