SpeakerCollection
SpeakerCollection(self, track_list=[])
A class to represent the vowel system of all speakers in a TextGrid.
Intended usage
It is a subclass of defaultdict, and can be keyed by the (file_name, group_name) tuple.
vowel_measurements = [VowelMeasurement(t) for t in fasttrack_tracks]
speakers = SpeakerCollection(vowel_measurements)Parameters
| Name | Type | Description | Default |
|---|---|---|---|
track_list |
list[VowelMeasurement] | A list of VowelMeasurements. |
[] |
Classes
| Name | Description |
|---|---|
| containing_class | A class for an entire vowel system. |
containing_class
SpeakerCollection.containing_class(self, track_list=EMPTY_LIST)
A class for an entire vowel system.
Intended Usage
It is a subclass of defaultdict, so it can be keyed by vowel class label.
vowel_measurements = [VowelMeasurement(t) for t in fasttrack_tracks]
vowel_system = VowelClassCollection(vowel_measurements)Parameters
| Name | Type | Description | Default |
|---|---|---|---|
track_list |
list[VowelMeasurement] | A list of VowelMeasurements. |
EMPTY_LIST |
Attributes
| Name | Type | Description |
|---|---|---|
| winners | list[OneTrack] | All winner tracks from the entire vowel system. |
| vowel_measurements | list[VowelMeasurement] | All VowelMeasurement objects within this vowel system |
| textgrid | AlignedTextGrid | The AlignedTextGrid associated with this vowel system. |
| winner_expanded_formants | NDArray[Shape[20, FormantN], Float] |
A cached property that returns the expanded formants for the winners. |
| winner_param | NDArray[Shape[Param, Formant, N], Float] |
An array of all parameters from all winners across the vowel system. |
| winner_maxformant | NDArray[Shape[1, N], Float] |
An array of the maximum formants of all winners across the vowel system |
| winner_param_mean | NDArray[Shape[1, FormantParam], Float] |
The mean of all DCT parameters across all formants for the winners in this vowel system. |
| winner_param_cov | NDArray[Shape[FormantParam, FormantParam], Float] |
The covariance of all parameters across all formants for the winners in this vowel system |
| winner_param_icov | NDArray[Shape[FormantParam, FormantParam], Float] |
The inverse of winner_param_cov. |
| winner_maxformant_mean | float | The mean maximum formant across all winners in this vowel system. |
| winner_maxformant_cov | NDArray[Shape[1, 1], Float] |
The covariance of the maximum formant across all winners in this vowel system. |
| winner_maxformant_icov | NDArray[Shape[1, 1], Float] |
The inverse of winner_maxformant_cov |
Classes
| Name | Description |
|---|---|
| containing_class | A class used to represent a vowel class. |
containing_class
SpeakerCollection.containing_class.containing_class(self, label='', vowel_measurements=lambda: []())
A class used to represent a vowel class.
Intended Usage
VowelClass subclasses collections.abc.Sequence, so it is indexable. While it can be created on its own, it is best to leave this up to either VowelClassCollection or SpeakerCollection.
vowel_measurements = [VowelMeasurement(t) for t in fasttrack_tracks]
vowel_class = VowelClass("ay", vowel_measurements)Parameters
| Name | Type | Description | Default |
|---|---|---|---|
label |
str | The vowel class label | '' |
tracks |
list[VowelMeasurement] | A list of VowelMeasurements | required |
Attributes
| Name | Type | Description |
|---|---|---|
| label | str | label of the vowel class |
| tracks | list | A list of VowelMeasurements |
| vowel_system | VowelClassCollection | The containing vowel system |
| winners | list[OneTrack] | A list of winner OneTracks from the vowel class |
| winner_param | NDArray[Shape[Param, Formant, N], Float] |
An np.array of winner DCT parameters from the vowel class |
| winner_param_mean | NDArray[Shape[ParamFormant, 1], Float] |
Mean of winner DCT parameters |
| winner_param_cov | NDArray[Shape[ParamFormant, ParamFormant], Float] |
Covariance of winner DCT parameters |
| winner_param_icov | NDArray[Shape[ParamFormant, ParamFormant], Float] |
Inverse covariance of winner DCT parameters |
Classes
| Name | Description |
|---|---|
| containing_class | A class used to represent a vowel measurement. |
containing_class
SpeakerCollection.containing_class.containing_class.containing_class(self, track, heuristic=Heuristic())
A class used to represent a vowel measurement.
