In many conventional coherent acoustical image formation algorithms using weight, delay, and summation operators, the computation complexity is mainly governed by the associated exponential and trigonometric operations which are commonly computed by power series expansion in computers. The reconstructed source distribution can be regarded as a result of accumulation of relative co-linear or random vectors. In this paper, we present the computation simplification by analyzing the phase density distribution of these vectors during the image formation process. Then we are able to reduce the computation complexity by replacing the exponential and trigonometric operators by real and symmetric weighting functions. The computation reduction will not only significantly enhance the potential for high-speed acoustical image reconstruction, but also simplify the filter structure for VLSI hardware implementation. There is no significant resolution degradation due to the algorithm simplification. This technique can be also applied to inverse scattering, spectral estimation, nondestructive evaluation, and beam forming.