fix RTP dump pipeline (91d82fe7) · Commits · IVAS Codec Public Collaboration / IVAS Codec

tests/conftest.py

+10 −8

Original line number	Diff line number	Diff line
		@@ -1358,7 +1358,7 @@ class PostRendFrontend:
		quiet_mode: Optional[bool] = True,
		trajectory_file: Optional[Path] = None,
		prbfi_file: Optional[Path] = None,
		framing: Optional[int] = None,
		render_frame_size: Optional[int] = None,
		add_option_list: Optional[list] = None,
		run_dir: Optional[Path] = None,
		) -> None:
		@@ -1368,6 +1368,14 @@ class PostRendFrontend:
		if quiet_mode:
		command.extend(["-q"])

		if render_frame_size:
		if render_frame_size not in [5, 10, 20]:
		raise ValueError(f"Incorrect render frame size specified {render_frame_size}!\n")

		command.extend(["-fr", str(render_frame_size)])
		else:
		command.extend(["-fr", self.fr])

		if input_file_format == "BINAURAL_SPLIT_PCM":
		# check if input sampling rate has been provided
		if not input_sampling_rate:
		@@ -1395,12 +1403,6 @@ class PostRendFrontend:

		command.extend(["-T", str(trajectory_file)])

		if framing:
		if framing not in [5, 10, 20]:
		raise ValueError(f"Incorrect framing size specified {framing}!\n")

		command.extend(["-fr", str(framing)])

		if add_option_list is not None:
		command.extend(add_option_list)

		@@ -1413,7 +1415,7 @@ class PostRendFrontend:
		"-o",
		str(output_file),
		"-fs",
		"48", # !!!!! should be eventually removed from the mandatory cmd-line parameters
		"48", # !!!!! it's a mandatory argument but should be eventually removed
		]

		cmd_str = textwrap.indent(" ".join(command), prefix="\t")

tests/rtp/test_rtp.py

+74 −22

Original line number	Diff line number	Diff line
		@@ -181,7 +181,10 @@ def test_rtp_bitstream_ivas_dtx(
		@pytest.mark.parametrize("bitrate", [48000])
		@pytest.mark.parametrize("bandwidth", ["FB"])
		@pytest.mark.parametrize("format", ["SBA"])
		@pytest.mark.parametrize("framesPerPacket", [1])
		@pytest.mark.parametrize("framesPerPacket", [8])
		@pytest.mark.parametrize("splitRenderingCodec", ["LCLD", "LC3PLUS"])
		@pytest.mark.parametrize("dof", [0, 1])
		@pytest.mark.parametrize("codecFrameSize", [5, 10, 20])
		@pytest.mark.parametrize("splitRendering_format", ["BINAURAL_SPLIT_CODED"])
		def test_rtp_bitstream_isar(
		test_info,
		@@ -193,6 +196,9 @@ def test_rtp_bitstream_isar(
		dut_decoder_frontend: DecoderFrontend,
		splitRendering_format: str,
		dut_postrend_frontend: PostRendFrontend,
		splitRenderingCodec: str,
		dof: int,
		codecFrameSize: int,
		):
		run_rtp_bitstream_tests(
		CODECS.IVAS,
		@@ -206,6 +212,9 @@ def test_rtp_bitstream_isar(
		dut_decoder_frontend,
		splitRendering_format,
		dut_postrend_frontend,
		splitRenderingCodec,
		dof,
		codecFrameSize,
		)

		def generateRequests(startTs: int, endTs: int) -> dict:
		@@ -596,6 +605,9 @@ class TVARGS:
		self.postRendHeadRotFile = (
		Path(ROOT_DIR).joinpath("scripts/trajectories/rotate_euler_quaternion_30s.csv").absolute()
		)
		# self.renderConfigFile = (
		# Path(ROOT_DIR).joinpath("tests/split_rendering/renderer_configs/framing/framing_lcld_0dof_20ms.txt").absolute()
		# )


		def add(self, fmt: str, inputFile: str, args: list[str] = []):
		@@ -618,6 +630,11 @@ class TVARGS:
		]
		return args

		def renderConfigFile(self, sr_codec: str, dof: int, frame_size: int):
		return (
		Path(ROOT_DIR).joinpath(f"tests/split_rendering/renderer_configs/framing/framing_{sr_codec}_{dof}dof_{frame_size}ms.txt").absolute()
		)

		def piFiles(self) -> tuple[Path]:
		return (self.sceneFile, self.deviceFile)

