The TongueTwister system supports the running of courses in correct pronunciation for students of English. The methodology for teaching words is based on the SuperMemo algorithm. Courses are run using the Moodle environment.
At the stage of preparation of the course, the system supports the collection and management of audio files using the MultiBaza module, which was developed by us.
The system architecture is shown in the diagram below.
Researchers from the Faculty of English at Adam Mickiewicz University planned an experiment to determine how the method of palatography might help in teaching the pronunciation of English words. This involves examination of the position of the articulatory organs – the tongue, lips and palate – using a special inserted device fitted with sensors. The university purchased the necessary equipment from two suppliers – from the UK (a slightly older version) and from Germany (a newer solution) – but neither of them was able to provide the software required to analyse the results.
For the two different types of testing equipment, we built software to compare the readings from the sensors with pronunciation models. For this purpose we applied advanced algorithms for measuring the similarity of objects. We built this solution into an e-learning system for foreign language teaching. Our solution lets students learn the pronunciation of new words, controlling the correctness of their speech by comparing the behaviour of their articulatory organs with model patterns.
Since it was deployed, we have supported the system on a pro bono publico basis, enabling the continuation of the experiments and constant expansion of the set of pronunciation models.
If you are seeking solutions for innovative research projects, it’s likely that we can help you. Let’s talk about it!