前奏,浑浑噩噩已经工作一年多,这一年多收获还是挺多的。逛园子应该有两年多了,工作后基本上是天天都会来园子逛逛,园子 里还是有很多牛人写了一些不错的博客,帮我解决很多问题。但是一直没写过博客,归根到底一个字“懒”,还有就是不知道该写 些什么...
今天把我最近研究讯飞语音东东,分享一下,不过有些还是前辈们提供的。之前公司让我做一个小的语音识别功能,一开始我就建议使用讯飞语音,个人觉得讯飞识别正确率还是可观的。可是老总说不能考 虑联网,还有就是钱的问题。想到微软自带语音识别引擎(基于win7)。第一次接触到语音识别,没什么头绪,只有收集相关资料 。最后成品出来了,但是识别效果不是那么满意,老总说那就将就用吧,我想他都那样说了,我也没多大意见...开年后老总老总 买了个iphone5玩了siri后,觉得我们现在那个语音太丑陋了,让换一个解决方案,识别率要达到90%。国内有好几家公司做语音识 别的比如科大讯飞、云知声、捷通华声以及紫冬锐意语音都做了一定开放。市面上我知道语音助手有百度语音助手、虫洞语音助手 、360语音助手以及科大讯飞语点;前两个我不知道采用那家公司的还是自己研发的不过都没开开放,360语音助手采用讯飞的。然 后我就继续提议使用讯飞语音SDK,于是乎同事们都下载讯飞语点来玩玩,老总说那就试试讯飞语音SDK。先做一个简单demo,看看识别效果,感觉识别率上能够满足要求。一般要的结果不光只是要把所说的话翻译成文字,而是需要的是 语义的理解:例如我要去北京,直接返回北京这个关键。目前讯飞还没把这个接口开放出来,公司负责人说今年会把这个开放出来 。那现在只能使用关键词识别语法,一种是直接是文本词库;另一种是ABNF语法。ABNF写法有点烦杂(语法文件里使用到的词句都是指定的,对于无法枚举的词句暂时没有很好的解决办法。),就直接采用文本词库(因为关键词有点多)。讯飞还有专门人负责关于讯飞语音问题解答(QQ群:153789256)。讯飞提供msc.dll这个DLL,调用DLL的封装:
////// MSCDLL入口封装 /// private class MscDll { #region MscDLL ////// 初始化MSC的ISR部分 /// /// 初始化时传入的字符串,以指定合成用到的一些配置参数,各个参数以“参数名=参数值”的形式出现,大小写不敏感,不同的参数之间以“,”或“\n”隔开,不设置任何值时可以传入NULL或空串: ///如果函数调用成功返回MSP_SUCCESS,否则返回错误代码,错误代码参见msp_errors [DllImport("msc.dll", CallingConvention = CallingConvention.StdCall)] public static extern int QISRInit(string configs); ////// 开始一个ISR会话 /// /// uri-list格式的语法,可以是一个语法文件的URL或者一个引擎内置语法列表。可以同时指定多个语法,不同的语法之间以“,” /// 隔开。进行语音听写时不需要语法,此参数设定为NULL或空串即可;进行语音识别时则需要语法,语法可以在此参数中指定,也可以随后调用QISRGrammarActivate指定识别所用的语法。 /// 本路ISR会话使用的参数,可设置的参数及其取值范围请参考《可设置参数列表_MSP20.xls》,各个参数以“参数名=参数值” 的形式出现,不同的参数之间以“,”或者“\n”隔开。 /// 如果函数调用成功则其值为MSP_SUCCESS,否则返回错误代码,错误代码参见msp_errors。几个主要的返回值: MSP_ERROR_NOT_INIT 未初始化 MSP_ERROR_INVALID_PARA 无效的参数; MSP_ERROR_NO_LICENSE 开始一路会话失败 ///MSC为本路会话建立的ID,用来唯一的标识本路会话,供以后调用其他函数时使用。函数调用失败则会返回NULL。 [DllImport("msc.dll", CallingConvention = CallingConvention.StdCall)] public static extern IntPtr QISRSessionBegin(string grammarList, string _params, ref int errorCode); ////// 传入语法 /// /// 由QISRSessionBegin返回过来的会话ID /// 语法字符串 /// 语法类型,可以是uri-list、abnf、xml等 /// 本次传入语法的权重,本参数在MSP 2.0中会被忽略。 ///如果函数调用成功返回MSP_SUCCESS,否则返回错误代码,错误代码参见msp_errors [DllImport("msc.dll", CallingConvention = CallingConvention.StdCall)] public static extern int QISRGrammarActivate(string sessionID, string grammar, string type, int weight); ////// 写入用来识别的语音 /// /// 由QISRSessionBegin返回过来的会话ID /// 音频数据缓冲区起始地址 /// 音频数据长度,其大小不能超过设定的max_audio_size /// 用来指明用户本次识别的音频是否发送完毕,可能值如下: /// MSP_AUDIO_SAMPLE_FIRST = 1 第一块音频 /// MSP_AUDIO_SAMPLE_CONTINUE = 2 还有后继音频 /// MSP_AUDIO_SAMPLE_LAST = 4 最后一块音频 /// 端点检测(End-point detected)器所处的状态,可能的值如下: /// MSP _EP_LOOKING_FOR_SPEECH = 0 还没有检测到音频的前端点。 /// MSP _EP_IN_SPEECH = 1 已经检测到了音频前端点,正在进行正常的音频处理。 /// MSP _EP_AFTER_SPEECH = 3 检测到音频的后端点,后继的音频会被MSC忽略。 /// MSP _EP_TIMEOUT = 4 超时。 /// MSP _EP_ERROR= 5 出现错误。 /// MSP _EP_MAX_SPEECH = 6 音频过大。 /// 识别器所处的状态 ///[DllImport("msc.dll", CallingConvention = CallingConvention.StdCall)] public static extern int QISRAudioWrite(string sessionID, byte[] waveData, uint waveLen, AudioStatus audioStatus, ref EpStatus epStatus, ref RecogStatus recogStatus); /// /// 获取识别结果 /// /// 由QISRSessionBegin返回过来的会话ID /// 识别结果的状态,其取值范围和含义请参考QISRAudioWrite的参数recogStatus /// 与服务器交互的间隔时间,可以控制和服务器的交互频度。单位为ms,建议取值为5000 /// 如果函数调用成功返回MSP_SUCCESS,否则返回错误代码,错误代码参见msp_errors ///函数执行成功并且获取到识别结果时返回识别结果,函数执行成功没有获取到识别结果时返回NULL [DllImport("msc.dll", CallingConvention = CallingConvention.StdCall)] public static extern IntPtr QISRGetResult(string sessionID, ref RecogStatus rsltStatus, int waitTime, ref int errorCode); ////// 结束一路会话 /// /// 由QISRSessionBegin返回过来的会话ID /// 结束本次会话的原因描述,用于记录日志,便于用户查阅或者跟踪某些问题 ///[DllImport("msc.dll", CallingConvention = CallingConvention.StdCall)] public static extern int QISRSessionEnd(string sessionID, string hints); /// /// 获取与识别交互相关的参数 /// /// 由QISRSessionBegin返回过来的会话ID /// 要获取的参数名称;支持同时查询多个参数,查询多个参数时,参数名称按“,” 或“\n”分隔开来 /// 获取的参数值,以字符串形式返回;查询多个参数时,参数值之间以“;”分开,不支持的参数将返回空的值 /// 参数值的长度 ///[DllImport("msc.dll", CallingConvention = CallingConvention.StdCall)] public static extern int QISRGetParam(string sessionID, string paramName, string paramValue, ref uint valueLen); /// /// 逆初始化MSC的ISR部分 /// ///[DllImport("msc.dll", CallingConvention = CallingConvention.StdCall)] public static extern int QISRFini(); /// /// 上传用户自定义词库 /// /// 由QISRSessionBegin返回过来的会话ID /// 词库名称 /// 词库数据采用是utf8格式 /// 词库大小 /// 参数值 /// 如果函数调用成功返回MSP_SUCCESS,否则返回错误代码,错误代码参见msp_errors ///函数执行成功并且返回exID(词库编号) [DllImport("msc.dll", CallingConvention = CallingConvention.StdCall)] public static extern IntPtr QISRUploadData(string sessionID, string dataName, byte[] userData, uint lenght, string paramValue, ref int errorCode); #endregion }
说明一下:“QISRUploadData”(上传词库)这个函数在开发文档里面没的,讯飞遗漏了。
具体实现(关键词识别,文本词库):
类:MscNet
View Code
1 #region 定义字段 2 3 //返回错误代号 4 private int ret = 0; 5 private RecoErrors Re = null; 6 ///7 /// 会话ID 8 /// 9 private string sess_id = null; 10 ///11 /// 参数 12 /// 13 private string param = null; 14 ///15 /// 语法 16 /// 17 private string grammar = null; 18 //错误消息通知托管 19 public delegate void delegdateOnerror(string Msg); 20 private string path = null; 21 ///22 /// 识别数据返回的事件 23 /// 24 public event