devbox
我家小花儿
这个屌丝很懒，什么也没留下！
关注作者
热门标签
热门文章
当前位置:   article > 正文

利用科大讯飞API实现语音识别，SWT实现客户端封装_科大讯飞的语音识别api免费可用多久啊

作者：我家小花儿 | 2024-03-12 19:18:23

赞

踩

科大讯飞的语音识别api免费可用多久啊

利用科大讯飞API来实现语音识别，利用Java SWT来封装界面。

科大讯飞API

语音识别的API可以免费试用5小时，许多厂家已经开放了语音识别的API例如百度，阿里等，这里使用科大讯飞的API来实现。其实也可以自己训练数据来实现语音识别的功能，只不过识别率可能不是太高，具体实现原理可以参考如下：日后有时间可以研究一下。

https://blog.ailemon.me/2018/08/29/asrt-a-chinese-speech-recognition-system/

https://github.com/nl8590687/ASRT_SpeechRecognition

声学模型通过采用卷积神经网络（CNN）和连接性时序分类（CTC）方法，使用大量中文语音数据集进行训练，将声音转录为中文拼音，并通过语言模型，将拼音序列转换为中文文本。

登录科大讯飞网址：https://www.xfyun.cn/services/lfasr

   

下载Java  SDK

   

新建应用，获取appId以及secret

       

在SDK中配置appId以及secret

  1. # APP ID
  2. app_id=
  3. # secret key
  4. secret_key=
  5. # we support both http and https prototype
  6. lfasr_host=http://raasr.xfyun.cn/api
  7. # file piece size
  8. file_piece_size=10485760
  9. # store path: this is not the store path for the result json file, but the path for the file piece during upload
  10. store_path=F://

Demo中给出一个测试用例：

使用过程如下：

初始化LFASRClient实例

  1.        // 初始化LFASRClient实例
  2.         LfasrClientImp lc = null;
  3.         try {
  4.             lc = LfasrClientImp.initLfasrClient();
  5.         } catch (LfasrException e) {
  6.             // 初始化异常，解析异常描述信息
  7.             Message initMsg = JSON.parseObject(e.getMessage(), Message.class);
  8.             System.out.println("ecode=" + initMsg.getErr_no());
  9.             System.out.println("failed=" + initMsg.getFailed());
  10.         }

上传语音文件

  1.  
  2.             // 上传音频文件
  3.             Message uploadMsg = lc.lfasrUpload(local_file, type, params);
  4.  
  5.             // 判断返回值
  6.             int ok = uploadMsg.getOk();
  7.             if (ok == 0) {
  8.                 // 创建任务成功
  9.                 task_id = uploadMsg.getData();

循环等待任务处理结果：

  1. // 循环等待音频处理结果
  2.         while (true) {
  3.             try {
  4.                 // 等待20s在获取任务进度
  5.                 Thread.sleep(sleepSecond * 1000);
  6.                 System.out.println("waiting ...");
  7.             } catch (InterruptedException e) {
  8.                 e.printStackTrace();
  9.             }
  10.             try {
  11.                 // 获取处理进度
  12.                 Message progressMsg = lc.lfasrGetProgress(task_id);
  13.  
  14.                 // 如果返回状态不等于0，则任务失败
  15.                 if (progressMsg.getOk() != 0) {
  16.                     System.out.println("task was fail. task_id:" + task_id);
  17.                     System.out.println("ecode=" + progressMsg.getErr_no());
  18.                     System.out.println("failed=" + progressMsg.getFailed());
  19.  
  20.                     return;
  21.                 } else {
  22.                     ProgressStatus progressStatus = JSON.parseObject(progressMsg.getData(), ProgressStatus.class);
  23.                     if (progressStatus.getStatus() == 9) {
  24.                         // 处理完成
  25.                         System.out.println("task was completed. task_id:" + task_id);
  26.                         break;
  27.                     } else {
  28.                         // 未处理完成
  29.                         System.out.println("task is incomplete. task_id:" + task_id + ", status:" + progressStatus.getDesc());
  30.                         continue;
  31.                     }
  32.                 }
  33.             } catch (LfasrException e) {
  34.                 // 获取进度异常处理，根据返回信息排查问题后，再次进行获取
  35.                 Message progressMsg = JSON.parseObject(e.getMessage(), Message.class);
  36.                 System.out.println("ecode=" + progressMsg.getErr_no());
  37.                 System.out.println("failed=" + progressMsg.getFailed());
  38.             }
  39.         }

