It is based on expf one by converting each line with the formula: new_val = (float) log10 (exp ((double) old_val))