EventHandlerDataArrived; 25 /// 26 /// 识别过程结束的事件 27 /// 28 public event EventHandler ISREnd; 29 ///30 /// 正在识别 31 /// 32 public event EventHandler Spoting; 33 34 #endregion 35 36 #region 构造函数 37 38 ///39 /// 构造函数,初始化引擎 40 /// 41 /// appid 42 /// 参数 43 /// 语法ID 44 /// 路径 45 public MscNet(string appid, string param, string path, string grammar) 46 { 47 this.path = path; 48 this.param = param; 49 this.grammar = grammar; 50 Re = new RecoErrors(); 51 //引擎初始化,只需初始化一次 52 ret = MscDll.QISRInit("appid=" + appid); 53 try 54 { 55 Re.GetError(ret); 56 } 57 catch (MscException ex) 58 { 59 RaiseError(ex.Message.ToString()); 60 } 61 } 62 #endregion 63 64 #region 公共方法 65 66 ///67 /// 开始识别语音 68 /// 69 /// 音频数据 70 public void InterpretVoice(byte[] buffer) 71 { 72 //用来指明用户本次识别的音频是否发送完毕 73 AudioStatus audioStatus = AudioStatus.ISR_AUDIO_SAMPLE_LAST; 74 //端点检测(End-point detected)器所处的状态 75 EpStatus ep_status = EpStatus.ISR_EP_NULL; 76 //识别器所处的状态 77 RecogStatus rec_status = RecogStatus.ISR_REC_NULL; 78 //识别结果的状态 79 RecogStatus rslt_status = RecogStatus.ISR_REC_NULL; 80 Loadgrammar(); 81 int ret = MscDll.QISRAudioWrite(sess_id, buffer, (uint)buffer.Length, audioStatus, ref ep_status, ref rec_status); 82 try 83 { 84 Re.GetError(ret); 85 do 86 { 87 if (rslt_status == RecogStatus.ISR_REC_STATUS_INCOMPLETE) 88 { 89 Spoting(this, new EventArgs());//通知正在识别 90 } 91 IntPtr p = MscDll.QISRGetResult(sess_id, ref rslt_status, 0, ref ret); 92 Re.GetError(ret); 93 if (p != IntPtr.Zero) 94 { 95 string tmp = PtrToStr(p); 96 DataArrived(this, new DataArrivedEventArgs(tmp));//激发识别数据到达事件 97 } 98 System.Threading.Thread.Sleep(500); 99 } while (rslt_status != RecogStatus.ISR_REC_STATUS_SPEECH_COMPLETE);100 }101 catch (MscException ex)102 {103 RaiseError(ex.Message);104 }105 finally106 {107 try108 {109 ret = MscDll.QISRSessionEnd(sess_id, string.Empty);110 Re.GetError(ret);111 ISREnd(this, new EventArgs());//通知识别结束 112 }113 catch (MscException ex)114 {115 RaiseError(ex.Message);116 }117 }118 }119 120 ///121 /// 上传词库词库采用是utf8格式122 /// 123 /// 词库名称124 /// 词库路径125 /// 参数126 ///返回词库编号 127 public string GetExID(string txtFile, string path, string param)128 {129 string filePath = Path.Combine(path, txtFile);130 string tmp = string.Empty;131 if (!string.IsNullOrEmpty(filePath))132 {133 try134 {135 sess_id = PtrToStr(MscDll.QISRSessionBegin(null, param, ref ret));136 Re.GetError(ret);137 try138 {139 using (FileStream fs = File.Open(filePath, FileMode.Open))140 {141 long len = fs.Length;142 byte[] buffer = new byte[len];143 fs.Read(buffer, 