Intended Usage
Certain properties of a VowelMeasurement instance are set by its membership within a VowelClass and that VowelClass’s membership in a VowelClassCollection. These memberships are best managed by passing a list of VowelMeasurements to SpeakerCollection.
vowel_measurements = [VowelMeasurement(t) for t in fasttrack_tracks]
speakers = SpeakerCollection(vowel_measurements)Parameters
| Name | Type | Description | Default |
|---|---|---|---|
track |
fasttrackpy.CandidateTracks |
A fasttrackpy.CandidateTrracks object | required |
heuristic |
Heuristic | A point measurement Heuristic to use. Defaults to Heuristic(). | Heuristic() |
Attributes
| Name | Type | Description |
|---|---|---|
| track | fasttrackpy.CandidateTracks |
an object of CandidateTracks class |
| candidates | list | list of candidates for the track |
| heuristic | Heuristic | an object of Heuristic class (default is Heuristic()) |
| vowel_class | VowelClass | The containing VowelClass object |
| formant_array | FormantArray | A FormantArray object |
| file_name | str | name of the file of the track |
| group | str | TierGroup of the track |
| id | str | id of the track |
| interval | aligned_textgrid.SequenceInterval |
interval of the track |
| label | str | label of the track |
| n_formants | int | number of formants in the track |
| optimized | int | The number of optimization iterations the vowel measurement has been through. |
| winner | OneTrack | fasttrackpy.OneTrack The winning formant track |
| winner_index | int | The index of the winning formant track |
| cand_param | NDArray[Shape[Param, Formant, Cand], Float] |
A array of the candidate DCT parameters. |
| cand_maxformant | NDArray[Shape[1, Cand], Float] |
An array of the candidate maximum formants. |
| cand_error | NDArray[Shape[Cand], Float] |
An array of the candidate smoothing error. |
| cand_error_logprob_vm | NDArray[Shape[Cand], Float] |
Conversion of the smooth error to log probabilities. The candidate with the lowest error = log(1), and the candidate with the largest error = log(0). |
| cand_param_(mahal/logprob)_speaker_byvclass | NDArray[Shape[Cand], Float] |
The mahalanobis distance (mahal) or associated log probability (logprob) for each candidate relative to the VowelClass for this speaker. These are calculated by drawing the relevant mean and covariance matrix from vm.vowel_class |
| cand_param_(mahal/logprob)_speaker_global | NDArray[Shape[Cand], Float] |
The mahalanobis distance (mahal) or associated log probability (logprob) for each candidate relative to all vowel measurements for this speaker. These are calculated by drawing the relevant mean and covariance matrix from vm.vowel_class.vowel_system |
| cand_param_(mahal/logprob)_corpus_byvclass | NDArray[Shape[Cand], Float] |
The mahalanobis distance (mahal) or associated log probability (logprob) for each candidate relative to this vowel class across all speakers. These are calculated by drawing the relevant mean and covariance matrix from vm.vowel_class.vowel_system.corpus |
| point_measure | pl.DataFrame |
A polars dataframe of the point measurement for this vowel. |
| vm_context | pl.DataFrame |
A polars dataframe of contextual information for the vowel measurement. |
Methods
| Name | Description |
|---|---|
| to_param_df | Return DataFrame of formant DCT parameters. |
| to_point_df | Return a DataFrame of point measurements |
| to_tracks_df | Return a DataFrame of the formant tracks |
to_param_df
SpeakerCollection.containing_class.containing_class.containing_class.to_param_df(output='log_param')
Return DataFrame of formant DCT parameters.
