We have built a new speech recognition exercise to give our students immediate feedback on their pronunciation skills. Built on industry-standard technology created by researchers at Carnegie Mellon university, our speech recognizer compares a recording of your voice against both a native speaker and non-native speaker model. This means that the technology recognises a correct input from a variety of different accents, making it a robust and flexible tool for pronunciation training.
We are in the process of rolling out these exercises across our English and Spanish courses at the moment!
When you come across a speech recognition exercise, listen to the phrase and press the microphone button to record your attempt at reproducing the phrase.
If you can't talk aloud at that moment, you can skip the exercise by pressing "I can't speak at the moment". You can come back and revisit this exercise at any time.