An immersive communication system requires rendering of auditory events including 3-D spatial sound to enable effective collaboration through an interactive sound field network. Rendering and morphing of room reverberation are crucial technology of auditory events modeling in virtual acoustics. We already conformed that a smoothed transfer function (TF) method for an inverse filtering (INF) to be robust against the TF variations. TF-smoothing could be done keeping the TF properties of minimum-phase and all-pass components. It assumes that the distance of the pole/zero location from the unit circle in the z-plane represents the reverberation condition for the TF. This paper describes the comparison with the unsmoothed TF and smoothed TF concerning to the magnitude response and both the minimum-phase and all-pass group delay, which is the delay of the envelopes between the input and output signals and depends on the frequency even for a nondispersive medium because of the reflection waves.