Disclosed, according to various embodiments of the present invention, are a method and a device, the device comprising: a plurality of cameras arranged at mutually different locations; a plurality of microphones arranged at mutually different locations; a memory; and a processor operatively connected to at least one among the plurality of cameras, the plurality of microphones, and the memory. The processor is set to: by using at least one of the plurality of cameras, determine whether at least one among a user wearing an electronic device or a partner engaged in conversation with the user is speaking or not; on the basis of the result of determining, set the directivity of at least one of the plurality of microphones; on the basis of the directivity that has been set, acquire audio from at least one of the plurality of microphones; acquire, from at least one of the plurality of cameras, an image containing the shape of the mouth of the user or the partner; and on the basis of the acquired audio and image, process speech of the speakers by using mutually different methods. Various embodiments are possible.