An audio signal processing device receives a plurality of audio signals via a left channel (L) and a right channel (R) so as to produce a composite signal L+R and a difference signal L−R. The composite signal L+R is changed in phase with an all-pass filter, while the difference signal L−R is changed in phase and frequency characteristic with a band-pass filter (e.g. a center frequency of 1 kHz). The band-pass filter has a gently curved frequency characteristic achieving a broad passing band. Additionally, a phase difference of 90 degrees is maintained between the all-pass filter and the band-pass filter over the entire audio frequency range. The composite signal and the difference signal are adjusted in their levels and then mixed together to produce a monaural signal achieving an audio surround effect for widely propagating sound into the surrounding space without degrading sound quality.