Audio Encoder Sinusoid Injection for Tonal Signal Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current audio codecs lack flexibility in encoding efficient and accurate approximations of audio signals, particularly performing poorly with tonal input signals that do not have a noise-like spectrum.
Innovation Solution
An encoder and decoder system that defines and processes single frequency components into subsets based on frequency and perceptual importance, using sinusoids to improve coding efficiency by selecting and representing key components, and generating indicators for these components to enhance encoding and decoding processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If spectral-band-replication (SBR) coding is used to encode high frequency regions by transposing low frequency regions, then bit rate savings are achieved, but coding accuracy deteriorates for tonal signals
Solution Approach 1:
The patent applies dynamics by adaptively switching between different coding modes (noise-like spectrum mode and tonal spectrum mode) based on the characteristics of the input signal. The encoder dynamically selects whether to use SBR coding or sinusoid injection coding, and adjusts parameters such as the number of sinusoids and their frequencies according to the signal's spectral properties, thereby optimizing both bit rate efficiency and coding accuracy for different signal types
Solution Approach 2:
The patent changes parameters by introducing sinusoid injection as an additional coding parameter alongside traditional SBR coding. By controlling the number, frequency, and amplitude of injected sinusoids based on spectral analysis, the system transforms the coding approach from purely parametric (SBR) to a hybrid method that can accurately represent both noise-like and tonal components, resolving the accuracy issue for tonal signals while maintaining bit rate efficiency
2Device complexity
If a fixed QMF filter bank with 32 bands is used for transposition, then the encoding process is simplified, but adaptability to different signal characteristics is reduced
Solution Approach 1:
The patent makes the encoding process dynamic by introducing adaptive elements that respond to signal characteristics. Although the QMF filter bank structure remains fixed, the system dynamically adjusts the coding strategy (SBR vs. sinusoid injection), the number of sinusoids to inject, their frequencies, and amplitudes based on spectral analysis. This dynamic adaptation allows the fixed filter bank to effectively handle diverse signal types without increasing its structural complexity
Solution Approach 2:
The patent segments the frequency spectrum into different regions and applies different coding strategies to different segments. By analyzing the spectral characteristics and identifying tonal components at specific frequencies, the system selectively applies sinusoid injection only where needed rather than uniformly across all bands. This segmented approach improves adaptability to different signal characteristics while avoiding unnecessary complexity in regions where simple SBR coding suffices
Data Source
AI summary
An apparatus including at least one processor and at least one memory including computer program code the at least one memory and the computer program code configured to, with the at least one processor, cause the apparatus at least to select at least two single frequency components; generate an indicator, the indicator being configured to represent the at least two single frequency components and is configured to be dependent on the frequency separation between the two single frequency components.