extend for more than one channel

    len_seg: int,

    thresh_low: float = -200,

    thresh_high: float = 0,

):

) -> np.ndarray:

    """

    Calculate Segmental SNR for test_sig to ref_sig as defined in ISO/IEC 14496-4

    """

    ss = list()

    denom_add = 10**-13 * len_seg

    n = 0

    segment_counter = np.zeros(ref_sig.shape[1])

    for ref_seg, test_seg in zip(

        audioarray.get_framewise(ref_sig, len_seg, zero_pad=True),

        audioarray.get_framewise(test_sig, len_seg, zero_pad=True),

    ):

        nrg_ref = np.sum(ref_seg**2)

        ref_power = 10 * np.log10(nrg_ref / len_seg)

        if ref_power < thresh_low or ref_power > thresh_high:

            continue

        nrg_ref = np.sum(ref_seg**2, axis=0)

        diff_seg = ref_seg - test_seg

        nrg_diff = np.sum(diff_seg**2)

        nrg_diff = np.sum(diff_seg**2, axis=0)

        ss_seg = np.log10(1 + nrg_ref / (denom_add + nrg_diff))

        # only sum up segments that fall inside the thresholds

        ref_power = 10 * np.log10(nrg_ref / len_seg)

        zero_mask = np.logical_or(ref_power < thresh_low, ref_power > thresh_high)

        ss_seg[zero_mask] = 0

        # increase segment counter only for channels that were not zeroed

        segment_counter += np.logical_not(zero_mask)

        ss.append(ss_seg)

        n += 1

    ssnr = 10 * np.log10(10 ** (np.sum(ss) / n) - 1)

    ssnr = 10 * np.log10(10 ** (np.sum(ss, axis=0) / segment_counter) - 1)

    return ssnr

    ssnr = get_ssnr(

        ref_sig, test_sig, len_seg, thresh_low=THRESH_LOW, thresh_high=THRESH_HIGH

    )

    print(ssnr)

    for i, s in enumerate(ssnr, start=1):

        print(f"Channel {i}: {s}")

    return 0

if __name__ == "__main__":

    args = parser.parse_args()

    main(args)

    sys.exit(main(args))