Dynamic controller for improving speech intelligibility
View Patent ↗A system improves the speech intelligibility and the speech quality of a speech segment. The system includes a dynamic controller that detects a background noise from an input by modeling a signal. A variable gain amplifier adjusts the variable gain of the amplifier in response to an output of dynamic controller. A shaping filter adjusts a speech signal by tilting portions of the speech signal of the dynamic controller.
1. A system that improves the speech intelligibility and the speech quality of a speech signal comprising:
a dynamic controller that is adapted to detect a background noise from an input comprising multiple signals by modeling, where the dynamic controller comprises:
a voice activity detector configured to identify which of the multiple signals comprises a speech signal, an unvoiced signal, or the background noise; and
a coherence estimator configured to estimate the spectral coherence between the unvoiced signal or the background noise and the speech signal by quantifying a quality of interference between the unvoiced signal or the background noise with the speech signal;
a variable gain amplifier adapted to adjust the variable gain in response to an output of dynamic controller; and
a shaping filter that adjusts the speech signal by tilting portions of the speech signal in response to the signal modeling of the dynamic controller.
2. The system that improves speech intelligibility and the speech quality of a speech signal of claim 1 where the modeling comprises approximating a linear relationship.
3. The system that improves speech intelligibility and the speech quality of a speech signal of claim 2 where the modeling comprises approximating an inverse linear relationship.
4. The system that improves speech intelligibility and the speech quality of a speech signal of claim 1 where the modeling comprises approximating an inverse linear relationship.
5. The system that improves speech intelligibility and the speech quality of a speech signal of claim 1 where the process comprises approximating a non-linear relationship.
6. The system that improves speech intelligibility and the speech quality of a speech signal of claim 1 where the dynamic controller further comprises:
a spectral estimator configured to estimate the signal-to-noise ratio of at least one of the multiple signals; and
an articulation estimator configured to measure the intelligibility of an output of the spectral estimator.
7. The system that improves speech intelligibility and the speech quality of a speech signal of claim 6 where the articulation estimator generates an articulation index that measures the intelligibility of the output from the spectral estimator, where the shaping filter adjusts the speech signal based on a comparison of the articulation index to a plurality of predetermined thresholds.