Frequency Translation by Spectral Envelope Warping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing hearing assistance devices face challenges in improving speech intelligibility for individuals with hearing impairments without distorting the natural sound quality, as current frequency translation methods often alter the pitch of sounds and introduce artifacts, especially in noisy environments.
Innovation Solution
The system employs high-frequency spectral envelope warping by estimating line spectral frequencies, applying warping functions, and scaling the transposed audio signal, while maintaining the low-frequency portion unaltered, using a combination of time and frequency domain processing to translate pole frequencies linearly or non-linearly, and randomizing the phase of the prediction error signal to reduce artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If frequency translation processing is applied to recode high-frequency sounds at lower frequencies, then speech intelligibility is improved for hearing-impaired individuals, but the pitch of the signal is altered and the natural sound quality is degraded
Solution Approach 1:
The audio signal is divided into multiple frequency bands using a filter bank, with only high-frequency bands (above a threshold frequency) subjected to frequency translation processing. Low-frequency bands are preserved without transposition, maintaining their original pitch and natural sound quality. This segmentation allows selective processing that improves speech intelligibility while minimizing pitch alteration artifacts.
Solution Approach 2:
Different processing strategies are applied to different frequency regions: high-frequency bands undergo frequency translation to improve intelligibility for hearing-impaired users, while low-frequency bands are left unprocessed to preserve natural pitch and sound quality. This local differentiation resolves the contradiction by applying frequency translation only where it provides benefit without degrading overall sound quality.
2Adaptability or versatility
If the entire signal is represented by a formant model and resynthesized after transposition, then frequency translation can be performed, but the entire signal is considerably altered and pitch is changed
Solution Approach 1:
Only the high-frequency portion of the signal is extracted and processed through frequency translation, while the low-frequency portion is excluded from transposition. The processed high-frequency components are then combined with the unprocessed low-frequency components. This extraction approach allows frequency translation capability to be applied selectively without altering the entire signal, thereby preserving pitch and reducing distortion.
3Ease of operation
If variable-rate playback mechanism is used for frequency transposition, then transposition can be implemented, but the entire spectrum is transposed and pitch is altered
Solution Approach 1:
The frequency spectrum is segmented into multiple bands, and frequency translation is applied only to high-frequency bands using a filter bank approach. This avoids the spectrum-wide pitch alteration inherent in variable-rate playback methods, as only the necessary high-frequency components are transposed while low-frequency components maintain their original pitch.
4Adaptability or versatility
If nonlinear frequency transposition function is applied, then frequency translation is achieved, but harmonic structures become inharmonic and artifacts are introduced
Solution Approach 1:
A linear frequency translation function is applied specifically to high-frequency bands rather than attempting complex nonlinear transposition across the entire spectrum. This partial action approach achieves the necessary frequency shifting for speech intelligibility while avoiding the inharmonic artifacts that result from nonlinear transposition functions, as linear translation preserves the harmonic relationships within each band.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Disclosed herein, among other things, is a system for frequency translation by high-frequency spectral envelope warping in hearing assistance devices. The present subject matter relates to improved speech intelligibility in a hearing assistance device using frequency translation by high-frequency spectral envelope warping. The system described herein implements an algorithm for performing frequency translation in an audio signal processing device for the purpose of improving perceived sound quality and speech intelligibility in an audio signal when presented using a system having reduced bandwidth relative to the original signal, or when presented to a hearing-impaired listener sensitive to only a reduced range of acoustic frequencies.