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

Deep learning-based train broadcast speech enhancement method and system

A technology for train broadcasting and voice enhancement, applied in voice analysis, instruments, etc., can solve problems such as inability to combine and increase the sound field, and achieve the effects of less difficulty, fullness, and low cost

Active Publication Date: 2022-03-01
SHENZHEN TONGGUANG TRACK TRAFFIC TECH
View PDF12 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The above technical solution provides an audio chip for the train environment, but it cannot combine the relevant information of the subway train passenger information system to automatically switch the sound field equalization parameters for different operating scenarios and dynamically increase the sound field. Therefore, this application proposes a new technical solution

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
  • Deep learning-based train broadcast speech enhancement method and system
  • Deep learning-based train broadcast speech enhancement method and system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0036] refer to figure 1 , the train announcement voice enhancement method based on deep learning includes:

[0037] S101. Obtain train operation information of the vehicle-mounted PIS system.

[0038] The above-mentioned vehicle-mounted PIS system is the passenger information system. In this application, the train operation information includes: speed information (vehicle speed), video information and designated identification information, which are used to determine the real-time scene of the train.

[0039] Moreover, when the vehicle speed is 0-30KM / H, it is identified as the starting scene; when the vehicle speed is 30-0KM / H, it is identified as the parking scene; other speeds are identified as the running scene; it should be noted that the above 30 is assigned to the low-speed threshold X The specific number can be selected according to the vehicle model and operating environment.

[0040] The above-mentioned video information is the video stream collected by the camer...

Embodiment 2

[0059] The difference with embodiment 1 is that this method also includes:

[0060] Record process information, bind time parameters, and save as audio transformation files;

[0061] Train preset neural network models using audio modification archives;

[0062] The trained neural network model is used to identify the subsequent real-time audio stream. If there is a record and the current environment of the train matches the recorded information, the processed audio stream in the record is called and sent to the on-board PIS system for playback.

[0063] It can be understood that the above process information is the key information during the execution of the method described in the embodiment, for example: the original audio stream, the corresponding real-time scene of the train, the processed audio stream, and the acoustic equalization parameters. By binding time parameters, a one-to-one correspondence can be established to know when, where, what scene, what kind of original...

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 train broadcast speech enhancement method and system based on deep learning. The method comprises the following steps: acquiring train operation information of a vehicle-mounted PIS system; train operation information is recognized, and a train real-time scene is determined; reading acoustic equalization parameters matched with the train real-time scene from a preset database; acquiring an audio stream of the vehicle-mounted PIS system; and based on the acoustic equalization parameter, processing the audio stream by using a preset central control sound effect algorithm, adjusting a sound sensing position, and sending the processed audio stream to the vehicle-mounted PIS system for playing. The method has the effects of automatically switching the sound field equalization parameters in combination with different operation scenes of the train and dynamically increasing the sound field.

Description

technical field [0001] The present application relates to the technical field of audio processing, in particular to a method and system for enhancing train broadcast voice based on deep learning. Background technique [0002] For the train, its station broadcast, message notification, etc. need to be notified by broadcasting. However, during the operation of the train, in addition to the noise from the internal passengers, there are also wind noise and mechanical noise outside the train, which interfere with the voice broadcast effect. [0003] The patent with application number CN201710105745.8 is a train voice amplification unit based on ARM+FPGA architecture. The train voice amplification unit includes three parts: a main control module, a codec amplification module and a communication module. The main control module includes the main control chip and peripheral devices, responsible for system initialization, audio storage and processing, and running application programs...

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
IPC IPC(8): G10L21/0232G10L25/30
CPCG10L21/0232G10L25/30
Inventor 李跃群张有利
Owner SHENZHEN TONGGUANG TRACK TRAFFIC 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

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com