2.0.0: Synthesis

Contents

2.0.0: Synthesis#

The final step in any AudibleLight data generation pipeline is synthesis, which refers to generating any outputs (e.g., audio, visuals, metadata).

Scene.generate#

Calling scene.generate can be used to do the following:

  • render IRs for all microphones;

  • convolve Event audio with IRs for all microphones;

  • combine Event + Ambience audio together WRT the Scene noise floor

  • save spatial audio for all microphones;

  • save DCASE-style metadata for all microphones;

  • save a custom AudibleLight JSON file for the Scene, which can be used to reproduce the Scene at a later date.

[1]:

import os

from audiblelight.core import Scene
from audiblelight import utils

First, we’ll generate a simple scene with one static and one moving Event, + some Ambience

[2]:

scene = Scene(
    duration=60,
    sample_rate=44100,
    backend="rlr",
    backend_kwargs=dict(
        mesh=utils.get_project_root() / "tests/test_resources/meshes/Oyens.glb"
    ),
    fg_path=utils.get_project_root() / "tests/test_resources/soundevents"
)
scene.add_microphone(microphone_type="ambeovr")
scene.add_event(event_type="static")
scene.add_event(event_type="moving")
scene.add_ambience(noise="gaussian")

CreateContext: Context created

Warning: initializing context twice. Will destroy old context and create a new one.

CreateContext: Context created

Warning: initializing context twice. Will destroy old context and create a new one.

CreateContext: Context created

2025-10-07 17:19:46.852 | INFO     | audiblelight.core:add_event:834 - Event added successfully: Static 'Event' with alias 'event000', audio file '/home/huw-cheston/Documents/python_projects/AudibleLight/tests/test_resources/soundevents/musicInstrument/8390.wav' (unloaded, 0 augmentations), 1 emitter(s).
Warning: initializing context twice. Will destroy old context and create a new one.
2025-10-07 17:20:34.473 | INFO     | audiblelight.core:add_event:834 - Event added successfully: Moving 'Event' with alias 'event001', audio file '/home/huw-cheston/Documents/python_projects/AudibleLight/tests/test_resources/soundevents/doorCupboard/35632.wav' (unloaded, 0 augmentations), 6 emitter(s).

CreateContext: Context created

Now, we can call scene.generate to dump audio, CSV files, and JSON files inside a given output directory:

[ ]:

os.makedirs("tmp")
scene.generate(output_dir="tmp")

Note that, to avoid saving any particular outputs, we can pass the following arguments:

[ ]:

scene.generate(
    audio=False,    # don't save audio
    metadata_json=False,    # don't save JSON
    metadata_dcase=False    # don't save CSV
)
Contents