Skip to content

Library Usage

This section details how to use AudioEnhancer as a Python package in your own codebase.

Facade Pattern

The package exposes the AudioEnhancer class, which serves as a Facade to hide the internal complexity of loaders, exporters, and builder patterns.

Additionally, AudioEnhancer implements the Singleton pattern, meaning you can retrieve the active instance from anywhere in your code using:

from audio_enhancer import AudioEnhancer

enhancer = AudioEnhancer.get_instance()

Loading and Exporting Audio

1. Load Audio

The facade's load_audio method internally uses pydub.AudioSegment.from_file. Because it uses FFmpeg under the hood, it supports decoding almost any audio format:

audio = enhancer.load_audio("path/to/input.mp3")

The method returns a pydub.AudioSegment object representing the raw audio samples and metadata.

2. Export Audio

Save your processed AudioSegment back to disk in the format of your choice:

enhancer.export_audio(audio, "path/to/output.flac", format="flac")

Supported formats include: wav, mp3, ogg, flac, m4a, etc.

Processing Pipelines

To apply processing steps (such as noise reduction and level adjustment) to an audio file, you construct a processing Pipeline.

A Pipeline coordinates one or more AudioProcessingStep strategies, running them sequentially.

Here is a typical flow:

from audio_enhancer import AudioEnhancer
from audio_enhancer.noise_reduction.spectral_gating import SpectralGatingNoiseReducer
from audio_enhancer.normalization.peak import PeakNormalizer

enhancer = AudioEnhancer.get_instance()
audio = enhancer.load_audio("recording.wav")

# 1. Access the builder
builder = enhancer.get_builder()

# 2. Configure steps
builder.add_step(SpectralGatingNoiseReducer())
builder.add_step(PeakNormalizer(target_dbfs=-2.0))

# 3. Build and execute pipeline
pipeline = builder.build()
processed_audio = pipeline.process(audio)

# 4. Save result
enhancer.export_audio(processed_audio, "processed_recording.wav", "wav")