An instrumental intelligibility metric based on information theory
This is an incremental improvement for speech processing researchers and engineers, offering a more accurate tool for evaluating speech quality.
The paper tackled the problem of measuring speech intelligibility by proposing SIIB, a metric that estimates shared information in bits per second, and found it highly correlated with intelligibility for noise-degraded and enhanced speech.
We propose a monaural intrusive instrumental intelligibility metric called speech intelligibility in bits (SIIB). SIIB is an estimate of the amount of information shared between a talker and a listener in bits per second. Unlike existing information theoretic intelligibility metrics, SIIB accounts for talker variability and statistical dependencies between time-frequency units. Our evaluation shows that relative to state-of-the-art intelligibility metrics, SIIB is highly correlated with the intelligibility of speech that has been degraded by noise and processed by speech enhancement algorithms.