sample.plots module

Plot functions for helping visualization

This module requires extra dependencies, which you can install with

pip install lim-sample[plots]

sample.plots.beatsdrop_comparison(model: SAMPLE, beatsdrops: Dict[str, BeatRegression], x: array, track_i: int = 0, fs: Optional[float] = None, transpose: bool = False, signal_hilbert_am: Union[bool, int] = False, warnings_ignore: bool = True, axs: Optional[Sequence[Axes]] = None, fig: Optional[Figure] = None)

Compare beat regression models

Parameters:

  • model (SAMPLE) – Fitted sample.sample.SAMPLE instance

  • beatsdrops (dict) – Dictionary of sample.beatsdrop.regression.BeatRegression values

  • x (array) – Original signal

  • track_i (int) – Index of the (beating) track of interest

  • fs (float) – Sampling frequency. If unspecified, it will be inferred by the model

  • transpose (bool) – Swap small multiples rows and columns

  • signal_hilbert_am (bool or int) – If True, then plot the Hilbert envelope of the signal, instead of the signal. If an int, then subsample the envelope by this factor

  • warnings_ignore (bool) – if True (default), then ignore warnings while fitting beat regression models

  • axs (sequence of axes) – Axes onto which to plot. If unspecified, they will be defined on fig

  • fig (figure) – Figure onto which to put axes. If unspecified, the current figure will be used

Returns:

Figure and axes for the plot

Return type:

figure, axes

sample.plots.resynthesis(x: ~numpy.ndarray, models: ~typing.Optional[~typing.Dict[str, ~sample.sample.SAMPLE]] = None, fs: ~typing.Optional[float] = None, original: bool = True, original_k: str = 'Original', axs: ~typing.Optional[~typing.Sequence[~matplotlib.axes._axes.Axes]] = None, fig: ~typing.Optional[~matplotlib.figure.Figure] = None, foreach: ~typing.Optional[~typing.Callable[[int, str, ~numpy.ndarray], ~typing.Any]] = None, wav_kws: ~typing.Optional[~typing.Dict[str, ~typing.Any]] = None, tf_kws: ~typing.Optional[~typing.Dict[str, ~typing.Any]] = None, tf_fun: ~typing.Callable[[~numpy.ndarray, float], ~typing.Tuple[~typing.Tuple[float, float], ~numpy.ndarray]] = <function _stft4resynthesis.<locals>._stft4resynthesis_>, db_floor: float = -90, **kwargs)

Plot a signal and its resynthesis, using different models

Parameters:

  • x (array) – Original signal

  • models (dict) – Dictionary of sample.sample.SAMPLE values

  • fs (float) – Sampling frequency. If unspecified, it will be inferred by the models

  • original (bool) – If True (default), also plot original signals

  • original_k (str) – Label for original signal

  • axs (sequence of axes) – Axes onto which to plot. If unspecified, they will be defined on fig

  • fig (figure) – Figure onto which to put axes. If unspecified, the current figure will be used

  • foreach (callable) – Function to be called for every model. It should take three inputs: the model index, the model label and the audio signal

  • wav_kws (dict) – Additional keyword arguments for wave plot (matplotlib.pyplot.plot())

  • tf_kws (dict) – Additional keyword arguments for time-frequency plot (sample.plots.tf_plot())

  • tf_fun (callable) – Function for getting a time-frequency representation from a signal. It should take an audio array as a positional argument and the sampling frequency fs as a keyword argument. It should return the frequency limits of the representation and the time-frequency matrix

  • db_floor (float) – Lower threshold for tf amplitude values, in dB. If None, the values will not be converted in dB

  • **kwargs – Keyword arguments for matplotlib.figure.Figure.subplots()

Returns:

Figure and list of axes for the plot

Return type:

figure, axes

sample.plots.sine_tracking_2d(s: Union[SinusoidalModel, SAMPLE], ax=None, **kwargs)

Plot sinusoidal tracks detected by the model on two axes, one for frequency and one for magnitude

Parameters:

  • m (sample.sms.sm.SinusoidalModel) – Trained sinusoidal model (or SAMPLE model)

  • ax – Axes list (optional)

  • **kwargs – Keyword arguments for matplotlib.pyplot.plot()

Returns:

The axes list

sample.plots.sine_tracking_3d(s: Union[SinusoidalModel, SAMPLE], ax=None)

Plot sinusoidal tracks detected by the model on one 3D axis

Parameters:
Returns:

The 3D axis

sample.plots.tf_plot(tf, tlim: Tuple[float, float] = (0, 1), flim: Tuple[float, float] = (0, 1), xlim: Optional[Tuple[float, float]] = None, ylim: Optional[Tuple[float, float]] = None, ax: Optional[Axes] = None, aspect_ratio: float = 1.3333333333333333, width: float = 8, **kwargs)

Plot a time-frequency matrix

Parameters:

  • tf (matrix) – Time-frequency matrix

  • tlim (float, float) – Extrema for time axis of matrix

  • flim (float, float) – Extrema for frequency axis of matrix

  • xlim (float, float) – Extrema for time axis of plot

  • ylim (float, float) – Extrema for frequency axis of plot

  • ax (Axes) – Axes onto which to plot the matrix

  • aspect_ratio (float) – Aspect ratio for image

  • width (float) – Width for figure size

  • **kwargs – Keyword arguments for imhsow()

Returns:

Axes where the matrix has been plotted

Return type:

Axes