# matplotlib.mlab#

Numerical Python functions written for compatibility with MATLAB commands with the same names. Most numerical Python functions can be found in the NumPy and SciPy libraries. What remains here is code for performing spectral computations and kernel density estimations.

## Spectral functions#

cohere

Coherence (normalized cross spectral density)

csd

Cross spectral density using Welch's average periodogram

detrend

Remove the mean or best fit line from an array

psd

Power spectral density using Welch's average periodogram

specgram

Spectrogram (spectrum over segments of time)

complex_spectrum

Return the complex-valued frequency spectrum of a signal

magnitude_spectrum

Return the magnitude of the frequency spectrum of a signal

angle_spectrum

Return the angle (wrapped phase) of the frequency spectrum of a signal

phase_spectrum

Return the phase (unwrapped angle) of the frequency spectrum of a signal

detrend_mean

Remove the mean from a line.

detrend_linear

Remove the best fit line from a line.

detrend_none

Return the original line.

class matplotlib.mlab.GaussianKDE(dataset, bw_method=None)[source]#

Bases: object

Representation of a kernel-density estimate using Gaussian kernels.

Parameters:
datasetarray-like

Datapoints to estimate from. In case of univariate data this is a 1-D array, otherwise a 2D array with shape (# of dims, # of data).

bw_methodstr, scalar or callable, optional

The method used to calculate the estimator bandwidth. This can be 'scott', 'silverman', a scalar constant or a callable. If a scalar, this will be used directly as kde.factor. If a callable, it should take a GaussianKDE instance as only parameter and return a scalar. If None (default), 'scott' is used.

Attributes:
datasetndarray

The dataset passed to the constructor.

dimint

Number of dimensions.

num_dpint

Number of datapoints.

factorfloat

The bandwidth factor, obtained from kde.covariance_factor, with which the covariance matrix is multiplied.

covariancendarray

The covariance matrix of dataset, scaled by the calculated bandwidth (kde.factor).

inv_covndarray

The inverse of covariance.

Methods

 kde.evaluate(points) (ndarray) Evaluate the estimated pdf on a provided set of points. kde(points) (ndarray) Same as kde.evaluate(points)
covariance_factor()[source]#
evaluate(points)[source]#

Evaluate the estimated pdf on a set of points.

Parameters:
points(# of dimensions, # of points)-array

Alternatively, a (# of dimensions,) vector can be passed in and treated as a single point.

Returns:
(# of points,)-array

The values at each point.

Raises:
ValueErrorif the dimensionality of the input points is different

than the dimensionality of the KDE.

scotts_factor()[source]#
silverman_factor()[source]#

Compute the angle of the frequency spectrum (wrapped phase spectrum) of x. Data is padded to a length of pad_to and the windowing function window is applied to the signal.

Parameters:
x1-D array or sequence

Array or sequence containing the data

Fsfloat, default: 2

The sampling frequency (samples per time unit). It is used to calculate the Fourier frequencies, freqs, in cycles per time unit.

windowcallable or ndarray, default: window_hanning

A function or a vector of length NFFT. To create window vectors see window_hanning, window_none, numpy.blackman, numpy.hamming, numpy.bartlett, scipy.signal, scipy.signal.get_window, etc. If a function is passed as the argument, it must take a data segment as an argument and return the windowed version of the segment.

sides{'default', 'onesided', 'twosided'}, optional

Which sides of the spectrum to return. 'default' is one-sided for real data and two-sided for complex data. 'onesided' forces the return of a one-sided spectrum, while 'twosided' forces two-sided.

The number of points to which the data segment is padded when performing the FFT. While not increasing the actual resolution of the spectrum (the minimum distance between resolvable peaks), this can give more points in the plot, allowing for more detail. This corresponds to the n parameter in the call to fft. The default is None, which sets pad_to equal to the length of the input signal (i.e. no padding).

Returns:
spectrum1-D array

The angle of the frequency spectrum (wrapped phase spectrum).

freqs1-D array

The frequencies corresponding to the elements in spectrum.

psd

Returns the power spectral density.

complex_spectrum

Returns the complex-valued frequency spectrum.

magnitude_spectrum

Returns the absolute value of the complex_spectrum.

angle_spectrum

Returns the angle of the complex_spectrum.

phase_spectrum

Returns the phase (unwrapped angle) of the complex_spectrum.

specgram

Can return the complex spectrum of segments within the signal.

matplotlib.mlab.cohere(x, y, NFFT=256, Fs=2, detrend=<function detrend_none>, window=<function window_hanning>, noverlap=0, pad_to=None, sides='default', scale_by_freq=None)[source]#

The coherence between x and y. Coherence is the normalized cross spectral density:

$C_{xy} = \frac{|P_{xy}|^2}{P_{xx}P_{yy}}$
Parameters:
x, y

Array or sequence containing the data

Fsfloat, default: 2

The sampling frequency (samples per time unit). It is used to calculate the Fourier frequencies, freqs, in cycles per time unit.

windowcallable or ndarray, default: window_hanning

A function or a vector of length NFFT. To create window vectors see window_hanning, window_none, numpy.blackman, numpy.hamming, numpy.bartlett, scipy.signal, scipy.signal.get_window, etc. If a function is passed as the argument, it must take a data segment as an argument and return the windowed version of the segment.

sides{'default', 'onesided', 'twosided'}, optional

Which sides of the spectrum to return. 'default' is one-sided for real data and two-sided for complex data. 'onesided' forces the return of a one-sided spectrum, while 'twosided' forces two-sided.

The number of points to which the data segment is padded when performing the FFT. This can be different from NFFT, which specifies the number of data points used. While not increasing the actual resolution of the spectrum (the minimum distance between resolvable peaks), this can give more points in the plot, allowing for more detail. This corresponds to the n parameter in the call to fft. The default is None, which sets pad_to equal to NFFT

NFFTint, default: 256

The number of data points used in each block for the FFT. A power 2 is most efficient. This should NOT be used to get zero padding, or the scaling of the result will be incorrect; use pad_to for this instead.

detrend{'none', 'mean', 'linear'} or callable, default: 'none'

The function applied to each segment before fft-ing, designed to remove the mean or linear trend. Unlike in MATLAB, where the detrend parameter is a vector, in Matplotlib it is a function. The mlab module defines detrend_none, detrend_mean, and detrend_linear, but you can use a custom function as well. You can also use a string to choose one of the functions: 'none' calls detrend_none. 'mean' calls detrend_mean. 'linear' calls detrend_linear.

scale_by_freqbool, default: True

Whether the resulting density values should be scaled by the scaling frequency, which gives density in units of 1/Hz. This allows for integration over the returned frequency values. The default is True for MATLAB compatibility.

noverlapint, default: 0 (no overlap)

The number of points of overlap between segments.

Returns:
Cxy1-D array

The coherence vector.

freqs1-D array

The frequencies for the elements in Cxy.

psd(), csd()

For information about the methods used to compute $$P_{xy}$$, $$P_{xx}$$ and $$P_{yy}$$.

Compute the complex-valued frequency spectrum of x. Data is padded to a length of pad_to and the windowing function window is applied to the signal.

Parameters:
x1-D array or sequence

Array or sequence containing the data

Fsfloat, default: 2

The sampling frequency (samples per time unit). It is used to calculate the Fourier frequencies, freqs, in cycles per time unit.

windowcallable or ndarray, default: window_hanning

A function or a vector of length NFFT. To create window vectors see window_hanning, window_none, numpy.blackman, numpy.hamming, numpy.bartlett, scipy.signal, scipy.signal.get_window, etc. If a function is passed as the argument, it must take a data segment as an argument and return the windowed version of the segment.

sides{'default', 'onesided', 'twosided'}, optional

Which sides of the spectrum to return. 'default' is one-sided for real data and two-sided for complex data. 'onesided' forces the return of a one-sided spectrum, while 'twosided' forces two-sided.