		@@ -634,6 +651,9 @@ def run_rtp_bitstream_tests(
		dut_decoder_frontend: DecoderFrontend,
		splitRendering_format: Optional[str] = None,
		dut_postrend_frontend: Optional[PostRendFrontend] = None,
		splitRenderingCodec: Optional[str] = None,
		dof: Optional[int] = None,
		codecFrameSize: Optional[int] = None,
		):
		tvArgs = TVARGS()
		tvArgs.add("MONO", "stv48n.wav")
		@@ -652,6 +672,16 @@ def run_rtp_bitstream_tests(
		):
		pytest.skip()

		if dof == 0:
		dut_decoder_frontend.fr = codecFrameSize
		else:
		dut_decoder_frontend.fr = 20

		renderConfigFile = tvArgs.renderConfigFile(splitRenderingCodec, dof, codecFrameSize)
		if not renderConfigFile.exists():
		# skip, if the corresponding render config file does not exist
		pytest.skip()

		print(
		"Test: dut_encoder_frontend={}, dtx={}, codec:={}, bitrate={}, bandwidth={}, caMode={}, format={},".format(
		dut_encoder_frontend._path, dtx, codec, bitrate, bandwidth, caMode, format
		@@ -678,6 +708,11 @@ def run_rtp_bitstream_tests(
		.joinpath(f"input-{codec}-{bitrate}-{caMode}-{format}-{dtx}.rtpdump")
		.absolute()
		)
		rtpdumpSplitOut = (
		Path(tmp_dir)
		.joinpath(f"split-{codec}-{bitrate}-{caMode}-{format}-{dtx}.rtpdump")
		.absolute()
		)
		splitOut = (
		Path(tmp_dir)
		.joinpath(f"split-{codec}-{bitrate}-{caMode}-{format}-{dtx}.cod")
		@@ -700,6 +735,7 @@ def run_rtp_bitstream_tests(
		)

		# Run WITHOUT rtpdump first to generate reference bitstream
		# ./IVAS_cod -sba +1 -dtx -rtpdump 8 48000 48 scripts/testv/WYZX_Harvard_Sentences_1-13s_48khz.wav sba.rtp
		dut_encoder_frontend.run(
		bitrate=bitrate,
		input_sampling_rate=48,
		@@ -722,18 +758,21 @@ def run_rtp_bitstream_tests(
		else:
		outMode = "STEREO" if format == "STEREO" else "BINAURAL"
		generatedPIData = generatePiData(0, 16000)
		generatedRequests = generateRequests(0, 1600)
		else:
		outMode = ""
		generatedPIData = dict()
		generatedRequests = dict()

		packer.packG192File(
		g192File=g192Out,
		rtpDumpOut=rtpdumpIn,
		piData=generatedPIData,
		requestsData=generateRequests(0, 1600),
		requestsData=generatedRequests,
		)

		if not split_rendering:
		# run with G.192 bitstream
		dut_decoder_frontend.run(
		output_config=outMode,
		output_sampling_rate=48,
		@@ -742,50 +781,63 @@ def run_rtp_bitstream_tests(
		add_option_list=[],
		)

		# run with RTP dump
		dut_decoder_frontend.run(
		output_config=outMode,
		output_sampling_rate=48,
		input_bitstream_path=rtpdumpIn,
		output_path=pcmOut,
		add_option_list=["-VOIP_HF_ONLY=1", "-PiDataFile", str(piDataOutJson)],
		)

		else:

		# run with G.192 bitstream
		# ./IVAS_dec -fr 20 -no_delay_cmp -render_config tests/split_rendering/renderer_configs/framing/framing_lcld_0dof_20ms.txt BINAURAL_SPLIT_CODED 48 sba.g192 outSR_0dof_20ms.cod
		dut_decoder_frontend.run(
		output_config=outMode,
		output_sampling_rate=48,
		input_bitstream_path=g192Out,
		output_path=splitOut,
		add_option_list=["-T", str(tvArgs.preRendHeadRotFile)],
		add_option_list=["-no_delay_cmp", "-render_config", str(renderConfigFile)],
		)

		# ./ISAR_post_rend -fr 20 -no_delay_cmp -fs 48 -i outSR_0dof_20ms.cod -if BINAURAL_SPLIT_CODED -o output_SR_0dof_20ms.g192.wav -T scripts/trajectories/rotate_euler_quaternion_30s.csv
		dut_postrend_frontend.run(
		input_file=splitOut,
		input_file_format=outMode,
		output_file=pcmOutG192,
		add_option_list=["-T", str(tvArgs.postRendHeadRotFile)],
		render_frame_size=dut_decoder_frontend.fr,
		add_option_list=["-no_delay_cmp", "-T", str(tvArgs.postRendHeadRotFile)],
		)

		if not split_rendering:
		dut_decoder_frontend.run(
		output_config=outMode,
		output_sampling_rate=48,
		input_bitstream_path=rtpdumpIn,
		output_path=pcmOut,
		add_option_list=["-VOIP_HF_ONLY=1", "-PiDataFile", str(piDataOutJson)],
		)
		else:
		import pdb
		pdb.set_trace()

		# run with RTP dump
		# ./IVAS_dec -rtpOutSR -VOIP_HF_ONLY=1 -no_delay_cmp -render_config tests/split_rendering/renderer_configs/framing/framing_lcld_0dof_20ms.txt BINAURAL_SPLIT_CODED 48 sba.rtp outSR_0dof_20ms.rtp
		dut_decoder_frontend.run(
		output_config=outMode,
		output_sampling_rate=48,
		input_bitstream_path=rtpdumpIn,
		output_path=splitOut,
		add_option_list=["-VOIP_HF_ONLY=1", "-PiDataFile", str(piDataOutJson), "-T", str(tvArgs.preRendHeadRotFile)],
		output_path=rtpdumpSplitOut,
		add_option_list=["-rtpOutSR", "-VOIP_HF_ONLY=1", "-no_delay_cmp", "-render_config", str(renderConfigFile), "-PiDataFile", str(piDataOutJson)],
		)

		# ./ISAR_post_rend -fr 20 -no_delay_cmp -fs 48 -s outSR_0dof_20ms.rtp.sr.txt -i outSR_0dof_20ms.rtp -if RTPDUMP -o output_SR_0dof_20ms.wav -T scripts/trajectories/rotate_euler_quaternion_30s.csv
		dut_postrend_frontend.run(
		input_file=splitOut,
		input_file_format=outMode,
		input_file=rtpdumpSplitOut,
		input_file_format="RTPDUMP",
		output_file=pcmOut,
		add_option_list=["-T", str(tvArgs.postRendHeadRotFile)],
		render_frame_size=dut_decoder_frontend.fr,
		add_option_list=["-no_delay_cmp", "-s", str(rtpdumpSplitOut.with_name(rtpdumpSplitOut.name + '.sr.txt')) , "-T", str(tvArgs.postRendHeadRotFile)],
		)

		decAudio, fs = readfile(pcmOut)
		g192Audio, Fs = readfile(pcmOutG192)
		decAudio /= 32768.0 # readfile reuturns 16 bit int
		g192Audio /= 32768.0 # readfile reuturns 16 bit int
		if not split_rendering:
		decAudio = decAudio[4 * 960 :]
		assert abs(decAudio.shape[0] - g192Audio.shape[0]) <= (
		4 * 960