Returns
| Type | Description |
|---|---|
pl.DataFrame |
A DataFrame of formant DCT parameters |
to_point_df
SpeakerCollection.containing_class.containing_class.containing_class.to_point_df()
Return a DataFrame of point measurements
Returns
| Type | Description |
|---|---|
pl.DataFrame |
A DataFrame of vowel point measures. |
to_tracks_df
SpeakerCollection.containing_class.containing_class.containing_class.to_tracks_df()
Return a DataFrame of the formant tracks
Returns
| Type | Description |
|---|---|
pl.DataFrame |
A dataframe with formant track data. |
Methods
| Name | Description |
|---|---|
| to_param_df | Return DataFrame of formant DCT parameters. |
| to_point_df | Return a DataFrame of point measurements |
| to_tracks_df | Return DataFrame of formanttracks. |
to_param_df
SpeakerCollection.containing_class.containing_class.to_param_df(output='log_param')
Return DataFrame of formant DCT parameters.
Returns
| Type | Description |
|---|---|
pl.DataFrame |
A DataFrame of formant DCT parameters |
to_point_df
SpeakerCollection.containing_class.containing_class.to_point_df()
Return a DataFrame of point measurements
Returns
| Type | Description |
|---|---|
pl.DataFrame |
A DataFrame of vowel point measures. |
to_tracks_df
SpeakerCollection.containing_class.containing_class.to_tracks_df()
Return DataFrame of formanttracks.
Returns
| Type | Description |
|---|---|
pl.DataFrame |
A DataFrame of formant tracks |
Methods
| Name | Description |
|---|---|
| to_param_df | Return DataFrame of formant DCT parameters. |
| to_point_df | Return a DataFrame of point measurements |
| to_tracks_df | Return a DataFrame of the formant tracks |
to_param_df
SpeakerCollection.containing_class.to_param_df(output='log_param')
Return DataFrame of formant DCT parameters.
Returns
| Type | Description |
|---|---|
pl.DataFrame |
A DataFrame of formant DCT parameters |
to_point_df
SpeakerCollection.containing_class.to_point_df()
Return a DataFrame of point measurements
Returns
| Type | Description |
|---|---|
pl.DataFrame |
A DataFrame of vowel point measures. |
to_tracks_df
SpeakerCollection.containing_class.to_tracks_df()
Return a DataFrame of the formant tracks
Returns
| Type | Description |
|---|---|
pl.DataFrame |
A dataframe with formant track data. |
Methods
| Name | Description |
|---|---|
| to_param_df | This will return a dataframe of the DCT parameters for all speakers. |
| to_point_df | This will return a DataFrame of point measurements |
| to_tracks_df | This will return a data frame of formant |
to_param_df
SpeakerCollection.to_param_df(output='log_param')
This will return a dataframe of the DCT parameters for all speakers. If output is passed param, it will be the DCT parameters in the original Hz. If passed log_param, it will be the DCT parameters over log(Hz).
Parameters
| Name | Type | Description | Default |
|---|---|---|---|
output |
Literal[‘param’, ‘log_param’] | Which set of DCT parameters to return. Defaults to “log_param”. | 'log_param' |
Returns
| Type | Description |
|---|---|
pl.DataFrame |
A DataFrame of DCT parameters for all speakers. |
to_point_df
SpeakerCollection.to_point_df()
This will return a DataFrame of point measurements for all speakers Returns: (pl.DataFrame): A DataFrame of vowel point measurements.
to_tracks_df
SpeakerCollection.to_tracks_df()
This will return a data frame of formant tracks for all speakers.
Returns
| Type | Description |
|---|---|
pl.DataFrame |
A dataframe of formant tracks for all speakers. |