0, (int)len);144 IntPtr p = MscDll.QISRUploadData(sess_id, "sces", buffer, (uint)fs.Length, "dtt=keylist", ref ret);145 Re.GetError(ret);146 if (p != IntPtr.Zero)147 {148 tmp = PtrToStr(p);149 }150 }151 }152 catch (FileNotFoundException ex)153 {154 RaiseError(ex.Message);155 }156 157 }158 catch (MscException ex)159 {160 RaiseError(ex.Message);161 }162 finally163 {164 try165 {166 ret = MscDll.QISRSessionEnd(sess_id, null);167 Re.GetError(ret);168 }169 catch (MscException ex)170 {171 RaiseError(ex.Message.ToString());172 }173 }174 }175 else176 {177 RaiseError("路径不正确!");178 }179 return tmp;180 }181 182 ///183 /// 对MSC的ISR部分进行逆初始化。184 /// 185 public void QISRFini()186 {187 try188 {189 ret = MscDll.QISRFini();190 Re.GetError(ret);191 }192 catch (MscException ex)193 {194 RaiseError(ex.Message.ToString());195 }196 }197 198 #endregion199 200 #region 受保护方法201 ///202 /// 加载语法 203 /// 204 private void Loadgrammar()205 {206 try207 {208 sess_id = PtrToStr(MscDll.QISRSessionBegin(grammar, param, ref ret));209 Re.GetError(ret);210 }211 catch (MscException ex)212 {213 RaiseError(ex.Message);214 }215 }216 217 ///218 /// 指针转字符串219 /// 220 /// 指向非托管代码字符串的指针221 ///返回指针指向的字符串 222 private string PtrToStr(IntPtr p)223 {224 Listlb = new List ();225 try226 {227 while (Marshal.ReadByte(p) != 0)228 {229 lb.Add(Marshal.ReadByte(p));230 p = p + 1;231 }232 }233 catch (AccessViolationException ex)234 {235 RaiseError(ex.Message);236 }237 return Encoding.Default.GetString(lb.ToArray());238 } 239 #endregion240 241 #region 事件242 /// 243 /// 错误通知事件244 /// 245 public event delegdateOnerror OnError;246 private void RaiseError(string Msg)247 {248 if (OnError != null)249 {250 OnError(Msg);251 }252 }253 ///254 /// 有识别数据返回的事件参数,包含了识别的文本结果255 /// 256 public class DataArrivedEventArgs : EventArgs257 {258 public string result;259 public DataArrivedEventArgs(string rs)260 {261 result = rs;262 }263 }264 #endregion
如果采用ABNF语法,只是与文本词库加载语法方式有点不一样:
View Code
1 ///2 /// 加载语法 3 /// 4 private void Loadgrammar() 5 { 6 try 7 { 8 9 //sess_id = PtrToStr(MscDll.QISRSessionBegin(grammar, param, ref ret)); 10 /*ABNF语法*/11 sess_id = PtrToStr(MscDll.QISRSessionBegin(null, "rst=plain,sub=asr,ssm=1,aue=speex,auf=audio/L16;rate=16000,cfd=350", ref ret));12 em.GetError(ret);13 string grammar1 = "#ABNF 1.0 GB2312;\n mode voice;\n language zh-cn;\n\n\n root $main;\n public $main = 我[想要]看$place1;\n $place1=足球;\n";14 ret = MscDll.QISRGrammarActivate(sess_id, grammar1, "abnf", 0);//将语法ID传入QISRSessionBegin15 /*end */16 em.GetError(ret); 17 }18 catch (MscException ex)19 {20 RaiseError(ex.Message.ToString());21 }22 }
