A small-footprint hands-free speakerphone apparatus includes a loudspeaker, at least one main microphone, at least one reference microphone, and openings for receiving sound. The loudspeaker is disposed within a speaker chamber. Each microphone is mounted in a microphone boot formed by an acoustic opaque resilient material. Each main microphone may be omni-directional or directional. Each reference microphone is omni-directional. An opening is formed in front of each microphone for receiving sound for the microphone. An opening is formed in back of each directional main microphone for receiving sound and forming directivity for the microphone. An opening is formed in the microphone boot for each reference microphone to receive sound from the loudspeaker. An opening may also be formed in the loudspeaker chamber for passing sound to the reference microphone. The apparatus may include one or more interface units to provide interface to external devices.