Bandwidth Extension of Narrowband Speech via Spectral Envelope Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing communication systems face limitations in extending audio bandwidths, particularly in noisy environments, where traditional methods fail to accurately reconstruct speech due to difficulty in modeling noise, leading to reduced intelligibility and distortion.
Innovation Solution
A system that extends narrowband speech signals into wideband spectra by generating a high-frequency spectrum and background noise spectrum, combining them using a summing circuit, with components like high-band generators, background noise generators, and parameter detectors to adjust spectral envelopes and phases, ensuring natural-sounding speech even in noisy conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional bandwidth extension methods are used to reconstruct speech, then speech quality is improved under ideal conditions, but intelligibility and quality decline in noisy environments due to difficulty in modeling noise effects
Solution Approach 1:
The system segments the speech signal processing into distinct frequency bands (low-band from 0-4kHz and high-band from 4-8kHz). The low-band generator processes the narrowband input separately from the high-band extension, allowing independent optimization of noise handling in each band. This segmentation enables the system to apply different processing strategies to different frequency regions, improving overall reliability in noisy conditions.
Solution Approach 2:
The patent introduces an intermediary spectral envelope model that bridges the narrowband input and wideband output. The spectral envelope serves as a mediator that captures the essential characteristics of the speech signal while filtering out noise. By modeling and extending the spectral envelope rather than directly processing the raw signal, the system can reconstruct high-frequency content while maintaining robustness against noise contamination.
2Manufacturing precision
If new telecommunication networks supporting larger bandwidths are built, then wideband speech transmission quality is improved, but network cost and establishment time increase significantly
Solution Approach 1:
The system creates a synthesized copy of the high-frequency speech spectrum based on the available low-frequency information. Instead of transmitting actual wideband signals through upgraded networks, the invention generates a high-band copy (4-8kHz) by extrapolating from the low-band input (0-4kHz) using spectral envelope modeling. This copying approach allows wideband quality to be achieved through signal processing rather than network infrastructure changes.
Solution Approach 2:
The patent changes the parameter representation of the speech signal from time-domain waveforms to frequency-domain spectral envelopes. By transforming the signal representation and operating in the frequency domain, the system can extend the bandwidth virtually through parameter manipulation rather than physical transmission. This parameter-based approach enables bandwidth extension using existing narrowband infrastructure.
3Manufacturing precision
If bandwidth extension is performed to improve speech quality, then perceived quality is enhanced, but artifacts may be introduced and processing complexity increases
Solution Approach 1:
The system applies partial bandwidth extension by focusing computational resources on extending only the critical high-frequency range (4-8kHz) rather than attempting to reconstruct the entire spectrum. The spectral envelope modeling applies extension selectively to frequency regions that most impact perceived quality, avoiding excessive processing in less critical bands. This partial action approach balances quality improvement with processing complexity.
Data Source
AI summary
A system extends the bandwidth of a narrowband speech signal into a wideband spectrum. The system includes a high-band generator that generates a high frequency spectrum based on a narrowband spectrum. A background noise generator generates a high frequency background noise spectrum based on a background noise within the narrowband spectrum. A summing circuit linked to the high-band generator and the background noise generator combines the high frequency spectrum and narrowband spectrum and the high frequency background noise spectrum.