在Java中實現語音聊天，你可以使用一些現成的庫和框架，例如JavaZoom的JSyn、JMF（Java Media Framework）或者使用WebRTC技術。下面是一個簡單的使用JMF實現語音聊天的示例：

1. 首先，確保你已經安裝了Java開發環境（JDK）和構建工具（如Maven或Gradle）。

2. 添加JMF依賴到你的項目中。如果你使用Maven，可以在pom.xml文件中添加以下依賴：

<dependency>
    <groupId>com.sun.media</groupId>
    <artifactId>jai_core</artifactId>
    <version>1.1.3</version>
</dependency>
<dependency>
    <groupId>com.sun.media</groupId>
    <artifactId>jai_imageio</artifactId>
    <version>1.1</version>
</dependency>
<dependency>
    <groupId>com.sun.media</groupId>
    <artifactId>soundbank</artifactId>
    <version>1.1.2</version>
</dependency>

3. 創建一個簡單的語音聊天程序，包括兩個客戶端和一個服務器端。

服務器端代碼（Server.java）：

import javax.media.*;
import javax.media.protocol.*;
import javax.media.control.*;
import java.io.*;
import java.net.*;

public class Server {
    public static void main(String[] args) throws Exception {
        ServerSocket serverSocket = new ServerSocket(12345);
        Socket socket = serverSocket.accept();
        AudioFormat format = new AudioFormat(16000, 16, 2, true, true);
        DataLine.Info info = new DataLine.Info(TargetDataLine.class, format);
        TargetDataLine line = (TargetDataLine) AudioSystem.getLine(info);
        line.open(format);
        line.start();

        Thread receiverThread = new Thread(() -> {
            try {
                InputStream in = socket.getInputStream();
                AudioInputStream audioStream = new AudioInputStream(in);
                AudioFormat decodedFormat = new AudioFormat(audioStream.getFormat().getSampleRate(),
                        audioStream.getFormat().getChannels(), audioStream.getFormat().getSampleSizeInBits() / 8,
                        audioStream.getFormat().isBigEndian(), audioStream.getFormat().getChannels());
                AudioInputStream decodedStream = AudioSystem.getAudioInputStream(decodedFormat, audioStream);
                line.stop();
                line.close();
                line = null;

                // Pass the decoded stream to the client
                // ...

            } catch (Exception e) {
                e.printStackTrace();
            }
        });
        receiverThread.start();

        // Send the captured audio stream to all connected clients
        // ...

        socket.close();
        serverSocket.close();
    }
}

客戶端代碼（Client.java）：

import javax.media.*;
import javax.media.protocol.*;
import javax.media.control.*;
import java.io.*;
import java.net.*;

public class Client {
    public static void main(String[] args) throws Exception {
        Socket socket = new Socket("localhost", 12345);
        AudioFormat format = new AudioFormat(16000, 16, 2, true, true);
        DataLine.Info info = new DataLine.Info(TargetDataLine.class, format);
        TargetDataLine line = (TargetDataLine) AudioSystem.getLine(info);
        line.open(format);
        line.start();

        OutputStream out = socket.getOutputStream();
        AudioInputStream audioStream = new AudioInputStream(line);
        AudioFormat encodedFormat = new AudioFormat(format.getSampleRate(), format.getChannels(),
                format.getSampleSizeInBits() / 8, format.isBigEndian(), format.getChannels());
        AudioInputStream encodedStream = AudioSystem.getAudioInputStream(encodedFormat, audioStream);

        Thread senderThread = new Thread(() -> {
            try {
                byte[] buffer = new byte[1024];
                int bytesRead;
                while ((bytesRead = encodedStream.read(buffer)) != -1) {
                    out.write(buffer, 0, bytesRead);
                }
            } catch (Exception e) {
                e.printStackTrace();
            }
        });
        senderThread.start();

        // Receive the decoded audio stream from the server
        // ...

        line.stop();
        line.close();
        socket.close();
    }
}

這個示例只是一個簡化的版本，實際應用中需要考慮更多的細節，例如處理多個客戶端的連接、音頻數據的編碼和解碼、錯誤處理和異常管理等。你還可以考慮使用更高級的庫，如WebRTC，來實現更復雜的語音聊天應用。