常量的枚举:
View Code
1 #region 错误代号 2 enum ErrorCode 3 { 4 MSP_SUCCESS= 0, 5 MSP_ERROR_FAIL = -1, 6 MSP_ERROR_EXCEPTION = -2, 7 8 /* General errors 10100(0x2774) */ 9 MSP_ERROR_GENERAL = 10100, /* 0x2774 */ 10 MSP_ERROR_OUT_OF_MEMORY = 10101, /* 0x2775 */ 11 MSP_ERROR_FILE_NOT_FOUND = 10102, /* 0x2776 */ 12 MSP_ERROR_NOT_SUPPORT = 10103, /* 0x2777 */ 13 MSP_ERROR_NOT_IMPLEMENT = 10104, /* 0x2778 */ 14 MSP_ERROR_ACCESS = 10105, /* 0x2779 */ 15 MSP_ERROR_INVALID_PARA = 10106, /* 0x277A */ 16 MSP_ERROR_INVALID_PARA_VALUE = 10107, /* 0x277B */ 17 MSP_ERROR_INVALID_HANDLE = 10108, /* 0x277C */ 18 MSP_ERROR_INVALID_DATA = 10109, /* 0x277D */ 19 MSP_ERROR_NO_LICENSE = 10110, /* 0x277E */ 20 MSP_ERROR_NOT_INIT = 10111, /* 0x277F */ 21 MSP_ERROR_NULL_HANDLE = 10112, /* 0x2780 */ 22 MSP_ERROR_OVERFLOW = 10113, /* 0x2781 */ 23 MSP_ERROR_TIME_OUT = 10114, /* 0x2782 */ 24 MSP_ERROR_OPEN_FILE = 10115, /* 0x2783 */ 25 MSP_ERROR_NOT_FOUND = 10116, /* 0x2784 */ 26 MSP_ERROR_NO_ENOUGH_BUFFER = 10117, /* 0x2785 */ 27 MSP_ERROR_NO_DATA = 10118, /* 0x2786 */ 28 MSP_ERROR_NO_MORE_DATA = 10119, /* 0x2787 */ 29 MSP_ERROR_SKIPPED = 10120, /* 0x2788 */ 30 MSP_ERROR_ALREADY_EXIST = 10121, /* 0x2789 */ 31 MSP_ERROR_LOAD_MODULE = 10122, /* 0x278A */ 32 MSP_ERROR_BUSY = 10123, /* 0x278B */ 33 MSP_ERROR_INVALID_CONFIG = 10124, /* 0x278C */ 34 MSP_ERROR_VERSION_CHECK = 10125, /* 0x278D */ 35 MSP_ERROR_CANCELED = 10126, /* 0x278E */ 36 MSP_ERROR_INVALID_MEDIA_TYPE = 10127, /* 0x278F */ 37 MSP_ERROR_CONFIG_INITIALIZE = 10128, /* 0x2790 */ 38 MSP_ERROR_CREATE_HANDLE = 10129, /* 0x2791 */ 39 MSP_ERROR_CODING_LIB_NOT_LOAD = 10130, /* 0x2792 */ 40 41 /* Error codes of network 10200(0x27D8)*/ 42 MSP_ERROR_NET_GENERAL = 10200, /* 0x27D8 */ 43 MSP_ERROR_NET_OPENSOCK = 10201, /* 0x27D9 */ /* Open socket */ 44 MSP_ERROR_NET_CONNECTSOCK = 10202, /* 0x27DA */ /* Connect socket */ 45 MSP_ERROR_NET_ACCEPTSOCK = 10203, /* 0x27DB */ /* Accept socket */ 46 MSP_ERROR_NET_SENDSOCK = 10204, /* 0x27DC */ /* Send socket data */ 47 MSP_ERROR_NET_RECVSOCK = 10205, /* 0x27DD */ /* Recv socket data */ 48 MSP_ERROR_NET_INVALIDSOCK = 10206, /* 0x27DE */ /* Invalid socket handle */ 49 MSP_ERROR_NET_BADADDRESS = 10207, /* 0x27EF */ /* Bad network address */ 50 MSP_ERROR_NET_BINDSEQUENCE = 10208, /* 0x27E0 */ /* Bind after listen/connect */ 51 MSP_ERROR_NET_NOTOPENSOCK = 10209, /* 0x27E1 */ /* Socket is not opened */ 52 MSP_ERROR_NET_NOTBIND = 10210, /* 0x27E2 */ /* Socket is not bind