Method for classifying digital audio automatically

An automatic classification and digital audio technology, applied in speech analysis, speech recognition, instruments, etc., can solve the problems of lack of audio classification methods, poor performance, complex calculations, etc., and achieve easy application, high accuracy, and simple calculation process Effect

Active Publication Date: 2014-06-11
UESTC COMSYS INFORMATION
View PDF4 Cites 33 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Acoustic perceptual features are some acoustically defined concepts that take into account the characteristics of human auditory perception, but usually the calculation is more complicated
Therefore, in the prior art, the methods of audio classification using different above-mentioned three features have achieved good results in one aspect, but poor performance in another, lacking a method that can integrate the respective advantages of the above-mentioned three features audio classification method for

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Method for classifying digital audio automatically
  • Method for classifying digital audio automatically
  • Method for classifying digital audio automatically

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0037] The present invention will be further described below in conjunction with the drawings and specific embodiments.

[0038] Such as figure 1 Shown is a flowchart of a method for implementing automatic digital audio classification according to an embodiment of the present invention, which specifically includes:

[0039] S1. Pre-processing the audio data, the pre-processing includes: pre-emphasis processing, frame processing and windowing processing;

[0040] S2. Extract audio features. The audio features include: spectrum centroid, spread spectrum, spectrum flatness, spectrum transition parameters, short-term energy, pitch frequency, and Mel frequency cepstral coefficient (MFCC) and MFCC first-order difference; where, the spectrum The flatness takes 23 coefficients to occupy 23 dimensions, the MFCC and its first-order difference each take 10 orders and a total of 20 dimensions, and the rest are each 1 dimension, forming a total of 48-dimensional feature vectors;

[0041] S3. Searc...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a method for classifying digital audio automatically. The method particularly comprises the following steps: pretreating an audio signal: pre-emphasis treatment, framing treatment and windowing treatment; extracting an audio characteristics comprising frequency spectrum mass center, spectrum spreading, frequency spectrum flatness, frequency spectrum transition parameters, short time energy, fundamental frequency and Mel frequency cepstrum coefficient (MFCC) and MFCC first order difference; indexing and classifying according to the audio characteristic, and finding out the similar audio. According to the invention, the problem of error caused by indexing audio by adopting a single audio characteristic in the prior art is solved; moreover, the method is simple in computation process, is easy to apply practically, and is high in indexing efficiency.

Description

Technical field [0001] The invention belongs to the technical field of audio retrieval, and specifically relates to the design of a method for realizing automatic classification of digital audio. Background technique [0002] Nowadays, there are a large amount of digital audio in online media. It is difficult to find files with the same or similar characteristics. The current common practice is to distinguish them based on text information such as file name, singer or scholar, label, etc. And doing so is very subjective, which makes the results inaccurate. Content-based audio feature extraction can overcome these shortcomings. Because the amount of audio data samples is too large to be directly used for automatic classification, corresponding features must be extracted from this numerical sequence. Common audio features can be roughly divided into three Types: time domain features, frequency domain features and acoustic perception features. [0003] The time domain feature only u...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G10L15/08
Inventor 陈科李世旭
Owner UESTC COMSYS INFORMATION
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap