LC-SBC Codec with Voice Activity Detection for Bandwidth Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing Low Complexity Sub-band Coder (LC-SBC) used in Bluetooth voice communication systems does not effectively reduce bandwidth usage and power consumption, as it lacks features like Voice Activity Detection (VAD) and Variable Rate encoding, which are essential for efficient voice compression in mobile communication.
Innovation Solution
An augmented version of LC-SBC is developed that incorporates Voice Activity Detection (VAD) and Comfort Noise Generation (CNG) techniques to differentiate between voice and noise frames, reducing bit rate by encoding indices of previously processed sub-band samples for noise frames, thereby minimizing bandwidth and power consumption without modifying the underlying logic or structure of LC-SBC.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If LC-SBC is used for wideband speech communication, then bandwidth usage is reduced compared to full audio coding, but voice activity detection and variable rate encoding features are missing, resulting in higher average bit rate and power consumption
Solution Approach 1:
The audio signal is divided into individual frames, and each frame is independently processed to determine whether it contains speech or noise. This segmentation allows the system to apply different encoding strategies to different frames, reducing overall bit rate while maintaining voice quality.
Solution Approach 2:
The codec dynamically adjusts the encoding mode based on the detected voice activity in each frame. When voice activity is detected, full encoding is applied; when noise is detected, reduced encoding or index-based representation is used. This dynamic adaptation reduces average power consumption without compromising speech quality.
2Quantity of substance
If voice activity detection is implemented to reduce bandwidth usage, then average bit rate decreases, but computational complexity increases
Solution Approach 1:
Voice activity detection is performed preliminarily on each frame before encoding to determine the appropriate encoding strategy. This preliminary classification allows the subsequent encoding process to be optimized, reducing the computational burden during actual data transmission.
Solution Approach 2:
The system changes the encoding parameters based on voice activity detection results. For noise frames, the system uses index-based representation instead of full sample encoding, significantly reducing the number of bits transmitted and the computational complexity of the encoding process.
3Reliability
If noise frames are encoded using full sub-band samples, then voice quality is maintained, but bandwidth consumption increases unnecessarily
Solution Approach 1:
Instead of encoding noise frames using full sub-band samples, the system creates a compact index that references previously stored noise frame samples. This copying approach allows the receiver to reconstruct noise frames using stored references, maintaining voice quality while significantly reducing bandwidth consumption.
Solution Approach 2:
The system discards redundant information in noise frames by using index-based representation instead of full sample encoding. The necessary information is recovered at the receiver end by referencing stored noise frame data, achieving compression without loss of quality.
Data Source
AI summary
A system and method for providing an augmented version of a Low-Complexity Sub-band Coder (LC-SBC) is described herein. In accordance with the method, a series of input audio samples representative of the frame are received. A series of sub-band samples is generated for each of a plurality of frequency sub-bands based on the input audio samples. A determination is made as to whether the frame is a voice frame or a noise frame. Responsive to a determination that the frame is a noise frame, an index representative of a previously-processed series of sub-band samples stored in a history buffer for at least one of the frequency sub-bands is encoded instead of encoding the series of sub-band samples generated for the frequency sub-band.