All FastTrack Options

Use of a config file

Instead of needing to include every option you wish to cusotomize as a command line entry, you can set them in a config file, and pass it to the fasttrack command with config. For example

# config.yml file
audio: speaker.wav
textgrid: speaker.TextGrid
dest: results
entry_classes: SequenceInterval
target_tier: phones
target_labels: "..1"
smoother_order: 6

# at the command line: 
fasttrack audio-textgrid --config config.yml

Command specific options

Just Audio Options

Option	Meaning
`file`	A single audio file to process
`dir`	A directory of audio files to process

Audio + TextGrid Options

Option	Meaning
`audio`	Path to an audio file to process
`textgrid`	Path to a textgrid to process

Shared Options

The Following options are general across most fasttrack commands.

Output Destinations

Option	Meaning
`output`	Output file name
`dest`	Output directory

If an output file name is not provided, fasttrack will try to use a reasonable name based on the input audio file name.

Output Options

Option	Meaning	Default
`which_output`	Whether to save just the winning track (`winner`) or all candidates (`all`)	`winner`
`data_output`	Whether to save the formant tracks (`formants`) or the smoothing parameters (`param`)	`formants`
`separate_output`	When processing a corpus, save each file/group to a separate file?	`False`

TextGrid Processing Options

Option	Meaning	Default
`entry_classes`	Format of the TextGrid.	`"Word\|Phone"`
`target_tier`	The tier to target for processing	`Phone`
`target_labels`	A regex for the labels to target for processing	`"[AEIOU]"`

Notes:

If your textgrid is not the output of a forced-aligner, change the entry_classes value to SequenceInterval and the value of target_tier to whatever the tier name is that you want to process.
Any string that can be interpreted as a regular expression can be passed to processing.

Audio Processing Options

Option	Meaning	Default
`min_duration`	The minumum duration segment to analyze	0.05 (s)
`min_max_formant`	The lowest maximum formant to try	4000(hz)
`max_max_formant`	The highest maximum formant to try	7000(hz)
`nstep`	The number of steps between the lowest and the highest maximum formant	20
`n_formants`	The number of formants to try to track in the audio	4
`window_length`	The formant analysis window length	0.025 (s)
`time_step`	The formant analysis step size	0.002 (s)
`pre_emphasis_from`	Pre-emphasis to be applied before formant tracking	50 (hz)

Smoother Options

Option	Meaning	Default
`smoother_method`	The smoothing method to use. It’s not recommended you change this from the default.	`dct_smooth_regression`
`smoother_order`	The “order” of the smooth. More is wigglier.	5
`loss_method`	Method for calculating the error between the smooth and measured formants. It’s not recommended that you change the default	`lmse`