Audio Encoder Bandwidth Control to Avoid Spectral Holes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Audio codecs face issues with perceptual degradation and audible artefacts due to incorrect bandwidth detection, leading to spectral holes and unnecessary noise generation in bandlimited audio signals, especially in speech signals.
Innovation Solution
An encoder apparatus with a bandwidth detector and controller that selectively configures a subgroup of frequency domain encoder tools to operate at different bandwidths based on signal characteristics, ensuring that only active frequency regions are processed, thereby avoiding spectral holes and reducing noise generation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If bandwidth detection is used to control coding tools, then audio quality is improved by avoiding artificial noise in inactive regions, but spectral holes may occur when detection is incorrect
Solution Approach 1:
The coding tools are divided into two groups: guided tools (TNS, noise filling) that are controlled by bandwidth detection results, and unguided tools (spectral quantizer, residual coder) that operate independently. This segmentation allows the guided tools to avoid spectral holes by operating only in active bands, while unguided tools maintain full bandwidth operation to ensure reliability even when detection is incorrect.
Solution Approach 2:
Different bandwidth configurations are applied to different groups of coding tools based on their specific requirements. Guided tools use bandwidth-adaptive configuration to improve local audio quality in active regions, while unguided tools maintain full bandwidth configuration to ensure overall spectral completeness and avoid local degradation.
2Object-generated harmful factors
If all coding tools are controlled by bandwidth detection, then noise generation is reduced in inactive regions, but side information requirements increase significantly
Solution Approach 1:
The set of coding tools is segmented into guided and unguided groups. Only the guided tools require bandwidth control information, reducing the amount of side information needed compared to controlling all tools. The unguided tools operate without additional bandwidth information, further reducing side information requirements.
Solution Approach 2:
The bandwidth detection mechanism serves multiple functions: it controls the guided coding tools to avoid noise generation, while the unguided tools provide a fallback mechanism that ensures reliable operation without requiring additional bandwidth information, effectively making the system robust without increasing side information overhead.
3Device complexity
If hard switching between bandwidth modes is used, then device complexity is reduced, but audible artefacts occur due to rectangular spectral holes
Solution Approach 1:
The coding tools are segmented such that guided tools can be hard-switched between bandwidth modes for simplicity, while unguided tools continuously operate at full bandwidth. This segmentation prevents the rectangular spectral holes that would result from hard-switching all tools, as the unguided tools fill in the spectral gaps.
Solution Approach 2:
Different switching strategies are applied to different tool groups: guided tools use hard switching for simplicity, while unguided tools maintain continuous full-bandwidth operation. This creates a hybrid approach where local quality is maintained in the unguided tool output, preventing audible artefacts while keeping overall device complexity low.
Data Source
Figure 1~2
Figure 3
Figure 3a
AI summary
There are provided examples for encoding and/or decoding information signals (e.g., audio signals). In one example, there is provided an encoder apparatus comprising: a plurality of frequency domain, FD, encoder tools for encoding an information signal, the information signal presenting a plurality of frames; and an encoder bandwidth detector and controller (39) configured to select a bandwidth for at least a subgroup (33, 36) of the plurality of FD encoder tools, the subgroup (33, 36) including less FD encoder tools than the plurality of FD encoder tools, on the basis of information signal characteristics so that at least one of the FD encoder tools of the subgroup (33, 36) has a different bandwidth with respect to at least one of the FD encoder tools which are not in the subgroup (33, 36). In one example, there is provided a decoder apparatus (40, 40a) comprising a plurality of FD decoder tools (43-48a) for decoding an information signal encoded in a bitstream, wherein: the FD decoder tools are divided: - in a subgroup comprising at least one FD decoder tool (43, 45); - in remaining FD decoder tools comprising at least one FD decoder tool (44, 46, 48a); wherein the decoder apparatus (40, 40a) is configured so as to choose a bandwidth for at least one of the plurality of decoder tools of the subgroup (43, 45) on the basis of bandwidth information included in the bitstream so that the at least one of the plurality of decoder tools of the subgroup (43, 45) performs signal processing a different bandwidth with respect to at least one of the remaining FD decoder tools of the plurality of decoder tools (44, 46, 48a).