Speech Decoder Temporal Envelope Shaping for SBR
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current speech encoding/decoding systems face challenges in accurately representing the temporal envelope of high frequency components, leading to pre-echo and post-echo artifacts, especially in bandwidth extension techniques like SBR, which can degrade the subjective quality of decoded signals without significantly reducing bit rate.
Innovation Solution
The system employs a method to analyze and quantify high frequency components, calculating temporal envelope supplementary information based on low frequency components, and multiplexing this information into a bit stream to be transmitted, allowing for improved representation of high frequency temporal envelopes without increasing bit rate, using a processor-based device with units for encoding and decoding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If temporal envelope supplementary information is calculated and multiplexed into the bit stream, then the temporal envelope of high frequency components is accurately represented, but the bit rate increases
Solution Approach 1:
The patent extracts only the essential temporal envelope supplementary information from the high frequency components rather than transmitting the complete high frequency signal. By calculating temporal envelope characteristics (such as envelope shape parameters, temporal variation indicators) and transmitting only these extracted features, the system achieves accurate temporal envelope representation while minimizing the additional bit rate requirement.
Solution Approach 2:
The patent applies different processing and information transmission strategies to different frequency bands. The low frequency components are encoded with standard codecs while the high frequency components use bandwidth extension techniques with selective temporal envelope information. This local differentiation allows optimized resource allocation where temporal envelope information is added only where needed (high frequencies) rather than uniformly across the entire signal spectrum.
2Productivity
If bandwidth extension techniques like SBR are used, then high frequency components are generated from low frequency components, but pre-echo and post-echo artifacts occur
Solution Approach 1:
The patent performs preliminary analysis of the temporal envelope characteristics of the original high frequency components before generating the bandwidth-extended signal. By calculating temporal envelope supplementary information (such as envelope shape parameters, temporal variation indicators) in advance and using this information to guide the bandwidth extension process, the system prepares the necessary control data beforehand to prevent pre-echo and post-echo artifacts from occurring during signal generation.
Solution Approach 2:
The patent uses the calculated temporal envelope supplementary information as feedback to adjust and refine the bandwidth extension process. The temporal envelope characteristics derived from the original signal are fed back into the high frequency generation process to constrain and shape the synthesized high frequency components, ensuring they match the temporal envelope of the original signal and thereby eliminating pre-echo and post-echo artifacts.
Data Source
AI summary
A linear prediction coefficient of a signal represented in a frequency domain is obtained by performing linear prediction analysis in a frequency direction by using a covariance method or an autocorrelation method. After the filter strength of the obtained linear prediction coefficient is adjusted, filtering may be performed in the frequency direction on the signal by using the adjusted coefficient, whereby the temporal envelope of the signal is shaped. This reduces the occurrence of pre-echo and post-echo and improves the subjective quality of the decoded signal, without significantly increasing the bit rate in a bandwidth extension technique in the frequency domain represented by SBR.