Adaptive Filter Speech Intelligibility Measurement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Ensuring accurate estimation and improvement of speech intelligibility from a loudspeaker in noisy environments, such as vehicles or indoor spaces, is challenging due to interference from various noise sources, and existing systems face limitations in sensitivity, computational complexity, and performance under low loudspeaker-power constraints.
Innovation Solution
A system using a microphone or microphone array with an adaptive filter to estimate the clean speech signal and measure background noise, applying a spectral mask to optimize speech intelligibility while maintaining signal distortion within prescribed levels, and employing a multi-microphone loudspeaker array for uniform intelligibility across an enclosure.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If adaptive filter is used to estimate clean speech signal, then measurement precision of speech intelligibility is improved, but device complexity increases
Solution Approach 1:
An adaptive filter is introduced as an intermediary component to estimate the clean speech signal from the noisy microphone input. The filter processes the signal through configurable tap delays and feedback coefficients, serving as a mediator between the noisy measurement and the desired clean signal representation, thereby improving measurement accuracy without requiring complete system redesign
Solution Approach 2:
The system adjusts parameters such as the number of tap delays, feedback coefficients, and filter coefficients to optimize the estimation accuracy. By changing these parameters based on the specific acoustic environment and noise conditions, the system achieves high measurement precision while keeping the overall device complexity manageable through parameter optimization rather than structural expansion
2Measurement precision
If spectral mask is applied to modify loudspeaker signal, then speech intelligibility is improved, but signal distortion increases
Solution Approach 1:
The spectral mask applies different processing characteristics to different frequency bands of the loudspeaker signal. By analyzing the spectrum and applying selective gains or filters to specific frequency ranges, the system improves speech intelligibility in the most critical frequency bands while minimizing distortion in other bands, achieving local optimization of quality
Solution Approach 2:
The spectral mask is dynamically adjusted based on the detected noise characteristics and speech content. The system adapts the mask parameters in real-time to balance intelligibility enhancement with distortion control, allowing the system to respond to changing acoustic conditions and maintain optimal performance without introducing excessive fixed distortion
3Adaptability or versatility
If multi-microphone array is used to improve uniform intelligibility, then adaptability to different positions is improved, but device complexity increases
Solution Approach 1:
The multi-microphone array divides the enclosure into different spatial zones and processes each zone independently. By segmenting the acoustic field and applying localized processing to each microphone channel, the system achieves uniform intelligibility across different positions while managing complexity through modular, independent channel processing rather than requiring a monolithic complex system
Solution Approach 2:
The microphone array system is designed to serve multiple functions: capturing acoustic signals, estimating speech intelligibility, characterizing noise fields, and guiding signal modification. This multi-functionality allows a single array system to address various acoustic challenges simultaneously, improving adaptability without proportionally increasing complexity
Data Source
AI summary
A method for accurately estimating and improving the speech intelligibility from a loudspeaker (LS) is disclosed. A microphone is placed in a desired position and using an adaptive filter, an estimate of the clean speech signal at the microphone is generated. By using the adaptive-filter estimate of the clean speech signal and measuring the background noise in the enclosure an accurate Speech Intelligibility Index (SII) or Articulation Index (AI) measurement at the microphone position is obtained. On the basis of the estimated speech intelligibility measurement, a decision can be made if the LS signal needs to be modified to improve the intelligibility.