to an address */ 53 MSP_ERROR_NET_NOTLISTEN = 10211, /* 0x27E3 */ /* Socket is not listenning */ 54 MSP_ERROR_NET_CONNECTCLOSE = 10212, /* 0x27E4 */ /* The other side of connection is closed */ 55 MSP_ERROR_NET_NOTDGRAMSOCK = 10213, /* 0x27E5 */ /* The socket is not datagram type */ 56 MSP_ERROR_NET_DNS=10214, 57 /* Error codes of mssp message 10300(0x283C) */ 58 MSP_ERROR_MSG_GENERAL = 10300, /* 0x283C */ 59 MSP_ERROR_MSG_PARSE_ERROR = 10301, /* 0x283D */ 60 MSP_ERROR_MSG_BUILD_ERROR = 10302, /* 0x283E */ 61 MSP_ERROR_MSG_PARAM_ERROR = 10303, /* 0x283F */ 62 MSP_ERROR_MSG_CONTENT_EMPTY = 10304, /* 0x2840 */ 63 MSP_ERROR_MSG_INVALID_CONTENT_TYPE = 10305, /* 0x2841 */ 64 MSP_ERROR_MSG_INVALID_CONTENT_LENGTH = 10306, /* 0x2842 */ 65 MSP_ERROR_MSG_INVALID_CONTENT_ENCODE = 10307, /* 0x2843 */ 66 MSP_ERROR_MSG_INVALID_KEY = 10308, /* 0x2844 */ 67 MSP_ERROR_MSG_KEY_EMPTY = 10309, /* 0x2845 */ 68 MSP_ERROR_MSG_SESSION_ID_EMPTY = 10310, /* 0x2846 */ 69 MSP_ERROR_MSG_LOGIN_ID_EMPTY = 10311, /* 0x2847 */ 70 MSP_ERROR_MSG_SYNC_ID_EMPTY = 10312, /* 0x2848 */ 71 MSP_ERROR_MSG_APP_ID_EMPTY = 10313, /* 0x2849 */ 72 MSP_ERROR_MSG_EXTERN_ID_EMPTY = 10314, /* 0x284A */ 73 MSP_ERROR_MSG_INVALID_CMD = 10315, /* 0x284B */ 74 MSP_ERROR_MSG_INVALID_SUBJECT = 10316, /* 0x284C */ 75 MSP_ERROR_MSG_INVALID_VERSION = 10317, /* 0x284D */ 76 MSP_ERROR_MSG_NO_CMD = 10318, /* 0x284E */ 77 MSP_ERROR_MSG_NO_SUBJECT = 10319, /* 0x284F */ 78 MSP_ERROR_MSG_NO_VERSION = 10320, /* 0x2850 */ 79 MSP_ERROR_MSG_MSSP_EMPTY = 10321, /* 0x2851 */ 80 MSP_ERROR_MSG_NEW_RESPONSE = 10322, /* 0x2852 */ 81 MSP_ERROR_MSG_NEW_CONTENT = 10323, /* 0x2853 */ 82 MSP_ERROR_MSG_INVALID_SESSION_ID = 10324, /* 0x2854 */ 83 84 /* Error codes of DataBase 10400(0x28A0)*/ 85 MSP_ERROR_DB_GENERAL = 10400, /* 0x28A0 */ 86 MSP_ERROR_DB_EXCEPTION = 10401, /* 0x28A1 */ 87 MSP_ERROR_DB_NO_RESULT = 10402, /* 0x28A2 */ 88 MSP_ERROR_DB_INVALID_USER = 10403, /* 0x28A3 */ 89 MSP_ERROR_DB_INVALID_PWD = 10404, /* 0x28A4 */ 90 MSP_ERROR_DB_CONNECT = 10405, /* 0x28A5 */ 91 MSP_ERROR_DB_INVALID_SQL = 10406, /* 0x28A6 */ 92 MSP_ERROR_DB_INVALID_APPID = 10407, /* 0x28A7 */ 93 94 /* Error codes of Resource 10500(0x2904)*/ 95 MSP_ERROR_RES_GENERAL = 10500, /* 0x2904 */ 96 MSP_ERROR_RES_LOAD = 10501, /* 0x2905 */ /* Load resource */ 97 MSP_ERROR_RES_FREE = 10502, /* 0x2906 */ /* Free resource */ 98 MSP_ERROR_RES_MISSING = 10503, /* 0x2907 */ /* Resource File Missing */ 99 MSP_ERROR_RES_INVALID_NAME = 10504, /* 0x2908 */ /* Invalid resource file name */100 