ELO-SPHERES Audiovisual Test (SEAT)

Implements the logic for running audiovisual intelligibility tests.

Targets Windows platform with Windows Subsystem for Linux already installed.

Installation

Open an anaconda shell

See https://docs.conda.io/projects/conda/en/latest/user-guide/install/index.html for installation instructions.

Get the code and test data using

(base) git clone https://github.com/ImperialCollegeLondon/sap-elospheres-audiovisual-test.git C:\gitwin\ImperialCollegeLondon\sap-elospheres-audiovisual-test

N.B. For now, the built-in tests have files with hardcoded paths. This is pretty ugly but avoids lots of complexity.

Create a new conda environment
```
(base) conda env create -n seat --file seat.yml
```
N.B. You can call the environment (-n <env_name>) whatever you like but seat will be used here.
Activate the new environment and change into the main seat directory
```
(base) conda activate seat
(seat) cd seat
```

Test the installation

These are just to check that everything installed properly. They are by no means exhaustive.

Unit tests

These all run in one go and don't require any user interaction

(seat) python -m unittest discover unit_tests

Setup audio

Run

(seat) python .\setup_jacktrip_wsl.py

Switch to the 'settings' tab and ensure appropriate values have been entered. Once satisfied, click the Start button. Switch to the console tab to inspect the output. The status at the bottom of the gui will change from Disconnected to Starting. Once the connection is established it will show 'Connected'

Press the Test button to play a metronome test tone on the left channel. Note that the metronome is a jack-enabled executable running under WSL. Press again to stop the test.

Assuming everything is ok, press the Stop button, return to the settings tab and press Save. These settings will be used by default from now on.

Demo tests

These demonstrate the functionality of a perceptual test.

Close any open consoles from previous tests and open a new anaconda prompt

Setup the environment

(base) conda activate seat
(seat) python .\start_jacktrip_wsl.py

Note that in this case jacktrip is left running. Once you have finished you can end jacktrip using

(seat) python .\stop_jacktrip_wsl.py

The first test demonstrates a signal detection task

(seat) python -m demo_tests.test_01_targettoneinnoise

The second test demonstrates a speech intelligibility task

python -m demo_tests.test_02_targetspeechtwomaskers

If the listeningeffortplayer app is running then the target speech video should be visible.

Real usage

Open the ListeningEffortPlayer app
Open anaconda prompt
- Activate the seat environment
- Run gui.py
- Choose the csv file which contains all the block definitions
- Select the block to run and add optional subject metadata
- Click 'Run' to start the block
- Experimenter selects the correctly reported keywords

cd C:\gitwin\ImperialCollegeLondon\sap-elospheres-audiovisual-test
git pull
conda env update -n seat --file seat.yml
conda activate seat
cd seat
python .\gui.py
#
# In the background jacktripcontrol will start with the previously saved settings
# Once the 'jtc...' button turns green and you have selected a block you can press 'Run'
# press the 'jtc...' button to change the settings and/or restart jacktrip.