The number of points to which the data segment is padded when performing the FFT. While not increasing the actual resolution of the spectrum (the minimum distance between resolvable peaks), this can give more points in the plot, allowing for more detail. This corresponds to the n parameter in the call to fft. The default is None, which sets pad_to equal to the length of the input signal (i.e. no padding).

Returns:
spectrum1-D array

The complex-valued frequency spectrum.

freqs1-D array

The frequencies corresponding to the elements in spectrum.

psd

Returns the power spectral density.

complex_spectrum

Returns the complex-valued frequency spectrum.

magnitude_spectrum

Returns the absolute value of the complex_spectrum.

angle_spectrum

Returns the angle of the complex_spectrum.

phase_spectrum

Returns the phase (unwrapped angle) of the complex_spectrum.

specgram

Can return the complex spectrum of segments within the signal.

matplotlib.mlab.csd(x, y, NFFT=None, Fs=None, detrend=None, window=None, noverlap=None, pad_to=None, sides=None, scale_by_freq=None)[source]#

Compute the cross-spectral density.

The cross spectral density $$P_{xy}$$ by Welch's average periodogram method. The vectors x and y are divided into NFFT length segments. Each segment is detrended by function detrend and windowed by function window. noverlap gives the length of the overlap between segments. The product of the direct FFTs of x and y are averaged over each segment to compute $$P_{xy}$$, with a scaling to correct for power loss due to windowing.

If len(x) < NFFT or len(y) < NFFT, they will be zero padded to NFFT.

Parameters:
x, y1-D arrays or sequences

Arrays or sequences containing the data

Fsfloat, default: 2

The sampling frequency (samples per time unit). It is used to calculate the Fourier frequencies, freqs, in cycles per time unit.

windowcallable or ndarray, default: window_hanning

A function or a vector of length NFFT. To create window vectors see window_hanning, window_none, numpy.blackman, numpy.hamming, numpy.bartlett, scipy.signal, scipy.signal.get_window, etc. If a function is passed as the argument, it must take a data segment as an argument and return the windowed version of the segment.

sides{'default', 'onesided', 'twosided'}, optional

Which sides of the spectrum to return. 'default' is one-sided for real data and two-sided for complex data. 'onesided' forces the return of a one-sided spectrum, while 'twosided' forces two-sided.

The number of points to which the data segment is padded when performing the FFT. This can be different from NFFT, which specifies the number of data points used. While not increasing the actual resolution of the spectrum (the minimum distance between resolvable peaks), this can give more points in the plot, allowing for more detail. This corresponds to the n parameter in the call to fft. The default is None, which sets pad_to equal to NFFT

NFFTint, default: 256

The number of data points used in each block for the FFT. A power 2 is most efficient. This should NOT be used to get zero padding, or the scaling of the result will be incorrect; use pad_to for this instead.

detrend{'none', 'mean', 'linear'} or callable, default: 'none'

The function applied to each segment before fft-ing, designed to remove the mean or linear trend. Unlike in MATLAB, where the detrend parameter is a vector, in Matplotlib it is a function. The mlab module defines detrend_none, detrend_mean, and detrend_linear, but you can use a custom function as well. You can also use a string to choose one of the functions: 'none' calls detrend_none. 'mean' calls detrend_mean. 'linear' calls detrend_linear.

scale_by_freqbool, default: True

Whether the resulting density values should be scaled by the scaling frequency, which gives density in units of 1/Hz. This allows for integration over the returned frequency values. The default is True for MATLAB compatibility.

noverlapint, default: 0 (no overlap)

The number of points of overlap between segments.

Returns:
Pxy1-D array

The values for the cross spectrum $$P_{xy}$$ before scaling (real valued)

freqs1-D array

The frequencies corresponding to the elements in Pxy

psd

equivalent to setting y = x.

References

Bendat & Piersol -- Random Data: Analysis and Measurement Procedures, John Wiley & Sons (1986)

matplotlib.mlab.detrend(x, key=None, axis=None)[source]#

Return x with its trend removed.