获取最终结果：

  1.        // 获取任务结果
  2.         try {
  3.             Message resultMsg = lc.lfasrGetResult(task_id);
  4.             // 如果返回状态等于0，则获取任务结果成功
  5.             if (resultMsg.getOk() == 0) {
  6.                 // 打印转写结果
  7.             	System.out.println(resultMsg.getData());
  8.                 System.out.println(Test.getFinalResult(resultMsg.getData()));
  9.             } else {
  10.                 // 获取任务结果失败
  11.                 System.out.println("ecode=" + resultMsg.getErr_no());
  12.                 System.out.println("failed=" + resultMsg.getFailed());
  13.             }
  14.         } catch (LfasrException e) {
  15.             // 获取结果异常处理，解析异常描述信息
  16.             Message resultMsg = JSON.parseObject(e.getMessage(), Message.class);
  17.             System.out.println("ecode=" + resultMsg.getErr_no());
  18.             System.out.println("failed=" + resultMsg.getFailed());
  19.         }

resultMsg.getData()返回一个json数组，里面有多个元素，在此将“onebest”元素取出拼接组成最终的输出文本。

  1. String str = "[{\"bg\":\"0\",\"ed\":\"2180\",\"onebest\":\"科大讯飞是中国最大！\",\"si\":\"0\",\"speaker\":\"0\","
  2. 				+ "\"wordsResultList\":[{\"alternativeList\":[],\"wc\":\"1.0000\",\"wordBg\":\"6\",\"wordEd\":\"114\",\"wordsName\":"
  3. 				+ "\"科大讯飞\",\"wp\":\"n\"},{\"alternativeList\":[],\"wc\":\"1.0000\",\"wordBg\":\"118\",\"wordEd\":\"147\",\"wordsName\""
  4. 				+ ":\"是\",\"wp\":\"n\"},{\"alternativeList\":[],\"wc\":\"1.0000\",\"wordBg\":\"148\",\"wordEd\":\"193\",\"wordsName\":\"中国\","
  5. 						+ "\"wp\":\"n\"},{\"alternativeList\":[],\"wc\":\"1.0000\",\"wordBg\":\"194\",\"wordEd\":\"213\",\"wordsName\":\"最\","
  6. 						+ "\"wp\":\"n\"},{\"alternativeList\":[],\"wc\":\"1.0000\",\"wordBg\":\"214\",\"wordEd\":\"218\",\"wordsName\":\"大\","
  7. 						+ "\"wp\":\"n\"},{\"alternativeList\":[],\"wc\":\"0.0000\",\"wordBg\":\"218\",\"wordEd\":\"218\",\"wordsName\":\"！\","
  8. 						+ "\"wp\":\"p\"},{\"alternativeList\":[],\"wc\":\"0.0000\",\"wordBg\":\"218\",\"wordEd\":\"218\",\"wordsName\":\"\","
  9. 						+ "\"wp\":\"g\"}]},{\"bg\":\"2190\",\"ed\":\"3080\",\"onebest\":\"的智能。\",\"si\":\"1\",\"speaker\":\"0\","
  10. 						+ "\"wordsResultList\":[{\"alternativeList\":[],\"wc\":\"1.0000\",\"wordBg\":\"15\",\"wordEd\":\"42\","
  11. 						+ "\"wordsName\":\"的\",\"wp\":\"n\"},{\"alternativeList\":[],\"wc\":\"1.0000\",\"wordBg\":\"47\",\"wordEd\":\"89\","
  12. 						+ "\"wordsName\":\"智能\",\"wp\":\"n\"},{\"alternativeList\":[],\"wc\":\"0.0000\",\"wordBg\":\"89\",\"wordEd\":\"89\","
  13. 						+ "\"wordsName\":\"。\",\"wp\":\"p\"},{\"alternativeList\":[],\"wc\":\"0.0000\",\"wordBg\":\"89\",\"wordEd\":\"89\","
  14. 						+ "\"wordsName\":\"\",\"wp\":\"g\"}]},{\"bg\":\"3090\",\"ed\":\"4950\",\"onebest\":\"语音技术提供商，\",\"si\":\"2\","
  15. 						+ "\"speaker\":\"0\",\"wordsResultList\":[{\"alternativeList\":[],\"wc\":\"1.0000\",\"wordBg\":\"4\",\"wordEd\":\"46\","
  16. 						+ "\"wordsName\":\"语音\",\"wp\":\"n\"},{\"alternativeList\":[],\"wc\":\"1.0000\",\"wordBg\":\"47\",\"wordEd\":\"92\","
  17. 						+ "\"wordsName\":\"技术\",\"wp\":\"n\"},{\"alternativeList\":[],\"wc\":\"1.0000\",\"wordBg\":\"93\",\"wordEd\":\"164\","
  18. 						+ "\"wordsName\":\"提供商\",\"wp\":\"n\"},{\"alternativeList\":[],\"wc\":\"0.0000\",\"wordBg\":\"164\",\"wordEd\":\"164\","
  19. 						+ "\"wordsName\":\"，\",\"wp\":\"p\"}]}]";
  1. public static String getFinalResult(String data){
  2. 		
  3. 		JSONArray ja = JSONArray.parseArray(data);
  4. 		StringBuilder sb = new StringBuilder();
  5. 		for(int i=0; i<ja.size(); i++){
  6. 			//System.out.println(ja.get(i));			
  7. 			sb.append(JSON.parseObject(ja.get(i).toString()).get("onebest"));
  8. 			//System.out.println(JSON.parseObject(ja.get(i).toString()).get("onebest"));
  9. 		}
  10. 		return sb.toString();
  11. 	}