MSP_ERROR_RES_INVALID_ID = 10505, /* 0x2909 */ /* Invalid resource ID */101 MSP_ERROR_RES_INVALID_IMG = 10506, /* 0x290A */ /* Invalid resource image pointer */102 MSP_ERROR_RES_WRITE = 10507, /* 0x290B */ /* Write read-only resource */103 MSP_ERROR_RES_LEAK = 10508, /* 0x290C */ /* Resource leak out */104 MSP_ERROR_RES_HEAD = 10509, /* 0x290D */ /* Resource head currupt */105 MSP_ERROR_RES_DATA = 10510, /* 0x290E */ /* Resource data currupt */106 MSP_ERROR_RES_SKIP = 10511, /* 0x290F */ /* Resource file skipped */107 108 /* Error codes of TTS 10600(0x2968)*/109 MSP_ERROR_TTS_GENERAL = 10600, /* 0x2968 */110 MSP_ERROR_TTS_TEXTEND = 10601, /* 0x2969 */ /* Meet text end */111 MSP_ERROR_TTS_TEXT_EMPTY = 10602, /* 0x296A */ /* no synth text */112 113 /* Error codes of Recognizer 10700(0x29CC) */114 MSP_ERROR_REC_GENERAL = 10700, /* 0x29CC */115 MSP_ERROR_REC_INACTIVE = 10701, /* 0x29CD */116 MSP_ERROR_REC_GRAMMAR_ERROR = 10702, /* 0x29CE */117 MSP_ERROR_REC_NO_ACTIVE_GRAMMARS = 10703, /* 0x29CF */118 MSP_ERROR_REC_DUPLICATE_GRAMMAR = 10704, /* 0x29D0 */119 MSP_ERROR_REC_INVALID_MEDIA_TYPE = 10705, /* 0x29D1 */120 MSP_ERROR_REC_INVALID_LANGUAGE = 10706, /* 0x29D2 */121 MSP_ERROR_REC_URI_NOT_FOUND = 10707, /* 0x29D3 */122 MSP_ERROR_REC_URI_TIMEOUT = 10708, /* 0x29D4 */123 MSP_ERROR_REC_URI_FETCH_ERROR = 10709, /* 0x29D5 */124 125 /* Error codes of Speech Detector 10800(0x2A30) */126 MSP_ERROR_EP_GENERAL = 10800, /* 0x2A30 */127 MSP_ERROR_EP_NO_SESSION_NAME = 10801, /* 0x2A31 */128 MSP_ERROR_EP_INACTIVE = 10802, /* 0x2A32 */129 MSP_ERROR_EP_INITIALIZED = 10803, /* 0x2A33 */130 131 /* Error codes of TUV */132 MSP_ERROR_TUV_GENERAL = 10900, /* 0x2A94 */133 MSP_ERROR_TUV_GETHIDPARAM = 10901, /* 0x2A95 */ /* Get Busin Param huanid*/134 MSP_ERROR_TUV_TOKEN = 10902, /* 0x2A96 */ /* Get Token */135 MSP_ERROR_TUV_CFGFILE = 10903, /* 0x2A97 */ /* Open cfg file */136 MSP_ERROR_TUV_RECV_CONTENT = 10904, /* 0x2A98 */ /* received content is error */137 MSP_ERROR_TUV_VERFAIL = 10905, /* 0x2A99 */ /* Verify failure */138 139 /* Error codes of IMTV */140 MSP_ERROR_IMTV_SUCCESS = 11000, /* 0x2AF8 */ /* 成功 */141 MSP_ERROR_IMTV_NO_LICENSE = 11001, /* 0x2AF9 */ /* 试用次数结束,用户需要付费 */142 MSP_ERROR_IMTV_SESSIONID_INVALID = 11002, /* 0x2AFA */ /* SessionId失效,需要重新登录通行证 */143 MSP_ERROR_IMTV_SESSIONID_ERROR = 11003, /* 0x2AFB */ /* SessionId为空,或者非法 */144 MSP_ERROR_IMTV_UNLOGIN = 11004, /* 0x2AFC */ /* 未登录通行证 */145 MSP_ERROR_IMTV_SYSTEM_ERROR = 11005, /* 0x2AFD */ /* 系统错误 */146 147 /* Error codes of HCR */148 MSP_ERROR_HCR_GENERAL = 11100,149 MSP_ERROR_HCR_RESOURCE_NOT_EXIST = 11101,150 151 /* Error codes of http 12000(0x2EE0) */152 MSP_ERROR_HTTP_BASE = 12000, /* 0x2EE0 */153 154 /*Error codes of ISV */155 MSP_ERROR_ISV_NO_USER = 13000, /* 32C8 */ /* the user doesn't exist */156 }157 #endregion158 159 #region ISR枚举常量160 public enum AudioStatus161 {162 ISR_AUDIO_SAMPLE_INIT = 0x00,163 ISR_AUDIO_SAMPLE_FIRST = 0x01,164 ISR_AUDIO_SAMPLE_CONTINUE = 0x02,165 ISR_AUDIO_SAMPLE_LAST = 0x04,166 ISR_AUDIO_SAMPLE_SUPPRESSED = 0x08,167 ISR_AUDIO_SAMPLE_LOST = 0x10,168 ISR_AUDIO_SAMPLE_NEW_CHUNK = 0x20,169 ISR_AUDIO_SAMPLE_END_CHUNK = 0x40,170 171 ISR_AUDIO_SAMPLE_VALIDBITS = 0x7f /* to validate the value of sample->status */172 }173 174 public enum EpStatus175 {176 ISR_EP_NULL = -1,177 ISR_EP_LOOKING_FOR_SPEECH = 0, ///还没有检测到音频的前端点178 ISR_EP_IN_SPEECH = 1, ///已经检测到了音频前端点,正在进行正常的音频处理。179 ISR_EP_AFTER_SPEECH = 3, ///检测到音频的后端点,后继的音频会被MSC忽略。180 ISR_EP_TIMEOUT = 4, ///超时181 ISR_EP_ERROR = 5, ///出现错误182 ISR_EP_MAX_SPEECH = 6 ///音频过大183 }184 185 public enum RecogStatus186 {187 ISR_REC_NULL = -1,188 ISR_REC_STATUS_SUCCESS = 0, ///识别成功,此时用户可以调用QISRGetResult来获取(部分)结果。189 ISR_REC_STATUS_NO_MATCH = 1, ///识别结束,没有识别结果190 ISR_REC_STATUS_INCOMPLETE = 2, ///正在识别中191 ISR_REC_STATUS_NON_SPEECH_DETECTED = 3, ///保留192 ISR_REC_STATUS_SPEECH_DETECTED = 4, ///发现有效音频193 ISR_REC_STATUS_SPEECH_COMPLETE = 5, ///识别结束194 ISR_REC_STATUS_MAX_CPU_TIME = 6, ///保留195 ISR_REC_STATUS_MAX_SPEECH = 7, ///保留196 ISR_REC_STATUS_STOPPED = 8, ///保留197 ISR_REC_STATUS_REJECTED = 9, ///保留198 ISR_REC_STATUS_NO_SPEECH_FOUND = 10 ///没有发现音频199 }200 #endregion201 }
自定义异常:
View Code
讯飞语音支持边录边上传,不过我这里采用是一次性上传。起初我采用的是边录边上传,不过感觉有数字混合后识别正常率不好(还没跟讯飞那边沟通。),最后才使用一次性上传,毕竟语音文件也不是大就200KB一下。讯飞语音识别不支持多线程识别。 1 [Serializable] //声明为可序列化的 因为要写入文件中 2 public class MscException : ApplicationException//由用户程序引发,用于派生自定义的异常类型 3 { 4 ///5 /// 默认构造函数 6 /// 7 public MscException() { } 8 public MscException(string message) 9 : base(message) { }10 public MscException(string message, Exception inner)11 : base(message, inner) { }12 } 13 ///14 /// 是否出错15 /// 16 internal class RecoErrors17 {18 public RecoErrors() { }19 ///20 /// 是否发生错误. 21 /// 22 /// 错误ID 23 public virtual void GetError(int id)24 {25 if (id != 0)26 { 27 var ex = new MscException(((ErrorCode)id).ToString("G"));28 // var ex = new MscException(Enum.GetName(typeof(ErrorCode),id));29 throw ex;30 }31 }32 }
我做的这个语音识别产品,做成服务端与客户端。服务端:放在一个能连接外网机子上提供语音识别(做了一个简单队列),客户端:将音频数据采集后发送到局域网内的语音识别服务端进行识别。
以上有些代码是借助别人的,第一次写大家尽量不要吐槽,不过可以给点意见。大家相互学习...