Abnormal Frame Detection in Speech Signals via Energy and Singularity Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for detecting speech distortion in wireless communications are either costly and time-consuming, such as manual subjective tests, or lack timely automatic detection capabilities, which is essential for assessing speech quality during transmission.

Innovation Solution

An abnormal frame detection method that processes speech signals by dividing them into subframes, calculating local energy values, performing singularity analysis through wavelet decomposition, and determining frames as abnormal based on threshold criteria to identify distortion.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual subjective test method is used to assess speech quality, then measurement precision is improved, but loss of time and productivity deteriorate

Engineering Contradiction:
Improvespeech quality assessment accuracyVSAvoidtesting efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent replaces manual subjective testing with an automated computer-based detection system. The system uses digital signal processing techniques including energy calculation, zero-crossing rate analysis, and correlation analysis to automatically identify abnormal frames in speech signals, eliminating the need for human testers while maintaining assessment accuracy.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system performs self-testing by automatically analyzing speech signals without requiring external human evaluation. The automated detection algorithm independently assesses speech quality by processing audio data through multiple computational steps, including frame-based energy calculation and pattern recognition, enabling the system to evaluate itself without human intervention.

Inventive Principle:
Principle #25Self-service

2Productivity

If automated detection method is implemented, then productivity is improved, but measurement precision deteriorates

Engineering Contradiction:
Improvedetection speedVSAvoidspeech distortion detection accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent divides the speech signal into discrete frames and further segments each frame into multiple subframes. This segmentation allows the system to analyze local characteristics of different portions of the signal independently, improving both the speed of processing and the precision of distortion detection by identifying specific abnormal segments within the overall speech signal.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system calculates local energy values and zero-crossing rates for each subframe individually, then compares these local characteristics against threshold values. This local quality analysis enables the system to detect subtle distortions in specific segments of the speech signal while maintaining overall processing efficiency through parallel evaluation of multiple segments.

Inventive Principle:
Principle #3Local quality

3Measurement precision

If complex analysis methods are used to improve detection accuracy, then measurement precision is improved, but device complexity increases

Engineering Contradiction:
Improveabnormal frame detection accuracyVSAvoidsignal processing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent performs preliminary calculations of energy values, zero-crossing rates, and correlation coefficients for each subframe before making the final abnormal frame determination. These preliminary analyses prepare the data in advance, allowing the system to use simple threshold comparison rules rather than complex real-time decision algorithms, thereby reducing overall system complexity while maintaining high detection accuracy.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3163574B1Method and apparatus for detecting abnormal frame
Publication Date: 2019.09.25 HUAWEI TECH CO LTD
  • EP3163574B1 patent drawingFigure 1
  • EP3163574B1 patent drawingFigure 2
  • EP3163574B1 patent drawingFigure 3~4

AI summary

An abnormal frame detection method and apparatus are provided. The method includes: obtaining a signal frame from a speech signal, and dividing the signal frame into at least two subframes (301); obtaining a local energy value of a subframe of the signal frame, and obtaining, according to the local energy value of the subframe, a first characteristic value used to indicate a local energy trend of the signal frame (302); performing singularity analysis on the signal frame to obtain a second characteristic value (303); and determining the signal frame as an abnormal frame if the first characteristic value meets a first threshold and the second characteristic value meets a second threshold (304). It is implemented whether distortion occurs in a speech signal is detected.