Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A quick audio retrieval method

An audio and fast technology, applied in special data processing applications, instruments, electrical digital data processing, etc., can solve the problem that the retrieval speed cannot meet the requirements of use, takes a long time, etc., achieve large frame shift, improve retrieval speed, large frame length effect

Inactive Publication Date: 2009-10-28
HARBIN INST OF TECH
View PDF0 Cites 9 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] Although the retrieval speed of the existing histogram method based on sub-band energy is significantly improved compared with other methods, it still takes a long time to query an audio segment in an extremely large audio library. The speed cannot meet the usage requirements

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
  • A quick audio retrieval method
  • A quick audio retrieval method
  • A quick audio retrieval method

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0034] Specific implementation mode one: as Figure 1~5 As shown, the method for fast audio retrieval described in this embodiment enters the retrieval stage after the preprocessing stage;

[0035] The process of the pretreatment stage is:

[0036] Step A1, extracting MFCC parameters from the audio library: performing feature extraction on the audio signal of the audio library to be queried to obtain feature information; the obtained feature information is represented by an a×m matrix, where a is an integer greater than 6 and less than or equal to 30; Both m and n are integers, m>>n, n≥4; during feature extraction, the frame length is 2-3s, and the frame shift is 0.2-0.3s;

[0037] According to the frame length and frame shift, the a×m matrix is ​​divided into several a×n matrices, each matrix is ​​a frame; n is an integer, n≥4;

[0038] Step A2, extracting MFCC parameters to the query audio segment (retrieval target): according to the frame length and frame shift described ...

specific Embodiment approach 2

[0080] Specific implementation mode two: as Figure 1~5 As shown, in step A1 of this embodiment, when extracting features, the frame length is 2.5s, and the frame shift is 0.25s, which can improve the retrieval speed while ensuring the retrieval accuracy. Other steps are the same as in the first embodiment.

specific Embodiment approach 3

[0081] Specific implementation mode three: as Figure 1~5 As shown, in step A1 described in this embodiment, a is set to 12. In this method, the 12th order MFCC obtained by 24 Mel filter banks and DCT transformation is adopted. Other steps are the same as in the first embodiment.

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

A quick audio retrieval method relates to an audio retrieval method. The invention aims at providing a quick audio retrieval method so as to further improve the speed of audio retrieval. The method of the invention enters in a retrieval stage after a pretreatment stage; and the process of the pretreatment stage comprises the following steps of: extracting and classifying MFCC parameters from an audio library, and extracting MFCC parameters from a query audio section (retrieval target) and classifying the query audio section according to a sorted classification; and the retrieval process of the retrieval stage comprises the following steps of: looking over the classifications, coarse retrieval and accurate retrieval of the query audio section. The retrieval method consumes 8.5h at the speed of 4ms and is approximately 3 times (Pentium M1.6GHz) faster than a classical method after speed conversion.

Description

technical field [0001] The invention relates to an audio retrieval method. Background technique [0002] In the field of audio retrieval, the existing classical algorithm is based on the histogram method of subband energy. This method uses a 7-channel IIR filter to obtain a series of subband energy parameters, and then generates a histogram, and then dynamically Filter out a large number of impossible matching points to achieve the purpose of fast retrieval. At present, the retrieval speed of this method has reached 0.3 seconds to retrieve 200 hours (CPU Pentium IV 2.0GHz). [1] [0003] Although the retrieval speed of the existing histogram method based on sub-band energy is significantly improved compared with other methods, it still takes a long time to query an audio segment in an extremely large audio library. The speed cannot meet the usage requirements. Undoubtedly, the faster the retrieval speed, the better. [0004] References: [1] Kimura&Kashino, A Quick Search ...

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): G06F17/30
Inventor 任广辉吴静龙
Owner HARBIN INST OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products