Parameters:
xarray or sequence

Array or sequence containing the data.

key{'default', 'constant', 'mean', 'linear', 'none'} or function

The detrending algorithm to use. 'default', 'mean', and 'constant' are the same as detrend_mean. 'linear' is the same as detrend_linear. 'none' is the same as detrend_none. The default is 'mean'. See the corresponding functions for more details regarding the algorithms. Can also be a function that carries out the detrend operation.

axisint

The axis along which to do the detrending.

detrend_mean

Implementation of the 'mean' algorithm.

detrend_linear

Implementation of the 'linear' algorithm.

detrend_none

Implementation of the 'none' algorithm.

matplotlib.mlab.detrend_linear(y)[source]#

Return x minus best fit line; 'linear' detrending.

Parameters:
y0-D or 1-D array or sequence

Array or sequence containing the data

detrend_mean

Another detrend algorithm.

detrend_none

Another detrend algorithm.

detrend

A wrapper around all the detrend algorithms.

matplotlib.mlab.detrend_mean(x, axis=None)[source]#

Return x minus the mean(x).

Parameters:
xarray or sequence

Array or sequence containing the data Can have any dimensionality

axisint

The axis along which to take the mean. See numpy.mean for a description of this argument.

detrend_linear

Another detrend algorithm.

detrend_none

Another detrend algorithm.

detrend

A wrapper around all the detrend algorithms.

matplotlib.mlab.detrend_none(x, axis=None)[source]#

Return x: no detrending.

Parameters:
xany object

An object containing the data

axisint

This parameter is ignored. It is included for compatibility with detrend_mean

detrend_mean

Another detrend algorithm.

detrend_linear

Another detrend algorithm.

detrend

A wrapper around all the detrend algorithms.

Compute the magnitude (absolute value) of the frequency spectrum of x. Data is padded to a length of pad_to and the windowing function window is applied to the signal.

Parameters:
x1-D array or sequence

Array or sequence containing the data

Fsfloat, default: 2

The sampling frequency (samples per time unit). It is used to calculate the Fourier frequencies, freqs, in cycles per time unit.

windowcallable or ndarray, default: window_hanning

A function or a vector of length NFFT. To create window vectors see window_hanning, window_none, numpy.blackman, numpy.hamming, numpy.bartlett, scipy.signal, scipy.signal.get_window, etc. If a function is passed as the argument, it must take a data segment as an argument and return the windowed version of the segment.

sides{'default', 'onesided', 'twosided'}, optional

Which sides of the spectrum to return. 'default' is one-sided for real data and two-sided for complex data. 'onesided' forces the return of a one-sided spectrum, while 'twosided' forces two-sided.

The number of points to which the data segment is padded when performing the FFT. While not increasing the actual resolution of the spectrum (the minimum distance between resolvable peaks), this can give more points in the plot, allowing for more detail. This corresponds to the n parameter in the call to fft. The default is None, which sets pad_to equal to the length of the input signal (i.e. no padding).

Returns:
spectrum1-D array

The magnitude (absolute value) of the frequency spectrum.

freqs1-D array

The frequencies corresponding to the elements in spectrum.

psd

Returns the power spectral density.

complex_spectrum

Returns the complex-valued frequency spectrum.

magnitude_spectrum

Returns the absolute value of the complex_spectrum.

angle_spectrum

Returns the angle of the complex_spectrum.

phase_spectrum

Returns the phase (unwrapped angle) of the complex_spectrum.

specgram

Can return the complex spectrum of segments within the signal.

Compute the phase of the frequency spectrum (unwrapped phase spectrum) of x. Data is padded to a length of pad_to and the windowing function window is applied to the signal.

Parameters:
x1-D array or sequence

Array or sequence containing the data

Fsfloat, default: 2

The sampling frequency (samples per time unit). It is used to calculate the Fourier frequencies, freqs, in cycles per time unit.

windowcallable or ndarray, default: window_hanning

A function or a vector of length NFFT. To create window vectors see window_hanning, window_none, numpy.blackman, numpy.hamming, numpy.bartlett, scipy.signal, scipy.signal.get_window, etc. If a function is passed as the argument, it must take a data segment as an argument and return the windowed version of the segment.

sides{'default', 'onesided', 'twosided'}, optional

Which sides of the spectrum to return. 'default' is one-sided for real data and two-sided for complex data. 'onesided' forces the return of a one-sided spectrum, while 'twosided' forces two-sided.

