UncorrelatedNoiseSource¶

class acoular.sources.UncorrelatedNoiseSource¶

Bases: SamplesGenerator

Simulate uncorrelated white or pink noise signals at multiple channels.

The UncorrelatedNoiseSource class generates noise signals (e.g., white or pink noise) independently at each channel. It supports a user-defined random seed for reproducibility and adapts the number of channels based on the provided microphone geometry. The output is generated block-by-block through the result() generator.

See also

acoular.signals.SignalGenerator

For defining noise types and properties.

acoular.microphones.MicGeom

For specifying microphone geometries.

Notes

• The type of noise is defined by the signal attribute, which must be an instance of a SignalGenerator-derived class that supports a seed parameter.

• Each channel generates independent noise, with optional pre-defined random seeds via the seed attribute.

• If no seeds are provided, they are generated automatically based on the number of channels and the signal seed.

Examples

To simulate uncorrelated white noise at multiple channels:

>>> from acoular import UncorrelatedNoiseSource, WNoiseGenerator, MicGeom
>>> import numpy as np
>>>
>>> # Define microphone geometry
>>> mic_positions = np.array([[0, 0, 0], [1, 0, 0], [0, 1, 0]]).T  # Three microphones
>>> mics = MicGeom(pos_total=mic_positions)
>>>
>>> # Define white noise generator
>>> noise_gen = WNoiseGenerator(sample_freq=51200, num_samples=1024, rms=1.0, seed=42)
>>>
>>> # Create the noise source
>>> noise_source = UncorrelatedNoiseSource(signal=noise_gen, mics=mics)
>>>
>>> # Generate noise output block-by-block
>>> for block in noise_source.result(num=256):
...     print(block.shape)
(256, 3)
(256, 3)
(256, 3)
(256, 3)

The output blocks contain noise signals for each of the 3 channels. The number of blocks depends on the total number of samples and the block size.

signal = Instance(NoiseGenerator, desc='type of noise')¶

Instance of a NoiseGenerator-derived class. For example: - WNoiseGenerator for white noise. - PNoiseGenerator for pink noise.

seed = CArray(dtype=uint32, desc='random seed values')¶

Array of random seed values for generating uncorrelated noise at each channel. If left empty, seeds will be automatically generated as np.arange(self.num_channels) + signal.seed. The size of the array must match the number of output channels.

num_channels = Delegate('mics', 'num_mics')¶

Number of output channels, automatically determined by the number of microphones defined in the mics attribute. Corresponds to the number of uncorrelated noise signals generated.

mics = Instance(MicGeom, desc='microphone geometry')¶

MicGeom object specifying the positions of microphones. This attribute is used to define the microphone geometry and the number of channels.

start_t = Float(0.0, desc='signal start time')¶

Start time of the generated noise signal in seconds. Determines the time offset for the noise output relative to the start of data acquisition. Default is 0.0.

start = Float(0.0, desc='sample start time')¶

Start time of data acquisition at the microphones in seconds. This value determines when the generated noise begins relative to the acquisition process. Default is 0.0.

num_samples = Delegate('signal')¶

Total number of samples in the noise signal, derived from the signal generator. This value determines the length of the output signal for all channels.

sample_freq = Delegate('signal')¶

Sampling frequency of the generated noise signal in Hz, derived from the signal generator. This value defines the temporal resolution of the noise output.

digest = Property( …¶

A unique identifier for the current state of the source, based on its properties. (read-only)

result(num=128)¶

Generate uncorrelated noise signals at microphones in blocks.

The result() method produces a Python generator that yields blocks of noise signals generated independently for each channel. This method supports customizable block sizes and ensures that the last block may have fewer samples if the total number of samples is not an exact multiple of the block size.

Parameters:

numint, optional

Number of samples per block to be yielded. Default is 128.

Yields:

numpy.ndarray

A 2D array of shape (num, num_channels) containing uncorrelated noise signals. The last block may be shorter if the total number of samples is not a multiple of num.

Raises:

ValueError

If the shape of the seed array does not match the number of channels.

Notes

• Each channel’s noise signal is generated using a unique random seed.

• The type and characteristics of the noise are defined by the signal attribute.