Gain Factor Limiting for Wideband Speech Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wideband speech coding techniques are computationally intensive and inefficient, particularly in mobile applications, as they require significant processing cycles and often lead to increased bandwidth, making them impractical for transmitting high-quality voice communications over narrowband channels without transcoding.
Innovation Solution
A method for wideband speech processing that calculates a gain factor based on the relation between subbands of a speech signal, allowing for efficient encoding and decoding by selecting appropriate quantization indices, enabling the transmission of narrowband and highband signals through narrowband channels without significant modification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If narrowband speech coding technique is scaled to cover wideband spectrum, then wideband speech quality is improved, but computational complexity increases significantly
Solution Approach 1:
The speech signal is divided into narrowband and highband portions that are coded separately. The narrowband portion (0-4 kHz) is encoded using conventional techniques, while the highband portion (4-8 kHz) is encoded separately with reference to the narrowband excitation signal, reducing overall computational complexity while maintaining wideband quality
Solution Approach 2:
The narrowband excitation signal serves as an intermediary to derive the highband excitation signal. By using the narrowband excitation as a reference and applying a highband gain factor, the system avoids direct complex processing of the entire wideband signal while still achieving highband reconstruction
2Measurement precision
If entire wideband spectrum is encoded to desired quality, then speech quality is improved, but bandwidth increases unacceptably
Solution Approach 1:
The highband portion of the speech signal is extracted and encoded separately from the narrowband portion. By encoding only the highband components with reference to the narrowband excitation rather than encoding the entire wideband signal independently, bandwidth usage is reduced while maintaining wideband quality
Solution Approach 2:
Instead of fully encoding the entire wideband spectrum independently, the system applies partial encoding to the highband portion using reference to the narrowband excitation signal. This partial action approach achieves acceptable wideband quality with reduced bandwidth requirements
3Measurement precision
If wideband speech coding is implemented, then speech quality is improved, but processing efficiency decreases
Solution Approach 1:
The speech coding process is segmented into narrowband and highband processing stages. The narrowband processing uses efficient conventional techniques, while the highband processing leverages the narrowband excitation reference, improving overall processing efficiency while maintaining wideband quality
Solution Approach 2:
The highband encoding process serves itself by using the narrowband excitation signal as a reference. This self-service approach eliminates the need for separate full wideband processing, improving processing efficiency while achieving wideband reconstruction
Data Source
AI summary
The range of disclosed configurations includes methods in which subbands of a speech signal are separately encoded, with the excitation of a first subband being derived from a second subband. Gain factors are calculated to indicate a time-varying relation between envelopes of the original first subband and of the synthesized first subband. The gain factors are quantized, and quantized values that exceed the pre-quantized values are re-coded.