SWT界面

直接使用有点费劲，想利用SWT来封装一个客户端，这里使用Eclipse来开发，首先安装SWT环境

参考地址如下：https://www.cnblogs.com/xinyan123/p/6225194.html

下载SWT插件：https://www.eclipse.org/windowbuilder/download.php

                                 

将安装包features以及plugins放入到eclipse安装目录对应文件夹下，重启eclipse

        

新建SWT工程

                                     

新建一个ApplicationWindow

                                    

可以使用图形化界面来进行界面UI设计

                                         

SWT核心实现：开始转换按钮的实现逻辑

  1.         //开始转换按钮
  2. 		Button startThansfer = new Button(container, SWT.NONE);
  3. 		startThansfer.addSelectionListener(new SelectionAdapter() {
  4. 			@Override
  5. 			public void widgetSelected(SelectionEvent e) {
  6. 				logDetailText.append(datePrefix + "开始转换........" + "\n");
  7. 				startThansfer.setEnabled(false);
  8. 				voicePath = voicePathText.getText();
  9. 				textPath = textPathText.getText();									
  10. 				int status = 0;					
  11. 				Callable<Integer> f = new TransferThread(logDetailText, countDownLatch, datePrefix, voicePath, textPath);
  12. 				//Callable<Integer> f = new TransferThreadAsyc(parent, logDetailText, countDownLatch, datePrefix, voicePath, textPath);
  13. 				try {
  14. 					status = f.call();
  15. 				} catch (Exception e2) {
  16. 					// TODO Auto-generated catch block
  17. 					e2.printStackTrace();
  18. 				}												
  19. 				try {
  20. 					countDownLatch.await();
  21. 				} catch (InterruptedException e1) {
  22. 					// TODO Auto-generated catch block
  23. 					e1.printStackTrace();
  24. 				}				
  25. 				if(status == 1){
  26. 					logDetailText.append(datePrefix + "转换完成" + "\n");
  27. 				}else{
  28. 					logDetailText.append(datePrefix + "转换失败" + "\n");
  29. 				}						
  30. 				startThansfer.setEnabled(true);
  31. 			}
  32. 		});

