High Frequency Compression for Speech Intelligibility
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Communication systems, especially wireless ones, face bandwidth limitations that lead to high frequency signal attenuation, affecting speech clarity, and existing solutions like speech/silence and voiced/unvoiced switches are unreliable and create artifacts.
Innovation Solution
A speech enhancement system that identifies and compresses speech segments, shifting them to specific frequency bands with adaptive gain adjustments to improve perceptual quality, and can interface with any bandlimited device, including wireless communication systems, to enhance speech before it is passed to another system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high frequency signals are passed through the communication system, then speech clarity is improved, but bandwidth limitations cause high frequency attenuation and signal loss
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting the compression ratio and frequency mapping based on the input signal characteristics. The system transforms high frequency components to lower frequencies using adjustable parameters, allowing optimization of speech clarity while working within bandwidth limitations. This resolves the contradiction by changing the frequency domain parameters adaptively rather than using fixed filtering.
2Loss of energy
If speech/silence and voiced/unvoiced switches are used to process unvoiced speech, then high frequency attenuation is overcome, but the system becomes unreliable and creates artifacts
Solution Approach 1:
The patent extracts only the necessary high frequency components that need enhancement and processes them separately through frequency compression. Instead of using switches that attempt to identify and process entire speech segments, the system selectively extracts and compresses high frequency content, avoiding the reliability issues of transition detection while still compensating for high frequency attenuation.
Solution Approach 2:
The frequency compression device acts as an intermediary that transforms high frequency components to lower frequencies without requiring speech/silence or voiced/unvoiced detection. This intermediary approach bypasses the unreliable switch-based processing while still achieving the goal of preserving high frequency speech information.
3Loss of energy
If high frequency components are compressed to lower frequency bands, then bandwidth limitations are overcome, but speech naturalness may be affected
Solution Approach 1:
The patent applies local quality by applying frequency compression selectively to specific high frequency bands rather than uniformly across the entire spectrum. Different compression ratios and mapping functions are applied to different frequency regions, preserving the natural characteristics of speech while efficiently utilizing the available bandwidth. This localized approach minimizes distortion while maximizing bandwidth efficiency.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A speech enhancement system that improves the intelligibility and the perceived quality of processed speech includes a frequency transformer and a spectral compressor. The frequency transformer converts speech signals from the time domain to the frequency domain. The spectral compressor compresses a pre-selected portion of the high frequency band and maps the compressed high frequency band to a lower band limited frequency range.