Audio Encoder Separating Impulse and Residual Signals at Low Bitrates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current audio coding technologies fail to achieve optimal quality for both general audio and speech signals, particularly at low bitrates, due to the inability of general audio coders to exploit speech source models and the limitations of speech coders in handling music signals, leading to inefficient bit allocation and increased power consumption in mobile devices.
Innovation Solution
A unified audio coding approach that separates audio signals into impulse-like and stationary portions, using an impulse coder for speech-like components and a filterbank-based coder for residual signals, allowing for adaptive bitrate control and efficient encoding of both types of signals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a general audio coder (perceptual audio coding) is used to encode speech signals at very low data rates, then the coding algorithm can shape spectral distortion according to masking threshold curves, but the coding quality is insufficient compared to dedicated LPC-based speech coders
Solution Approach 1:
The audio signal is segmented into two distinct components: impulse-like portions (speech) and stationary portions (music/general audio). Each segment is processed by a specialized coder optimized for its characteristics, with a switch selecting the appropriate coding path based on signal type detection.
Solution Approach 2:
The encoding system is designed to handle multiple signal types (speech and general audio) within a single unified framework. The switchable architecture allows the same system to adaptively select between LPC-based coding for speech and perceptual audio coding for music, achieving universal applicability across different audio content.
2Productivity
If an LPC-based speech coder is used to encode general music signals, then the coding algorithm can efficiently model speech production, but the coding quality is insufficient due to inability to shape spectral envelope according to masking threshold
Solution Approach 1:
The system dynamically adapts its coding approach based on the characteristics of the input signal. A detection mechanism identifies whether the current signal segment is speech-like or music-like, and the switch accordingly selects the optimal coding algorithm, enabling dynamic optimization of coding quality for varying audio content.
Solution Approach 2:
The coding parameters and algorithm selection change based on the detected signal characteristics. For speech signals, LPC parameters are used; for music signals, perceptual coding parameters are applied. This parameter adaptation allows the system to optimize coding quality for each signal type independently.
3Manufacturing precision
If separate speech and audio coding systems are used, then each can be optimized for its specific signal type, but the device complexity increases and power consumption rises in mobile devices
Solution Approach 1:
Two previously separate coding systems (LPC speech coder and perceptual audio coder) are merged into a single unified encoding device. A switch and detection mechanism coordinate the two coders, allowing them to share common infrastructure while maintaining specialized functionality, thereby reducing overall system complexity compared to maintaining completely separate systems.
Solution Approach 2:
The encoding system automatically detects the signal type and self-selects the appropriate coding algorithm without external intervention. This self-service capability eliminates the need for complex manual configuration or external control systems, reducing operational complexity while maintaining optimized coding quality.
4Manufacturing precision
If separate speech and audio coding systems are used, then each can be optimized for its specific signal type, but power consumption increases in mobile devices
Solution Approach 1:
The power consumption is segmented and optimized by processing only the necessary signal components with the appropriate coder. Since the switch directs speech signals to the LPC coder and music signals to the perceptual coder, each processor operates at full efficiency only when needed, avoiding wasted computational power and reducing overall energy consumption compared to running both systems continuously.
Solution Approach 2:
The system discards the output of one coder when the other is selected, avoiding redundant processing. The switch mechanism ensures that only the necessary coding path is active at any given time, eliminating wasted energy on processing that would not contribute to the final output, thereby reducing power consumption in mobile devices.
Data Source
AI summary
An audio encoder for encoding an audio signal includes an impulse extractor for extracting an impulse-like portion from the audio signal. This impulse-like portion is encoded and forwarded to an output interface. Furthermore, the audio encoder includes a signal encoder which encodes a residual signal derived from the original audio signal so that the impulse-like portion is reduced or eliminated in the residual audio signal. The output interface forwards both, the encoded signals, i.e., the encoded impulse signal and the encoded residual signal for transmission or storage. On the decoder-side, both signal portions are separately decoded and then combined to obtain a decoded audio signal.