Skip to content

Getting started

Files needed

Your input structure

Your files should be organized in the following types of structure.

The input data may consist of audio, text, or a combination of both modalities. It is not required to provide all modalities for each participants. It is also possible to provide multiple audio or text files for one participant.

Each files should have a distinct name.

Two example of a good input:

/project/
|-- data/
|   |-- audio/
|   |   |-- sub-ABC123_task-naming01.wav
|   |   |-- sub-ABC456_task-naming03.wav
|   |-- text/
|   |   |-- sub-ABC123_task-speech.txt
|   |   |-- sub-ABC456_task-speech.txt

/project/
|-- data/
|   |-- audio/
|   |   |-- sub-ABC123/
|   |   |   |-- sub-ABC123_task-naming01.wav
|   |   |-- sub-ABC456/
|   |   |   |-- sub-ABC456_task-naming03.wav
|   |-- text/
|   |   |-- sub-ABC123/
|   |   |   |-- sub-ABC123_task-speech.txt
|   |   |-- sub-ABC456/
|   |   |   |-- sub-ABC456_task-speech.txt
|   |   |-- sub-ABC456_task-otherspeech.txt

The next example is a bad input because of filenames collision for the last two text files:

/project/
|-- data/
|   |-- audio/
|   |   |-- sub-ABC123/
|   |   |   |-- sub-ABC123_task-naming01.wav
|   |   |-- sub-ABC456/
|   |   |   |-- sub-ABC456_task-naming03.wav
|   |-- text/
|   |   |-- sub-ABC123/
|   |   |   |-- sub-ABC123_task-speech.txt
|   |   |-- sub-ABC456/
|   |   |   |-- sub-ABC456_task-speech.txt
|   |   |-- sub-ABC456_task-speech.txt

participant_ids.csv

This CSV file must contain at least 4 columns:

  • participant_id is required for the pipeline to find your audio and text files. These filenames must begin by the participant_id.
  • language: 2 choices, en or fr.
  • sex: 2 choices, male or female.
  • task: 3 choices, cookie, picnic or nothing.

Example:

participant_id language sex task
sub-ABC123 en male cookie
sub-ABC456 en female picnic

nextflow.config

Here the content of your nextflow.config file:

apptainer.enabled = true
params {
    audio_folder = "/project/data/audio"
    text_folder = "/project/data/text"
    container_dir = "/your/folder/to/save/containers"
}

See Installation for the container_dir parameter.

If you are a member of the UNF, you only need to specify the audio_folder and text_folder parameters.

Running

nextflow run lingualab/speechmetryflow -r {last_release_or_tag} --input participant_ids.csv

Replace the -r option with the release you want to use.

If you are a member of the UNF, do not forget to add -profile unf_elm.

output

The pipeline produces csv files in the results/Statistics directory:
OPENSMILE/population_opensmile_metrics_{set}.csv metrics compute with opensmile
LINGUALAB_AUDIO/population_lingualab_audio.csv metrics compute with parselmouth
UHMOMETER/population_uhmometer_metrics.csv metrics compute with uhm-o-meter
LINGUALAB_TEXT/population_textdescriptives_text.csv metrics compute with TextDescriptives
LINGUALAB_TEXT/population_speechmetryflow_text.csv metrics compute within speechmetryflow, See Text Features section