JAST Construction Task software guide
A breakdown of tasks involved in setting up, running and analysing experiments, and links to the various software and tools available for each task
NOTE: Most of the software will present some help information if you use a "-help" or "--help" argument or supply an invalid argument.
All the software below runs on Windows and some will also work on Mac and Unix. All source code is available for programmers wishing to port the code to another platform, but programmers should check the libraries are available for their target platform.
Most software expects to be run from the directory containing the asc, xml and dtd files for the experiment, and the font.ttf file from the software/jast_install directory. The batch files mentioned below have the text $JCT_HOME which should be replaced by the install location of the JCT software (usually something like C:\JCT ). Batch files that need editing this way will contain editthis as part of their filename and should be copied to the same directory as your data files before editing. The estimated run times given here are very rough, and depend on a number of factors but should give you an idea of which processes you can run and wait for and which ones you will want to run overnight.
The following table lists the main tasks involved in designing, running and analysing experiments. For each task, the appropriate software and tools are indicated.
| Task | Software location (in the JCT Software distribution) |
Further Info/Links | Estimated Time to run per participant |
|---|---|---|---|
| Record video of the experiment | We used Camtasia and Tsunami TMPG | Camtasia Studio Tsunami MPG Video Settings |
Real-time x2 |
| Run the construction experiment | software/jast_install | JCT Software User Documentation |
Real-time |
| Create new construction tasks for experiment | software/JASTModelBuilder | JASTModelBuilder Documentation |
- |
| Modify the construction experiment software | software/jast_src/JCT |
JCT Software System Documentation (software design and libraries) JCT Software Developer Documentation (quick reference for adding features to an experiment) |
- |
| Generate videos from eyetracker data | software/JASTEDF2MPG/asc2mpg.bat | Usage: asc2mpg
infile.asc outfile.mpg |
Real-time x2 |
| Do drift correction | software/JASTEDF2MPG/ffdrift.exe | Usage: ffdrift.exe outfile.mpg --ascfile
infile.asc --offfile
offsets.driftoffsetsThen apply driftoffset file using reformatasc.py ascfile.asc -d offsets.driftoffsets [options] |
- |
| Convert data to GDF or ELAN format | software/JASTAnalyser | Usage: [Either use dist/asc2gdf.exe or asc2gdf.py -
see allasc2elan .bat and.asc2elan.bat
files for details] asc2gdf server.asc client.asc
outfile.gdf asc2elan server.asc client.asc outfile.elan
If you omit a filename, a dialog box asking for 3 filenames will appear. |
Overnight |
| Convert data to NXT format (requires GDF data) | software/gdf2nxt | Usage: gdf2nxt [options] [-t tiers] [-o
observations]
[input_path] will output into "codings" subdirectory of current directory, overwriting any files already there. See the data module in CVS for examples of metadata files that handle this data. Requires a conditions.csv file to specify trial and dyad conditions for import into NXT. See gdf2nxt documentation for further information. |
Real-time x2 |
| Analyse and export data from NXT to tab-delimited format. | software/gdf2nxt/FunctionQuery (also available in NXT) | FunctionQuery usage JAST NXT sample queries JAST NXT metadata structure NXT documentation |
Seconds to hours depending on query complexity |
| Perform cross-recurrence analysis (requires GDF data) | software/JASTAnalyser/cross-correlation | Usage: cross-correlation [options]
infile.gdf [> outfile.csv]Further information is available on Eye-Eye Lag Analysis and Cross Recurrence Programming. |
Max: Analysis window width x trials per file |
| Split cross-correlation CSV data into conditions | software/JASTAnalyser/reformat-cc | Usage: reformat-cc [options] infile.csv
Will create 8 new csv files based on condition information in conditions.csv. These 8 files correspond to the 8 combinations of the 3 boolean variables representing whether eye gaze is displayed, whether mouse position is displayed and whether speech is allowed, and are only valid for jcte2. See also Eye-Eye Lag Analysis and Specification of the Experimental Conditions for Each Trial. |
< 10 Minutes |
|
Add timings to NXT named-entities / referring expressions (Requires wavesurfer) |
software/gdf2nxt/n2w.sh Use wavesurfer with JAST-NE.conf, speech.wav and output files. software/gdf2nxt/w2n.sh (may need to be run twice) |
Download Wavesurfer | < 2 minutes for conversion + transcriber time |
