实时通讯

src/main/java/com/realtime/config/ClientVoiceHandler.java

@@ -0,0 +1,90 @@
+package com.realtime.config;
+
+import com.realtime.service.DoubaoVoiceService;
+import lombok.extern.slf4j.Slf4j;
+import org.bytedeco.javacv.FFmpegFrameGrabber;
+import org.bytedeco.javacv.FFmpegFrameRecorder;
+import org.bytedeco.javacv.Frame;
+import org.springframework.stereotype.Component;
+import org.springframework.web.socket.BinaryMessage;
+import org.springframework.web.socket.CloseStatus;
+import org.springframework.web.socket.WebSocketSession;
+import org.springframework.web.socket.handler.BinaryWebSocketHandler;
+
+import java.io.ByteArrayInputStream;
+import java.io.ByteArrayOutputStream;
+import java.io.EOFException;
+@Slf4j
+@Component
+public class ClientVoiceHandler extends BinaryWebSocketHandler {
+    private final DoubaoVoiceService doubaoVoiceService;
+
+    public ClientVoiceHandler(DoubaoVoiceService doubaoVoiceService) {
+        this.doubaoVoiceService = doubaoVoiceService;
+    }
+
+    @Override
+    public void afterConnectionEstablished(WebSocketSession session) throws Exception {
+        // 客户端会话
+        // 建立与豆包 API 的连接
+
+        System.out.println(session);
+        doubaoVoiceService.connect(session);
+    }
+
+    @Override
+    protected void handleBinaryMessage(WebSocketSession session, BinaryMessage message) throws Exception {
+        System.out.println("2");
+        // 接收客户端的音频数据（如 PCM 片段）
+        byte[] audioData = message.getPayload().array();
+        // 预处理：确保格式符合豆包 API 要求（如采样率、位深）
+        byte[] processedData = preprocessAudio(audioData);
+        // 发送给豆包 API
+        doubaoVoiceService.sendAudio(processedData);
+    }
+
+    @Override
+    public void afterConnectionClosed(WebSocketSession session, CloseStatus status) throws Exception {
+        // 关闭与豆包 API 的连接
+        doubaoVoiceService.close();
+    }
+
+    private byte[] preprocessAudio(byte[] audioData) {
+        if (audioData == null || audioData.length == 0) {
+            log.warn("接收到空的音频数据，跳过处理");
+            return new byte[0]; // 返回空数组，避免后续错误
+        }
+
+        ByteArrayOutputStream outputStream = new ByteArrayOutputStream();
+
+        try (FFmpegFrameGrabber grabber = new FFmpegFrameGrabber(new ByteArrayInputStream(audioData));
+             FFmpegFrameRecorder recorder = new FFmpegFrameRecorder(outputStream, 1)) {
+
+            grabber.start();
+            // 打印输入音频信息（调试用）
+            log.debug("输入音频格式：{}，采样率：{}，声道：{}",
+                    grabber.getFormat(), grabber.getSampleRate(), grabber.getAudioChannels());
+
+            recorder.setFormat("s16le");
+            recorder.setSampleRate(16000);
+            recorder.setAudioChannels(1);
+            recorder.start();
+
+            Frame frame;
+            int frameCount = 0;
+            while ((frame = grabber.grab()) != null) {
+                recorder.record(frame);
+                frameCount++;
+            }
+            log.debug("成功处理 {} 帧音频", frameCount);
+
+            recorder.stop();
+            grabber.stop();
+            return outputStream.toByteArray();
+
+        } catch (Exception e) {
+            log.error("音频预处理失败", e);
+            return new byte[0];
+        }
+    }
+}