The number of points to which the data segment is padded when performing the FFT. While not increasing the actual resolution of the spectrum (the minimum distance between resolvable peaks), this can give more points in the plot, allowing for more detail. This corresponds to the n parameter in the call to fft. The default is None, which sets pad_to equal to the length of the input signal (i.e. no padding).

Returns:
spectrum1-D array

The phase of the frequency spectrum (unwrapped phase spectrum).

freqs1-D array

The frequencies corresponding to the elements in spectrum.

psd

Returns the power spectral density.

complex_spectrum

Returns the complex-valued frequency spectrum.

magnitude_spectrum

Returns the absolute value of the complex_spectrum.

angle_spectrum

Returns the angle of the complex_spectrum.

phase_spectrum

Returns the phase (unwrapped angle) of the complex_spectrum.

specgram

Can return the complex spectrum of segments within the signal.

matplotlib.mlab.psd(x, NFFT=None, Fs=None, detrend=None, window=None, noverlap=None, pad_to=None, sides=None, scale_by_freq=None)[source]#

Compute the power spectral density.

The power spectral density $$P_{xx}$$ by Welch's average periodogram method. The vector x is divided into NFFT length segments. Each segment is detrended by function detrend and windowed by function window. noverlap gives the length of the overlap between segments. The $$|\mathrm{fft}(i)|^2$$ of each segment $$i$$ are averaged to compute $$P_{xx}$$.

If len(x) < NFFT, it will be zero padded to NFFT.

Parameters:
x1-D array or sequence

Array or sequence containing the data

Fsfloat, default: 2

The sampling frequency (samples per time unit). It is used to calculate the Fourier frequencies, freqs, in cycles per time unit.

windowcallable or ndarray, default: window_hanning

A function or a vector of length NFFT. To create window vectors see window_hanning, window_none, numpy.blackman, numpy.hamming, numpy.bartlett, scipy.signal, scipy.signal.get_window, etc. If a function is passed as the argument, it must take a data segment as an argument and return the windowed version of the segment.

sides{'default', 'onesided', 'twosided'}, optional

Which sides of the spectrum to return. 'default' is one-sided for real data and two-sided for complex data. 'onesided' forces the return of a one-sided spectrum, while 'twosided' forces two-sided.

The number of points to which the data segment is padded when performing the FFT. This can be different from NFFT, which specifies the number of data points used. While not increasing the actual resolution of the spectrum (the minimum distance between resolvable peaks), this can give more points in the plot, allowing for more detail. This corresponds to the n parameter in the call to fft. The default is None, which sets pad_to equal to NFFT

NFFTint, default: 256

The number of data points used in each block for the FFT. A power 2 is most efficient. This should NOT be used to get zero padding, or the scaling of the result will be incorrect; use pad_to for this instead.

detrend{'none', 'mean', 'linear'} or callable, default: 'none'

The function applied to each segment before fft-ing, designed to remove the mean or linear trend. Unlike in MATLAB, where the detrend parameter is a vector, in Matplotlib it is a function. The mlab module defines detrend_none, detrend_mean, and detrend_linear, but you can use a custom function as well. You can also use a string to choose one of the functions: 'none' calls detrend_none. 'mean' calls detrend_mean. 'linear' calls detrend_linear.

scale_by_freqbool, default: True

Whether the resulting density values should be scaled by the scaling frequency, which gives density in units of 1/Hz. This allows for integration over the returned frequency values. The default is True for MATLAB compatibility.

noverlapint, default: 0 (no overlap)

The number of points of overlap between segments.

