formatting (e15eb5ad) · Commits · IVAS Codec Public Collaboration / IVAS Processing Scripts

ivas_processing_scripts/audiotools/wrappers/masaAnalyzer.py

+3 −1

Original line number	Diff line number	Diff line
		@@ -73,7 +73,9 @@ def masaAnalyzer(
		binary = find_binary("masaAnalyzer")

		# enforce metadata_out_filename to be a Path object
		if metadata_out_filename is not None and not isinstance(metadata_out_filename, Path):
		if metadata_out_filename is not None and not isinstance(
		metadata_out_filename, Path
		):
		metadata_out_filename = Path(metadata_out_filename)

		if num_tcs not in [1, 2]:

ivas_processing_scripts/audiotools/wrappers/reverb.py

+6 −5

Original line number	Diff line number	Diff line
		@@ -186,7 +186,7 @@ def reverb_stereo(
		y_right = reverb(input, IR_right, align=align)

		# combine into stereo output
		y = audio.fromtype('STEREO')
		y = audio.fromtype("STEREO")
		y.fs = input.fs
		y.audio = np.column_stack([y_left.audio, y_right.audio])

		@@ -252,7 +252,7 @@ def reverb_foa(
		y_z = reverb(input, IR_z, align=align)

		# combine into FOA output
		y = audio.fromtype('FOA')
		y = audio.fromtype("FOA")
		y.fs = input.fs
		y.audio = np.column_stack([y_w.audio, y_x.audio, y_y.audio, y_z.audio])

		@@ -303,7 +303,7 @@ def reverb_hoa2(
		ych.append(reverb(input, IR, align=align))

		# combine into HOA2 output
		y = audio.fromtype('HOA2')
		y = audio.fromtype("HOA2")
		y.fs = input.fs
		y.audio = np.column_stack(
		[
		@@ -321,6 +321,7 @@ def reverb_hoa2(

		return y


		def reverb_hoa3(
		input: Audio,
		hoa3_IR: Audio,
		@@ -365,7 +366,7 @@ def reverb_hoa3(
		ych.append(reverb(input, IR, align=align))

		# combine into HOA3 output
		y = audio.fromtype('HOA3')
		y = audio.fromtype("HOA3")
		y.fs = input.fs
		y.audio = np.column_stack(
		[

ivas_processing_scripts/generation/init.py

+3 −2

Original line number	Diff line number	Diff line
		@@ -31,6 +31,7 @@
		#

		import logging

		import yaml

		from ivas_processing_scripts.constants import (
		@@ -41,10 +42,10 @@ from ivas_processing_scripts.constants import (
		from ivas_processing_scripts.generation import (
		config,
		generate_ambi_items,
		generate_stereo_items,
		generate_ismN_items,
		generate_omasa_items,
		generate_osba_items,
		generate_stereo_items,
		)
		from ivas_processing_scripts.utils import create_dir

ivas_processing_scripts/generation/generate_ambi_items.py

+29 −15

Original line number	Diff line number	Diff line
		@@ -41,8 +41,15 @@ import numpy as np
		from ivas_processing_scripts.audiotools import audio, audioarray, audiofile, convert
		from ivas_processing_scripts.audiotools.convert.objectbased import convert_objectbased
		from ivas_processing_scripts.audiotools.convert.scenebased import convert_scenebased
		from ivas_processing_scripts.audiotools.wrappers.bs1770 import get_loudness, loudness_norm
		from ivas_processing_scripts.audiotools.wrappers.reverb import reverb_foa, reverb_hoa2, reverb_hoa3
		from ivas_processing_scripts.audiotools.wrappers.bs1770 import (
		get_loudness,
		loudness_norm,
		)
		from ivas_processing_scripts.audiotools.wrappers.reverb import (
		reverb_foa,
		reverb_hoa2,
		reverb_hoa3,
		)
		from ivas_processing_scripts.generation import config
		from ivas_processing_scripts.utils import apply_func_parallel

		@@ -201,7 +208,9 @@ def generate_ambi_scene(
		N_inputs = len(np.atleast_1d(scene["input"]))

		# initialize output dirs
		output_filename = Path(scene["output"]).parent / (cfg.use_output_prefix + Path(scene["output"]).name)
		output_filename = Path(scene["output"]).parent / (
		cfg.use_output_prefix + Path(scene["output"]).name
		)

		dir_path = output_filename.parent
		if dir_path and not dir_path.exists():
		@@ -211,7 +220,6 @@ def generate_ambi_scene(
		y = audio.SceneBasedAudio(cfg.format)

		for i in range(N_inputs):

		# parse parameters from the scene description
		source_file = np.atleast_1d(scene["input"])[i]
		IR_file = np.atleast_1d(scene["IR"])[i]
		@@ -239,7 +247,9 @@ def generate_ambi_scene(
		logger.info(f"Convolving {source_file} with {IR_file}")

		# get input filename and IR filename
		input_filename = Path(source_file).parent / (cfg.use_input_prefix + Path(source_file).name)
		input_filename = Path(source_file).parent / (
		cfg.use_input_prefix + Path(source_file).name
		)
		IR_filename = Path(IR_file).parent / (cfg.use_IR_prefix + Path(IR_file).name)

