ertk.preprocessing.vad_trim.VADTrimmerConfig

class ertk.preprocessing.vad_trim.VADTrimmerConfig(method: Method = Method.librosa, top_db: int = 60, energy_thresh: float = 2.5, freq_thresh: float = 150, sf_thresh: float = 10, window: float = 0.01, debug: bool = False, min_speech: int = 5, min_silence: int = 10)

Bases: ERTKConfig

VAD trimming configuration.

__init__(method: Method = Method.librosa, top_db: int = 60, energy_thresh: float = 2.5, freq_thresh: float = 150, sf_thresh: float = 10, window: float = 0.01, debug: bool = False, min_speech: int = 5, min_silence: int = 10) → None

Methods

__init__([method, top_db, energy_thresh, ...])

Inherited Methods

default()	Create default config.
from_config(config)	Create config object from any compatible config.
from_file(path)	Create config from YAML file and optionlly override some values.
merge_with_args(args)	Merge config with command-line arguments.
merge_with_config(config)	Merge other config into this config.
to_dictconfig()	Convert config to DictConfig.
to_file(path)	Write config to YAML file.
to_string()	Generate YAML string representation of config.

Attributes

debug	Whether to plot debug figures.
energy_thresh	Energy threshold for MH2009 VAD.
freq_thresh	Frequency threshold for MH2009 VAD.
method	VAD trimming method.
min_silence	Minimum number of consecutive silence frames to keep.
min_speech	Minimum number of consecutive speech frames to keep.
sf_thresh	Spectral flatness threshold for MH2009 VAD.
top_db	Top decibel threshold for librosa VAD.
window	Window size in seconds for MH2009 VAD.

debug: bool = False

Whether to plot debug figures.

energy_thresh: float = 2.5

Energy threshold for MH2009 VAD.

freq_thresh: float = 150

Frequency threshold for MH2009 VAD.

method: Method = 'librosa'

VAD trimming method.

min_silence: int = 10

Minimum number of consecutive silence frames to keep.

min_speech: int = 5

Minimum number of consecutive speech frames to keep.

sf_thresh: float = 10

Spectral flatness threshold for MH2009 VAD.

top_db: int = 60

Top decibel threshold for librosa VAD.

window: float = 0.01

Window size in seconds for MH2009 VAD.