clarity.enhancer.multiband_compressor package

Submodules

Module contents

class clarity.enhancer.multiband_compressor.Compressor(threshold: float = 0.0, ratio: float = 1.0, attack: float = 15.0, release: float = 100.0, makeup_gain: float = 0.0, knee_width: float = 0.0, sample_rate: float = 44100.0)[source]

Bases: object

Based in the compressor from [1]. Code adapted from JUCE C++ source code. Optimization using IIR filters.

References: [1] Giannoulis, D., Massberg, M., & Reiss, J. D. (2012). Digital dynamic range compressor design - A tutorial and analysis. Journal of the Audio Engineering Society, 60(6), 399-408.

Example: >>> import librosa >>> import matplotlib.pyplot as plt

>>> signal, sr = librosa.load(
...     librosa.ex("brahms"),
...     sr=None,
...     duration=10,
...     mono=False
... )
>>> if signal.ndim == 1:
>>>      signal = signal[np.newaxis, :]

>>> compressor = Compressor(
...    threshold=-30.0,
...    ratio=4.0,
...    attack=10.0,
...    release=100.0,
...    makeup_gain=1.25,
...    sample_rate=sr,
...    knee_width=10.0,
...)

>>> compressed_signal = compressor(signal)
>>> fig, axes = plt.subplots(2, 1)
>>> axes[0].specgram(signal[0], Fs=sr, NFFT=512, noverlap=256)
>>> axes[0].set_title("original signal")
>>> axes[1].specgram(compressed_signal[0], Fs=sr, NFFT=512, noverlap=256)
>>> axes[1].set_title("Compressed signal")
>>> plt.yticks([x for x in range(0, int(sr / 2), 1000)])
>>> plt.tight_layout()
>>> plt.show()
>>> plt.close()

>>> plt.figure()
>>> plt.plot(signal[0])
>>> plt.plot(compressed_signal[0])
>>> plt.show()
class clarity.enhancer.multiband_compressor.Crossover(freq_crossover: list | ndarray | int, sample_rate: float = 44100)[source]

Bases: object

A class to compute crossover filter for two or more crossover frequencies. This is based on [1]

References: [1] D’Appolito, J. A. (1984, October). Active realization of multiway all-pass crossover systems. In Audio Engineering Society Convention 76. Audio Engineering Society.

Example: >>> xover_freqs = np.array( … [250, 500, 1000, 1500, 2000, 3000, 4000, 5000, 6000, 8000] …) * np.sqrt(2) >>> xover = Crossover(xover_freqs) >>> signal = np.random.randn(1000) >>> filtered_signal = xover(signal) >>> xover.plot_filter()

FILTER_ORDER = 4
plot_filter()[source]

Method to plot the frequency response of the filter. This can help to validate the Class is generating the expected filters

xover_component(signal: ndarray, filter_idx: int, len_xover_freqs: int, axis: int = -1) → ndarray[source]

Apply the crossover filter to the signal.

Parameters:

  • signal (np.ndarray) – The input signal.

  • filter_idx (int) – The index of the filter to apply.

  • len_xover_freqs (int) – The number of crossover frequencies.

  • axis (int) – The axis along which to apply the filter. Default is -1. `More information in scipy.signal.lfilter` documentation.

Returns:

The filtered signal.

Return type:

np.ndarray

class clarity.enhancer.multiband_compressor.MultibandCompressor(crossover_frequencies: int | float | list | np.ndarray, sample_rate: float = 44100, compressors_params: dict | None = None)[source]

Bases: object

Multiband Compressor.

set_compressors(attack: list | float = 15.0, release: list | float = 100.0, threshold: list | float = 0.0, ratio: list | float = 1.0, makeup_gain: list | float = 0.0, knee_width: list | float = 0.0) → None[source]

Set the compressors parameters.

Parameters can be a float or a list with the same length as center_frequencies.

Parameters:

  • attack (list | float) – The attack time in milliseconds.

  • release (list | float) – The release time in milliseconds.

  • threshold (list | float) – The threshold in dB.

  • ratio (list | float) – The ratio.

  • gain (list | float) – The make-up gain in dB.

  • knee_width (list | float) – The knee width in dB.

Returns:

The compressors.

Return type:

list[Compressor]

Raises:

  • ValueError – If the parameters are not the same length as

  • crossover frequencies + 1. –