Voice Alignment via Loss and Discontinuity Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing voice alignment methods for abnormal voice recognition in communications networks suffer from significant errors, requiring multiple algorithms and processes to achieve accuracy, which is inefficient and often inaccurate due to limitations in user privacy protection policies.
Innovation Solution
A voice alignment method that detects voice loss and discontinuity in test voices before alignment, selecting an appropriate algorithm based on detection results to align the test voice with the original voice, using techniques such as inserting or deleting silent statements to synchronize start and end time domain locations, thereby improving alignment efficiency and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing voice alignment methods are used to align original voice and test voice, then voice alignment can be performed, but the alignment accuracy is low and requires multiple algorithms and processing times
Solution Approach 1:
The patent applies preliminary action by performing detection of voice loss and discontinuity before the alignment operation. The detection unit identifies whether the test voice has voice loss or discontinuity issues, and this detection result guides the selection of appropriate alignment algorithms. This preliminary detection step prevents盲目 application of multiple complex algorithms, thereby improving alignment accuracy while reducing unnecessary processing complexity.
2Measurement precision
If multiple algorithms and processing times are used to overcome alignment errors, then alignment accuracy may improve, but processing efficiency decreases
Solution Approach 1:
The patent implements dynamics by dynamically selecting alignment algorithms based on the detection results. Instead of statically applying multiple algorithms in fixed sequences, the system adaptively chooses the most appropriate algorithm according to the specific voice conditions (whether voice loss or discontinuity is detected). This dynamic approach maintains high alignment accuracy while significantly improving processing efficiency by avoiding unnecessary algorithm applications.
3Reliability
If user privacy protection policies are strictly followed, then user privacy is protected, but the ability to recognize abnormal voices is restricted
Solution Approach 1:
The patent applies segmentation by dividing the voice analysis into separate functional components: a detection unit that identifies voice loss and discontinuity, and an alignment unit that performs alignment based on detection results. This segmentation allows the system to operate within privacy constraints by processing voices locally without requiring access to sensitive user data, while still maintaining effective abnormal voice recognition capability through the specialized detection and alignment mechanisms.
Data Source
Figure 1~2
Figure 3~5
Figure 6~7
AI summary
This application relates to artificial intelligence, and provides a voice alignment method, including: obtaining an original voice and a test voice, where the test voice is a voice generated after the original voice is transmitted over a communications network; performing loss detection and/or discontinuity detection on the test voice, where the loss detection is used to determine whether the test voice has a voice loss compared with the original voice, and the discontinuity detection is used to determine whether the test voice has voice discontinuity compared with the original voice; and aligning the test voice with the original voice based on a result of the loss detection and/or the discontinuity detection, to obtain an aligned original voice and an aligned test voice, where the result of the loss detection and/or the discontinuity detection is used to indicate a manner of aligning the test voice with the original voice. According to the voice alignment method provided in this application, the voice alignment method is determined based on the result of the loss detection and/or the discontinuity detection, and voice alignment may be performed based on a specific status of the test voice by using a most suitable method, thereby improving efficiency of the voice alignment.