The present disclosure relates to a sound image localization apparatus that includes head related transfer function (HRTF) filters that receive a monaural audio signal and generate a first channel output signal using a first head related transfer function describing a change in sound at right/left eardrum caused when a sound occurs from a first direction; low-pass filters that cut high frequency components of the monaural audio signal and pass low frequency components of the monaural signal; delay circuits that delay an output of the low-pass filters by first/second delay amount; and mixers that mix an output of the HRTF filters and an output of the delay circuits and output an audio signal.