An analog integrated-circuit vocal tract

Area of Science:

• Electrical Engineering
• Acoustics
• Speech Processing

Background:

• The human vocal tract is a complex biomechanical system responsible for speech production.
• Previous models often lacked the efficiency and real-time processing capabilities for advanced applications.
• Integrated circuits offer a pathway to miniaturize and enhance the functionality of vocal tract models.

Purpose of the Study:

• To develop the first experimental analog integrated-circuit vocal tract.
• To enable efficient and physiologically accurate speech synthesis.
• To explore applications in noise-robust speech recognition.

Main Methods:

• Mapping fluid dynamics to electrical circuit parameters (volume velocity to current, pressure to voltage).
• Implementing a 16-stage cascade of two-port pi-elements for a tunable transmission line.
• Utilizing a nonlinear resistor to model vocal tract airflow, including laminar and turbulent flow.
• Integrating a current source to simulate the glottal source.

Main Results:

• Achieved high signal-to-noise ratios (SNR) of 64-66 dB for formant resonances.
• Demonstrated the synthesis of speech signals using compact, linearly interpolatable articulatory parameters.
• The analog vocal tract chip operates at a low power consumption of 275 muW.

Conclusions:

• The analog vocal tract chip represents a significant advancement in speech synthesis and processing.
• Its physiological model and electrical implementation offer potential for noise-robust speech recognition systems.
• This technology paves the way for more efficient and compact speech-related electronic devices.