Audio Signal Decoding with Frame-Independent Context Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio and speech coding technologies face challenges in reducing redundant bit information and ensuring decoding efficiency, particularly when random access to audio streams is required, due to the reliance on previous frame information for accurate decoding.
Innovation Solution
The proposed solution involves a speech audio coding apparatus that selectively uses frequency domain coding or linear prediction domain coding in the low frequency band and enhanced spectral band replication in the high frequency band, combining tools from advanced audio coding and AMR-WB+ codecs, to reduce redundant bit information and enable decoding regardless of random access.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If context-based lossless coding is performed using previous frame information, then decoding precision is improved, but random access capability deteriorates
Solution Approach 1:
The patent segments the audio signal into independent frames, where each frame can be decoded independently using only its own quantization indices without requiring previous frame information. This segmentation enables random access to any frame in the audio stream while maintaining decoding precision through frame-independent context-based lossless coding.
Solution Approach 2:
The patent performs preliminary encoding of each frame's quantization indices during the encoding phase, storing them in a format that enables direct decoding without needing subsequent or previous frame data. This preliminary preparation of encoded data allows random access points throughout the audio stream to be decoded immediately.
2Measurement precision
If detailed fine structures are coded in the low frequency band, then audio quality is improved, but bit rate increases
Solution Approach 1:
The patent applies different coding strategies to different frequency bands: detailed fine structure coding is applied locally to the low frequency band where it is most needed for audio quality, while the high frequency band uses a more efficient coding approach that is less bit-intensive. This local differentiation of coding quality optimizes the balance between audio quality and bit rate.
Solution Approach 2:
The patent changes the coding parameters based on frequency band characteristics. In the low frequency band, parameters are set to preserve detailed fine structures with higher precision. In the high frequency band, parameters are adjusted to use more efficient compression, reducing bit rate while maintaining acceptable quality through perceptual coding principles.
3Productivity
If spectral band replication is used to generate high frequency signals, then bandwidth efficiency is improved, but decoding complexity increases
Solution Approach 1:
The patent uses spectral band replication to create a simplified copy of the low frequency signal spectrum and places it in the high frequency band. This copying approach efficiently generates high frequency content from low frequency information, improving bandwidth efficiency. The decoding complexity is managed by using a systematic copying process rather than requiring complex high frequency analysis and synthesis.
Data Source
AI summary
A method for coding and decoding an audio signal or speech signal and an apparatus adopting the method are provided.