You don't - they use head-related transfer function filters (the Fourier transform of head-related impulse response) to make it work from a single pair of speakers.
You don't - they use head-related transfer function filters (the Fourier transform of head-related impulse response) to make it work from a single pair of speakers.