Returns:
Pxx1-D array

The values for the power spectrum $$P_{xx}$$ (real valued)

freqs1-D array

The frequencies corresponding to the elements in Pxx

specgram

specgram differs in the default overlap; in not returning the mean of the segment periodograms; and in returning the times of the segments.

magnitude_spectrum

returns the magnitude spectrum.

csd

returns the spectral density between two signals.

References

Bendat & Piersol -- Random Data: Analysis and Measurement Procedures, John Wiley & Sons (1986)

matplotlib.mlab.specgram(x, NFFT=None, Fs=None, detrend=None, window=None, noverlap=None, pad_to=None, sides=None, scale_by_freq=None, mode=None)[source]#

Compute a spectrogram.

Compute and plot a spectrogram of data in x. Data are split into NFFT length segments and the spectrum of each section is computed. The windowing function window is applied to each segment, and the amount of overlap of each segment is specified with noverlap.

Parameters:
xarray-like

1-D array or sequence.

Fsfloat, default: 2

The sampling frequency (samples per time unit). It is used to calculate the Fourier frequencies, freqs, in cycles per time unit.

windowcallable or ndarray, default: window_hanning

A function or a vector of length NFFT. To create window vectors see window_hanning, window_none, numpy.blackman, numpy.hamming, numpy.bartlett, scipy.signal, scipy.signal.get_window, etc. If a function is passed as the argument, it must take a data segment as an argument and return the windowed version of the segment.

sides{'default', 'onesided', 'twosided'}, optional

Which sides of the spectrum to return. 'default' is one-sided for real data and two-sided for complex data. 'onesided' forces the return of a one-sided spectrum, while 'twosided' forces two-sided.

The number of points to which the data segment is padded when performing the FFT. This can be different from NFFT, which specifies the number of data points used. While not increasing the actual resolution of the spectrum (the minimum distance between resolvable peaks), this can give more points in the plot, allowing for more detail. This corresponds to the n parameter in the call to fft. The default is None, which sets pad_to equal to NFFT

NFFTint, default: 256

The number of data points used in each block for the FFT. A power 2 is most efficient. This should NOT be used to get zero padding, or the scaling of the result will be incorrect; use pad_to for this instead.

detrend{'none', 'mean', 'linear'} or callable, default: 'none'

The function applied to each segment before fft-ing, designed to remove the mean or linear trend. Unlike in MATLAB, where the detrend parameter is a vector, in Matplotlib it is a function. The mlab module defines detrend_none, detrend_mean, and detrend_linear, but you can use a custom function as well. You can also use a string to choose one of the functions: 'none' calls detrend_none. 'mean' calls detrend_mean. 'linear' calls detrend_linear.

scale_by_freqbool, default: True

Whether the resulting density values should be scaled by the scaling frequency, which gives density in units of 1/Hz. This allows for integration over the returned frequency values. The default is True for MATLAB compatibility.

noverlapint, default: 128

The number of points of overlap between blocks.

modestr, default: 'psd'
What sort of spectrum to use:
'psd'

Returns the power spectral density.

'complex'

Returns the complex-valued frequency spectrum.

'magnitude'

Returns the magnitude spectrum.

'angle'

Returns the phase spectrum without unwrapping.

'phase'

Returns the phase spectrum with unwrapping.

Returns:
spectrumarray-like

2D array, columns are the periodograms of successive segments.

freqsarray-like

1-D array, frequencies corresponding to the rows in spectrum.

tarray-like

1-D array, the times corresponding to midpoints of segments (i.e the columns in spectrum).

psd

differs in the overlap and in the return values.

complex_spectrum

similar, but with complex valued frequencies.

magnitude_spectrum

similar single segment when mode is 'magnitude'.

angle_spectrum

similar to single segment when mode is 'angle'.

phase_spectrum

similar to single segment when mode is 'phase'.

Notes

detrend and scale_by_freq only apply when mode is set to 'psd'.

matplotlib.mlab.window_hanning(x)[source]#

Return x times the Hanning (or Hann) window of len(x).

window_none

Another window algorithm.

matplotlib.mlab.window_none(x)[source]#

No window function; simply return x.

window_hanning