Method for segmenting audio frequencies on basis of lay reliability variation trend

Abstract

The invention discloses a method for segmenting audio frequencies on the basis of lay reliability variation trend. The method includes particular steps of 1), detecting layer jump points in fixed-length analysis windows on the basis of reliability variation trend; 2), segmenting the audio frequencies on the basis of detection on the jump points in the fixed-length analysis windows. The method has the advantages that accumulative errors due to split-window detection can be prevented by the aid of the fixed-length analysis windows, the acoustic jump points are detected in the fixed-length analysis windows according to the reliability variation trend, and accordingly limitation problems of excessive redundant points and hard decision can be solved; the problem of mixing detection of existing jump points can be solved by the aid of the fixed-length analysis windows in top-down layer detection modes, redundant segmentation points can be effectively reduced, and accordingly the comprehensive performance can be improved.

Description

【Technical field】 [0001] The invention belongs to the field of audio signal processing, in particular to an audio segmentation method based on the variation trend of hierarchical reliability. 【Background technique】 [0002] In recent years, in the field of signal processing, in terms of signal segmentation processing, content-based audio signal segmentation, also known as jump point detection, refers to the phenomenon that there are differences between auditory features when a continuous audio signal stream changes. Audio is divided into segments, each of which is either consistent in content or similar in acoustic characteristics. At present, content-based audio signal segmentation methods are mainly divided into three categories: distance-based, model-based, and model-selection-based segmentation methods. Among them, the distance-based segmentation method uses the distance between samples of adjacent windows to measure the similarity of adjacent audio segments. It has the...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the problems in the existing segmentation method that the setting of the threshold leads to redundant or missed detection of the segmentation point, the cumulative error caused by the window detection method, and the limitation of the hard judgment method, the present invention proposes a new signal audio segmentation method based on Audio Segmentation Method Based on Variation Trend of Hierarchical Reliability

Images