A sound localization apparatus capable of localizing a sound image at a given position, in a simple configuration is provided with: a first signal processor for convoluting an input audio signal with a first impulse response corresponding to a path from a reference sound source position to the listener's left ear to generate a left-channel audio signal for localization; a second signal processor for convoluting the input audio signal with a second impulse response corresponding to a path from the reference sound source position to the listener's right ear to generate a right-channel audio signal for localization; and a third signal processor for applying a third impulse response so as to localize a sound image obtained by reproducing the audio signals for localization at a position different from the reference sound source position.