最近在研究语音识别，使用百度的sdk。发现只有识别的部分，而我需要保存音频文件，并且实现当有声音传入时自动生成音频文件。
先上代码：
public class EngineeCore {
String filePath = "E:\\voice\\voice_cache.wav";
AudioFormat audioFormat;
TargetDataLine targetDataLine;
boolean flag = true;
private void stopRecognize() {
flag = false;
targetDataLine.stop();
targetDataLine.close();
}private AudioFormat getAudioFormat() {
float sampleRate = 16000;
// 8000,11025,16000,22050,44100
int sampleSizeInBits = 16;
// 8,16
int channels = 1;
// 1,2
boolean signed = true;
// true,false
boolean bigEndian = false;
// true,false
return new AudioFormat(sampleRate, sampleSizeInBits, channels, signed, bigEndian);
}// end getAudioFormat
private void startRecognize() {
try {
// 获得指定的音频格式
audioFormat = getAudioFormat();
DataLine.Info dataLineInfo = new DataLine.Info(TargetDataLine.class, audioFormat);
targetDataLine = (TargetDataLine) AudioSystem.getLine(dataLineInfo);
// Create a thread to capture the microphone
// data into an audio file and start the
// thread running. It will run until the
// Stop button is clicked. This method
// will return after starting the thread.
flag = true;
new CaptureThread().start();
} catch (Exception e) {
e.printStackTrace();
} // end catch
}// end captureAudio method
class CaptureThread extends Thread {
public void run() {
AudioFileFormat.Type fileType = null;
File audioFile = new File(filePath);
fileType = AudioFileFormat.Type.WAVE;
//声音录入的权值
int weight = 2;
//判断是否停止的计数
int downSum = 0;
ByteArrayInputStream bais = null;
ByteArrayOutputStream baos = new ByteArrayOutputStream();
AudioInputStream ais = null;
try {
targetDataLine.open(audioFormat);
targetDataLine.start();
byte[] fragment = new byte[1024];
ais = new AudioInputStream(targetDataLine);
while (flag) {
targetDataLine.read(fragment, 0, fragment.length);
//当数组末位大于weight时开始存储字节（有声音传入），一旦开始不再需要判断末位
if (Math.abs(fragment[fragment.length-1]) > weight || baos.size() > 0) {
baos.write(fragment);
System.out.println("守卫："+fragment[0]+",末尾："+fragment[fragment.length-1]+",lenght"+fragment.length);
//判断语音是否停止
if(Math.abs(fragment[fragment.length-1])<=weight){
downSum++;
}else{
System.out.println("重置奇数");
downSum=0;
} //计数超过20说明此段时间没有声音传入(值也可更改)
if(downSum>20){
System.out.println("停止录入");
break;
}
}
}
//取得录音输入流
audioFormat = getAudioFormat();
byte audioData[] = baos.toByteArray();
bais = new ByteArrayInputStream(audioData);
ais = new AudioInputStream(bais, audioFormat, audioData.length / audioFormat.getFrameSize());
//定义最终保存的文件名
System.out.println("开始生成语音文件");
AudioSystem.write(ais, AudioFileFormat.Type.WAVE, audioFile);
downSum = 0;
stopRecognize();
} catch (Exception e) {
e.printStackTrace();
} finally {
//关闭流
try {
ais.close();
bais.close();
baos.reset();
} catch (IOException e) {
e.printStackTrace();
}
}
}// end run
}// end inner class CaptureThread
接下来测试
public static void main(String args[]) {
EngineeCore engineeCore = new EngineeCore();
engineeCore.startRecognize();
}
当有较高的声音传入麦克风时，targetDataLine读取的字节数组首位或末位绝对值会变大（位置取决于音频格式中的一些参数，如bigEndian）。传入音量低，绝对值会变小
录音开始。从targetDataLine中读取的音频数据被保存在ByteArrayOutputStream中。一段时间音量一直低于权值时，认为无声音传入，结束录音。从ByteArrayOutputStream取出字节数组，
转为音频保存在本地文件中。
从targetDataLine读取的字节数组不能直接用于百度等语音识别，需要先转为音频文件，然后读取音频文件生成的字节数组，才可用于语音识别。
