On Speech Quality Estimation of Phase-Aware Single-Channel Speech Enhancement