转换执行线程工作类：

  1. package com.voice.text;
  2.  
  3. import java.io.FileOutputStream;
  4. import java.util.HashMap;
  5. import java.util.concurrent.Callable;
  6. import java.util.concurrent.CountDownLatch;
  7.  
  8. import org.eclipse.swt.widgets.Text;
  9.  
  10. import com.alibaba.fastjson.JSON;
  11. import com.iflytek.msp.cpdb.lfasr.client.LfasrClientImp;
  12. import com.iflytek.msp.cpdb.lfasr.exception.LfasrException;
  13. import com.iflytek.msp.cpdb.lfasr.model.LfasrType;
  14. import com.iflytek.msp.cpdb.lfasr.model.Message;
  15. import com.iflytek.msp.cpdb.lfasr.model.ProgressStatus;
  16. import com.iflytek.voicecloud.lfasr.demo.Test;
  17.  
  18. public class TransferThread implements Callable<Integer> {
  19. 	
  20. 	private Text logDetailText;
  21. 	private CountDownLatch countDownLatch;
  22. 	private LfasrType type = LfasrType.LFASR_STANDARD_RECORDED_AUDIO;
  23. 	private int sleepSecond = 20;
  24. 	private String datePrefix;
  25. 	private String voicePath;
  26. 	private String textPath;
  27. 	
  28. 	public TransferThread(Text logDetailText, CountDownLatch countDownLatch, String datePrefix, String voicePath, String textPath) {
  29. 		this.logDetailText = logDetailText;
  30. 		this.countDownLatch = countDownLatch;
  31. 		this.datePrefix = datePrefix;
  32. 		this.voicePath = voicePath;
  33. 		this.textPath = textPath;
  34. 	}
  35. 	@Override
  36. 	public Integer call() throws Exception {
  37. 		// 初始化LFASRClient实例
  38.         LfasrClientImp lc = null;
  39.         try {
  40.             lc = LfasrClientImp.initLfasrClient();
  41.         } catch (LfasrException e) {
  42.             // 初始化异常，解析异常描述信息
  43.             Message initMsg = JSON.parseObject(e.getMessage(), Message.class);
  44.             logDetailText.append(datePrefix + "ecode=" + initMsg.getErr_no() + "\n");
  45.             System.out.println("ecode=" + initMsg.getErr_no());
  46.             logDetailText.append(datePrefix + "failed=" + initMsg.getFailed() + "\n");
  47.             //System.out.println(datePrefix + "failed=" + initMsg.getFailed());
  48.             countDownLatch.countDown();
  49.             return -1;
  50.         }
  51.  
  52.         // 获取上传任务ID
  53.         String task_id = "";
  54.         HashMap<String, String> params = new HashMap<String, String>();
  55.         params.put("has_participle", "true");
  56.         //合并后标准版开启电话版功能
  57.         //params.put("has_seperate", "true");
  58.         try {
  59.             // 上传音频文件
  60.             Message uploadMsg = lc.lfasrUpload(voicePath, type, params);
  61.  
  62.             // 判断返回值
  63.             int ok = uploadMsg.getOk();
  64.             if (ok == 0) {
  65.                 // 创建任务成功
  66.                 task_id = uploadMsg.getData();
  67.                 //System.out.println("创建任务成功  task_id=" + task_id);
  68.                 logDetailText.append(datePrefix + "创建任务成功  task_id=" + task_id + "\n");
  69.             } else {
  70.                 // 创建任务失败-服务端异常
  71.                 //System.out.println(datePrefix + "ecode=" + uploadMsg.getErr_no());
  72.                 logDetailText.append(datePrefix + "ecode=" + uploadMsg.getErr_no() + "\n");
  73.                 //System.out.println(datePrefix + "failed=" + uploadMsg.getFailed());
  74.                 logDetailText.append(datePrefix + "failed=" + uploadMsg.getFailed() + "\n");
  75.                 countDownLatch.countDown();
  76.                 return -1;
  77.             }
  78.         } catch (LfasrException e) {
  79.             // 上传异常，解析异常描述信息
  80.             Message uploadMsg = JSON.parseObject(e.getMessage(), Message.class);
  81.             //System.out.println(datePrefix + "ecode=" + uploadMsg.getErr_no());
  82.             logDetailText.append(datePrefix + "ecode=" + uploadMsg.getErr_no() + "\n");
  83.             //System.out.println(datePrefix + "failed=" + uploadMsg.getFailed()); 
  84.             logDetailText.append(datePrefix + "failed=" + uploadMsg.getFailed() + "\n"); 
  85.             countDownLatch.countDown();
  86.             return -1;
  87.         }
  88.  
  89.         // 循环等待音频处理结果
  90.         while (true) {
  91.             try {
  92.                 // 等待20s在获取任务进度
  93.                 Thread.sleep(sleepSecond * 1000);
  94.                 //System.out.println("waiting ...");
  95.                 logDetailText.append(datePrefix + "failed=" + "waiting ..." + "\n");
  96.             } catch (InterruptedException e) {
  97.                 e.printStackTrace();
  98.             }
  99.             try {
  100.                 // 获取处理进度
  101.                 Message progressMsg = lc.lfasrGetProgress(task_id);
  102.  
  103.                 // 如果返回状态不等于0，则任务失败
  104.                 if (progressMsg.getOk() != 0) {
  105.                     //System.out.println("task was fail. task_id:" + task_id);
