Audio Signal Classification Accuracy vs Complexity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio encoding technologies face challenges in accurately classifying speech and music signals, leading to inefficient coding modes and increased complexity, with previous solutions prone to errors and complexity.
Innovation Solution
An improved classification method involving an initial classification followed by a finer classification using parameters like voicing, modified correlation, and long-term pitch gain to accurately distinguish speech from music, reducing misclassification and complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If closed-loop mode decisions (e.g., AMR-WB+, USAC) are used, then classification accuracy is improved, but device complexity increases
Solution Approach 1:
The classification process is segmented into two distinct stages: an initial classification stage that provides a first classification decision, and a finer classification stage that refines the decision based on additional parameters. This segmentation allows the system to achieve high accuracy through the finer classification stage while the initial stage provides a baseline that reduces the burden on more complex processing, effectively managing overall system complexity.
2Device complexity
If open-loop mode decisions (e.g., AMR-WB+, EVRC-WB) are used, then device complexity is reduced, but classification accuracy deteriorates
Solution Approach 1:
The initial classification stage performs a preliminary classification action that provides a first classification decision before the finer classification stage. This preliminary action establishes a baseline classification that can be refined subsequently, allowing the system to achieve high accuracy without requiring all complex processing to occur simultaneously, thus balancing accuracy and complexity.
3Measurement precision
If finer classification is performed for all frames, then classification accuracy is improved, but processing time increases
Solution Approach 1:
The classification process is made dynamic by conditionally applying the finer classification stage only when necessary. The system determines whether to perform the more time-consuming finer classification based on the initial classification results and other frame characteristics, thereby achieving high accuracy when needed while minimizing processing time when the initial classification is sufficient.
4Measurement precision
If multiple parameters are used in finer classification, then classification accuracy is improved, but device complexity increases
Solution Approach 1:
Different levels of classification quality are applied locally to different parts of the processing pipeline. The initial classification stage uses a simpler set of parameters for a first classification decision, while the finer classification stage applies additional parameters (such as voicing, modified correlation, signal activity, and long term pitch gain) only when needed. This local differentiation of quality allows high accuracy where required while maintaining lower complexity elsewhere.
Data Source
Figure 1A~1B
Figure 2
Figure 3
AI summary
Improved audio classification is provided for encoding applications. An initial classification is performed, followed by a finer classification, to produce speech classifications and music classifications with higher accuracy and less complexity than previously available. Audio is classified as speech or music on a frame by frame basis. If the frame is classified as music by the initial classification, that frame undergoes a second, finer classification to confirm that the frame is music and not speech (e.g., speech that is tonal and/or structured that may not have been classified as speech by the initial classification). Depending on the implementation, one or more parameters may be used in the finer classification. Example parameters include voicing, modified correlation, signal activity, and long term pitch gain.