It's About Time: Minimizing Hardware and Software Latencies in Speech Research With Real-Time Auditory Feedback

Area of Science:

• Speech production research
• Auditory-motor integration
• Human-computer interaction

Background:

• Auditory feedback perturbation paradigms are used to study speech production, auditory-motor integration, and learning.
• Real-time manipulation of acoustic speech output requires precise measurement of hardware and software latencies.
• Inconsistent latency reporting in existing literature hinders scientific reproducibility.

Purpose of the Study:

• To investigate and quantify hardware and software latencies in real-time auditory feedback perturbation systems.
• To identify factors contributing to total feedback loop latency.
• To provide recommendations for improving the accuracy and reproducibility of latency measurements in speech research.

Main Methods:

• Measured hardware and software latencies using the Audapter software for formant perturbation studies.
• Tested various audio interfaces, computers, and audio drivers.
• Evaluated the impact of nondefault Audapter parameter settings on software latency.

Main Results:

• Hardware latencies, often overlooked, were found to be significant (up to 15 ms) for certain audio interfaces.
• Total feedback loop latencies were frequently greater than reported in previous studies.
• Nondefault Audapter parameters reduced software latency without compromising formant tracking accuracy.

Conclusions:

• Audio interface choice and software parameter optimization critically influence total feedback loop latency.
• Accurate measurement and reporting of combined hardware and software latency are essential for scientific rigor.
• Minimizing feedback loop latency is recommended for future speech research utilizing real-time auditory feedback perturbations.