		# read source file
		@@ -280,7 +290,9 @@ def generate_ambi_scene(
		# adjust the signal length (trim from the end or pad with zeros) to align its length with the previous signal(s)
		N_pad = y.audio.shape[0] - x.audio.shape[0]
		if N_pad != 0:
		x.audio = audioarray.trim(x.audio, x.fs, limits=[0, -N_pad], samples=True)
		x.audio = audioarray.trim(
		x.audio, x.fs, limits=[0, -N_pad], samples=True
		)

		# superimpose
		y.audio += x.audio
		@@ -305,7 +317,9 @@ def generate_ambi_scene(
		binaudio = audio.fromtype("BINAURAL")
		binaudio.fs = y.fs
		convert_scenebased(y, binaudio)
		binaural_output_filename = output_filename.with_name(output_filename.stem + "_BINAURAL" + output_filename.suffix)
		binaural_output_filename = output_filename.with_name(
		output_filename.stem + "_BINAURAL" + output_filename.suffix
		)
		audiofile.write(
		binaural_output_filename,
		binaudio.audio,

ivas_processing_scripts/generation/generate_ismN_items.py

+32 −22

Original line number	Diff line number	Diff line
		@@ -33,12 +33,16 @@ import csv
		import logging
		from itertools import groupby, repeat
		from math import floor
		import numpy as np
		from pathlib import Path

		import numpy as np

		from ivas_processing_scripts.audiotools import audio, audioarray, audiofile
		from ivas_processing_scripts.audiotools.convert.objectbased import convert_objectbased
		from ivas_processing_scripts.audiotools.wrappers.bs1770 import get_loudness, loudness_norm
		from ivas_processing_scripts.audiotools.wrappers.bs1770 import (
		get_loudness,
		loudness_norm,
		)
		from ivas_processing_scripts.generation import config
		from ivas_processing_scripts.utils import apply_func_parallel

		@@ -186,7 +190,9 @@ def generate_ismN_scene(

		# initialize output dirs
		ism_format = f"ISM{N_inputs}"
		output_filename = Path(scene["output"]).parent / (cfg.use_output_prefix + Path(scene["output"]).name)
		output_filename = Path(scene["output"]).parent / (
		cfg.use_output_prefix + Path(scene["output"]).name
		)

		dir_path = output_filename.parent
		if dir_path and not dir_path.exists():
		@@ -197,7 +203,6 @@ def generate_ismN_scene(

		# repeat for all source files
		for i in range(N_inputs):

		# parse parameters from the scene description
		source_file = (
		scene["input"][i] if isinstance(scene["input"], list) else scene["input"]
		@@ -236,7 +241,9 @@ def generate_ismN_scene(
		logger.info(f"Encoding {source_file} at position(s) {source_azi},{source_ele}")

		# get input filename
		input_filename = Path(source_file).parent / (cfg.use_input_prefix + Path(source_file).name)
		input_filename = Path(source_file).parent / (
		cfg.use_input_prefix + Path(source_file).name
		)

		# generate ISM metadata .csv filename (should end with .wav..0.csv, .wav.1.csv, ...)
		y.metadata_files.insert(i, str(output_filename.with_suffix(f".{i}.csv")))
		@@ -268,7 +275,9 @@ def generate_ismN_scene(
		# pad ISM signal with zeros to have the same length as the MASA signal
		N_pad = y.audio.shape[0] - x.audio.shape[0]
		if N_pad != 0:
		x.audio = audioarray.trim(x.audio, x.fs, limits=[0, -N_pad], samples=True)
		x.audio = audioarray.trim(
		x.audio, x.fs, limits=[0, -N_pad], samples=True
		)

		# append ISM signal to the ISM object
		y.audio = np.append(y.audio, x.audio, axis=1)
		@@ -309,7 +318,7 @@ def generate_ismN_scene(
		azi = np.arange(
		float(eval(source_azi[0])),
		float(eval(source_azi[0])) + N_frames * float(eval(source_azi[1])),
		float(eval(source_azi[1]))
		float(eval(source_azi[1])),
		)
		else:
		# replicate static azimuth value N_frames times
		@@ -336,7 +345,7 @@ def generate_ismN_scene(
		ele = np.arange(
		float(eval(source_ele[0])),
		np.sign(float(eval(source_ele[1]))) * 90,
		float(eval(source_ele[1]))
		float(eval(source_ele[1])),
		)[:N_frames]

		# repeat the last elevation value, if array is shorter than N_frames
		@@ -381,10 +390,11 @@ def generate_ismN_scene(
		binaudio = audio.fromtype("BINAURAL")
		binaudio.fs = y.fs
		convert_objectbased(y, binaudio)
		binaural_output_filename = output_filename.with_name(output_filename.stem + "_BINAURAL" + output_filename.suffix)
		binaural_output_filename = output_filename.with_name(
		output_filename.stem + "_BINAURAL" + output_filename.suffix
		)
		audiofile.write(
		binaural_output_filename,
		binaudio.audio,
		binaudio.fs,
		)