pull ISAR VOIP test from the main branch

#!/usr/bin/env python3

"""

   (C) 2022-2025 IVAS codec Public Collaboration with portions copyright Dolby International AB, Ericsson AB,

   Fraunhofer-Gesellschaft zur Foerderung der angewandten Forschung e.V., Huawei Technologies Co. LTD.,

   Koninklijke Philips N.V., Nippon Telegraph and Telephone Corporation, Nokia Technologies Oy, Orange,

   Panasonic Holdings Corporation, Qualcomm Technologies, Inc., VoiceAge Corporation, and other

   contributors to this repository. All Rights Reserved.

   This software is protected by copyright law and by international treaties.

   The IVAS codec Public Collaboration consisting of Dolby International AB, Ericsson AB,

   Fraunhofer-Gesellschaft zur Foerderung der angewandten Forschung e.V., Huawei Technologies Co. LTD.,

   Koninklijke Philips N.V., Nippon Telegraph and Telephone Corporation, Nokia Technologies Oy, Orange,

   Panasonic Holdings Corporation, Qualcomm Technologies, Inc., VoiceAge Corporation, and other

   contributors to this repository retain full ownership rights in their respective contributions in

   the software. This notice grants no license of any kind, including but not limited to patent

   license, nor is any license granted by implication, estoppel or otherwise.

   Contributors are required to enter into the IVAS codec Public Collaboration agreement before making

   contributions.

   This software is provided "AS IS", without any express or implied warranties. The software is in the

   development stage. It is intended exclusively for experts who have experience with such software and

   solely for the purpose of inspection. All implied warranties of non-infringement, merchantability

   and fitness for a particular purpose are hereby disclaimed and excluded.

   Any dispute, controversy or claim arising under or in relation to providing this software shall be

   submitted to and settled by the final, binding jurisdiction of the courts of Munich, Germany in

   accordance with the laws of the Federal Republic of Germany excluding its conflict of law rules and

   the United Nations Convention on Contracts on the International Sales of Goods.

"""

import pytest

from tempfile import TemporaryDirectory

from pathlib import Path

import filecmp

from tests.split_rendering.utils import *

from tests.split_rendering.constants import SCRIPTS_DIR, TESTV_DIR

from tests.test_be_for_jbm_neutral_dly_profile import (

    INPUT_FILES,

    get_options_cod,

)

from pyaudio3dtools import audioarray, audiofile

IN_FORMATS = [

    "MC_5_1",

    "ISM4",

    "FOA",

    "MASA2TC",

]

DELAY_PROFILES = ["dly_error_profile_0.dat", "dly_error_profile_5.dat"]

# Compares PCM output and tracefile from a VoIP BINAURAL_SPLIT_PCM chain with equivalent BINAURAL

# chain to ensure time-scaling and other JBM operations are BE between the two.

@pytest.mark.parametrize("in_format", IN_FORMATS)

@pytest.mark.parametrize("delay_profile", DELAY_PROFILES)

def test_voip_be_splitrend_vs_binaural(

    in_format,

    delay_profile,

    dut_encoder_frontend,

    dut_decoder_frontend,

    ivas_bitrate=128000,

):

    with TemporaryDirectory() as tmp_dir:

        tmp_dir = Path(tmp_dir)

        sampling_rate_khz = 48

        delay_profile_path = SCRIPTS_DIR / "dly_error_profiles" / delay_profile

        delay_profile_id = int(delay_profile[-5])

        # run encoder

        bitstream_file = (tmp_dir / f"{in_format}-dly{delay_profile_id}.192").absolute()

        dtx = False

        wav_in = TESTV_DIR / INPUT_FILES[in_format]

        dut_encoder_frontend.run(

            ivas_bitrate,

            sampling_rate_khz,

            wav_in,

            bitstream_file,

            add_option_list=get_options_cod(in_format, dtx),

            run_dir=tmp_dir,

        )

        def run_decoder(out_format):

            options = []

            # With CLDFB pose correction (default with BINAURAL_SPLIT_PCM), a 20 ms audio frame is

            # rendered with only one head position (first of the 4 per frame). If we want to compare

            # the output from BINAURAL_SPLIT_PCM to output from BINAURAL, the head trajectory must

            # be static.

            head_traj = Path(SCRIPTS_DIR / "trajectories/const000.csv")

            options.extend(["-T", str(head_traj)])

            wav_out = (

                tmp_dir

                / f"{in_format}-{ivas_bitrate}-{out_format}-dly{delay_profile_id}.wav"

            ).absolute()

            trace_out = wav_out.with_suffix(".trace")

            options.extend(["-Tracefile", str(trace_out), "-no_delay_cmp"])

            if out_format == "BINAURAL_SPLIT_PCM":

                isar_md_file = wav_out.with_suffix(".isarmd")

                options.extend(["-om", str(isar_md_file)])

            else:

                isar_md_file = None

            dut_decoder_frontend.run(

                out_format,

                sampling_rate_khz,

                bitstream_file,

                wav_out,

                netsim_profile=delay_profile_path,

                add_option_list=options,

            )

            return wav_out, trace_out, isar_md_file

        wav_out_bin, trace_out_bin, _ = run_decoder("BINAURAL")

        wav_out_sr, trace_out_sr, _ = run_decoder("BINAURAL_SPLIT_PCM")

        # Note regarding delay alignment: both output audio files contain the same decoder delay.

        #

        #  - When outputting to BINAURAL with -no_delay_cmp, decoder delay is present in the audio

        #    output, as expected.

        #

        #  - When outputting to BINAURAL_SPLIT_PCM, decoder delay is never compensated in output

        #    audio (irrespective of the -no_delay_cmp flag). The delay value is saved in the ISAR

        #    metadata file and compensated at the post-rendering stage.

        audio_sr, _ = audiofile.readfile(str(wav_out_sr))

        audio_bin, _ = audiofile.readfile(str(wav_out_bin))

        # Ensure audio and tracefiles are BE

        audio_cmp_result = audioarray.compare(

            audio_bin, audio_sr, fs=sampling_rate_khz * 1000, per_frame=False

        )

        tracefiles_equal = filecmp.cmp(trace_out_bin, trace_out_sr)

        failed = not audio_cmp_result["bitexact"] or not tracefiles_equal

        if failed:

            message = []

            if not audio_cmp_result["bitexact"]:

                message.append(

                    "Difference found between delay-aligned BINAURAL audio and BINAURAL_SPLIT_PCM audio! "

                    f"Max abs diff: {audio_cmp_result['max_abs_diff']}"

                )

            if not tracefiles_equal:

                message.append(

                    "Difference found between BINAURAL tracefile and BINAURAL_SPLIT_PCM tracefile!"

                )

            pytest.fail("; ".join(message))