  106.                     //System.out.println("ecode=" + progressMsg.getErr_no());
  107.                     //System.out.println("failed=" + progressMsg.getFailed());
  108.                     logDetailText.append(datePrefix + "task was fail. task_id:" + task_id + "\n");
  109.                     logDetailText.append(datePrefix + "ecode=" + progressMsg.getErr_no() + "\n");
  110.                     logDetailText.append(datePrefix + "failed=" + progressMsg.getFailed() + "\n");
  111.                     countDownLatch.countDown();
  112.                     return -1;
  113.                 } else {
  114.                     ProgressStatus progressStatus = JSON.parseObject(progressMsg.getData(), ProgressStatus.class);
  115.                     if (progressStatus.getStatus() == 9) {
  116.                         // 处理完成
  117.                         //System.out.println(datePrefix + "task was completed. task_id:" + task_id + "\n");
  118.                         logDetailText.append(datePrefix + "task was completed. task_id:" + task_id + "\n");
  119.                         break;
  120.                     } else {
  121.                         // 未处理完成
  122.                         //System.out.println(datePrefix + "task is incomplete. task_id:" + task_id + ", status:" + progressStatus.getDesc() + "\n");
  123.                         logDetailText.append(datePrefix + "task is incomplete. task_id:" + task_id + ", status:" + progressStatus.getDesc() + "\n");
  124.                         continue;
  125.                     }
  126.                 }
  127.             } catch (LfasrException e) {
  128.                 // 获取进度异常处理，根据返回信息排查问题后，再次进行获取
  129.                 Message progressMsg = JSON.parseObject(e.getMessage(), Message.class);
  130.                 //System.out.println(datePrefix + "ecode=" + progressMsg.getErr_no() + "\n");
  131.                 //System.out.println(datePrefix + "failed=" + progressMsg.getFailed() + "\n");
  132.                 logDetailText.append(datePrefix + "ecode=" + progressMsg.getErr_no() + "\n");
  133.                 logDetailText.append(datePrefix + "failed=" + progressMsg.getFailed() + "\n");
  134.             }
  135.         }
  136.  
  137.         // 获取任务结果
  138.         try {
  139.             Message resultMsg = lc.lfasrGetResult(task_id);
  140.             // 如果返回状态等于0，则获取任务结果成功
  141.             if (resultMsg.getOk() == 0) {
  142.                 // 打印转写结果
  143.             	String result = Test.getFinalResult(resultMsg.getData());
  144.             	String output = textPath + "\\" + System.currentTimeMillis() + ".txt";
  145.             	FileOutputStream f = new FileOutputStream(output);
  146.             	f.write(result.getBytes());
  147.                 //System.out.println(result);
  148.                 logDetailText.append(datePrefix + "结果存放路径： " + output + "\n");
  149.                 logDetailText.append(datePrefix + "最终转换结果： " + "\n");
  150.                 logDetailText.append(datePrefix + result + "\n");
  151.             } else {
  152.                 // 获取任务结果失败
  153.                 //System.out.println(datePrefix + "ecode=" + resultMsg.getErr_no() + "\n");
  154.                 //System.out.println(datePrefix + "failed=" + resultMsg.getFailed() + "\n");
  155.                 logDetailText.append(datePrefix + "ecode=" + resultMsg.getErr_no() + "\n");
  156.                 logDetailText.append(datePrefix + "failed=" + resultMsg.getFailed() + "\n");
  157.                 countDownLatch.countDown();
  158.                 return -1;
  159.             }
  160.         } catch (LfasrException e) {
  161.             // 获取结果异常处理，解析异常描述信息
  162.             Message resultMsg = JSON.parseObject(e.getMessage(), Message.class);
  163.             //System.out.println(datePrefix + "ecode=" + resultMsg.getErr_no() + "\n");
  164.             //System.out.println(datePrefix + "failed=" + resultMsg.getFailed() + "\n");
  165.             logDetailText.append(datePrefix + "ecode=" + resultMsg.getErr_no() + "\n");
  166.             logDetailText.append(datePrefix + "failed=" + resultMsg.getFailed() + "\n");
  167.             countDownLatch.countDown();
  168.             return -1;
  169.         }
  170. 		countDownLatch.countDown();
  171. 		return 1;
  172. 	}		
  173. }

整合代码，实现最终效果如下：

                                         

代码位置：https://github.com/ChenWenKaiVN/VoiceToText

下一阶段优化方向

1.主线程会出现假死现象。需要深入研究一下SWT UI线程与非UI线程的运行机制。

https://blog.csdn.net/dollyn/article/details/38582743/

2.研究一下进度条的问题，显示转换进度。

3.配置界面需要与SDK配置文件进一步相结合，许多变量还是写死在SDK配置文件中。

4.研究一下可执行jar的打包方法，将JRE一起加入到可执行jar中。

5.研究一下语音识别的技术原理

https://github.com/nl8590687/ASRT_SpeechRecognition

声明：本文内容由网友自发贡献，不代表【wpsshop博客】立场，版权归原作者所有，本站不承担相应法律责任。如您发现有侵权的内容，请联系我们。转载请注明出处：https://www.wpsshop.cn/w/我家小花儿/article/detail/225480
推荐阅读
相关标签

Copyright © 2003-2013 www.wpsshop.cn 版权所有，并保留所有权利。

  

闽ICP备14008679号