Figure 5
From: Optimizing Automatic Speech Recognition for Low-Proficient Non-Native Speakers
The distribution of phonetic distances between all response pairs to all questions.
Figure 5 | EURASIP Journal on Audio, Speech, and Music Processing