Online speech synthesis
Turn text information into sound information and give your app a "mouth". We offer a wide selection of highly featured pronunciation people (sound library). Its synthetic sound in the tone, nature and other aspects of performance are close to or even more than the vocal. This TTS experience achieves a truly commercially available standard
The voice of the speech synthesis is still very bull p, not only the basic pronunciation of people, but also the quality of the pronunciation of people, characteristics of the pronunciation of people, the star pronunciation of people, of course, if you have special requirements can also be customized.
Here we choose the basic pronunciation of people do simple javaweb integration test, because other options to apply, think or forget, and so the process go through.
Platform Environment
JDK1.7, TOMCAT8, Eclipse, Flight JDK, win+ffmpeg (test), linux+docker+ffmpeg (production)
Description: True, win platform under FFmpeg installation use or is very easy, directly download the compression package without installation, Java directly invoke the execution of commands. Linux under a variety of dependent compilation can put your little machine to die, and also a variety of compilation errors, and then decisively use Docker, the only headache is that the environment is really clean, various commands do not support, of course, this is the advantage of Docker.
Flow chart
Web integration
Iflytek provides us with a simple SDK, the Iflytek MSC Development Guide-java. Of course, if you want to have a flying account, registration, create applications and so on here do not repeat, as long as the last can get to a app_id can.
Win+ffmpeg (Test)
Speech synthesis needs dynamic link library support, according to their own system to put Msc64.dll or Msc32.dll to the specified directory, you can use System.getproperty ("Java.library.path") to view, placed in any directory.
- According to their own system to download the corresponding FFmpeg, decompression can, directly call the bin directory Ffmpeg.exe can be. -
Linux+docker+ffmpeg (production) Get ffmpeg image
docker pull jrottenberg/ffmpeg
Create and run the container
docker run -it --name app_ffmpeg -p 8080:8080 -v /home/app_ffmpeg/:/mnt/app/ --entrypoint=‘bash‘ jrottenberg/ffmpeg
Note: The various Yum, wget, and Vim are not present in the Docker container, so the majority configuration is obtained from the host machine and then synchronously copied into the container.
Installing the configuration JDK
Oracle to get the certification must be downloaded, here we download and manually uploaded to the/home/app_ffmpeg directory.
# 复制配置文件到宿机docker cp 4f131c866092:/etc/profile /home/app_ffmpeg/
Edit Profile, append the following configuration
#set java environmentJAVA_HOME=/mnt/app/jdk1.7.0_80JRE_HOME=/mnt/app/jdk1.7.0_80/jreCLASS_PATH=.:$JAVA_HOME/lib/dt.jar:$JAVA_HOME/lib/tools.jar:$JRE_HOME/libPATH=$PATH:$JAVA_HOME/bin:$JRE_HOME/binexport JAVA_HOME JRE_HOME CLASS_PATH PATH
# 复制配置文件到容器docker cp /home/app_ffmpeg/profile 4f131c866092:/etc/
Entry into the container, effective configuration
# 进入容器docker exec -it app_ffmpeg bash# 使配置生效source /etc/profile# 检查JDK是否安装成功java -version
Installation configuration Tomcat
If the Tomcat boot card master does not move
Locate the Jdk1.x.x_xx/jre/lib/security/java.security file, find the Securerandom.source setting in the file, and change it to:
securerandom.source=file:/dev/./urandom
If the tomcat output is garbled in Chinese
localelocale -aLANG=C.UTF-8 (有的是zh_CN.UTF-8,不过我在本地没发现这种编码)source /etc/profile
Configure the news fly Dynamic Library
Depending on your system version, upload libmsc32.so or libmsc64.so to the/lib/and/lib64/directories, respectively.
Open Source Projects
Https://gitee.com/52itstyle/xufei_msc
Demo Address
http://xunfei.52itstyle.com/xufei_msc/
Voice Javaweb speech Synthesis Solution