首页
学习
活动
专区
工具
TVP
发布
社区首页 >专栏 >c#语音识别(wav to text)

c#语音识别(wav to text)

作者头像
冰封一夏
发布2019-09-10 18:06:34
1.4K0
发布2019-09-10 18:06:34
举报

最近在搞一个语音识别的项目,wav to 文字,一开始写的代码在使用多线程的时候会出现识别超时的现象,不过后来终于google到解决方法。下面给出代码,不过遗憾的是资源释放不完全,希望给位给点建议。

public class SpRecognizer
    {
        private SpInProcRecoContext m_wavRecoContext;
        private ISpeechRecoGrammar m_Grammar;
        private SpFileStream m_infile;
        public CtiLine ctiLine;
        private string result = string.Empty;
        SpInprocRecognizer recognizer;

        public SpRecognizer()
        {
            try
            {
                recognizer = new SpInprocRecognizer();
                m_wavRecoContext = (SpInProcRecoContext)recognizer.CreateRecoContext();
                m_wavRecoContext.RetainedAudio = SpeechRetainedAudioOptions.SRAORetainAudio;
                m_infile = new SpFileStreamClass();
                m_infile.Format.GetWaveFormatEx();
            }
            catch (Exception ex)
            {
                SearchLog.Log.AppendLog("RecognizeWav", ex.ToString());
            }
        }

        public void OpenWAV()
        {
            try
            {
                m_Grammar = m_wavRecoContext.CreateGrammar(0);
                m_Grammar.DictationLoad("", SpeechLoadOption.SLOStatic);
                m_wavRecoContext.Recognition += new _ISpeechRecoContextEvents_RecognitionEventHandler(RecoContext_Recognition);
                m_wavRecoContext.EndStream += new _ISpeechRecoContextEvents_EndStreamEventHandler(RecoContext_EndRecognition);
                m_infile.Open(ctiLine.MStation.RecorderFilePath, SpeechStreamFileMode.SSFMOpenForRead, false);
                m_wavRecoContext.Recognizer.AudioInputStream = m_infile;
                m_Grammar.DictationSetState(SpeechRuleState.SGDSActive);
            }
            catch (Exception ex)
            {
                SearchLog.Log.AppendLog("OpenWAV", ex.ToString());
            }
        }

        void RecoContext_Recognition(int StreamNumber, object StreamPosition, SpeechRecognitionType RecognitionType, ISpeechRecoResult Result)
        {
            if (m_infile != null)
            {
                result += Result.PhraseInfo.GetText(0, -1, true);
            }
        }

        void RecoContext_EndRecognition(int StreamNumber, object StreamPosition, bool f)
        {
            ctiLine.MStation.MobileIsStation = IsMobileStation();
            SearchLog.Log.AppendLog("(" + ctiLine.MStation.MobileNo + ")" + "Recognize Result", result);       
            m_wavRecoContext.Recognizer.AudioInputStream = null;
            m_Grammar.DictationUnload();
            ctiLine.IsRecognized = true;        
            m_infile.Close();        
            m_infile = null;
            m_wavRecoContext = null;
            m_Grammar = null;
            recognizer = null;
        }
本文参与 腾讯云自媒体分享计划,分享自作者个人站点/博客。
原始发表:2012-10-31 ,如有侵权请联系 cloudcommunity@tencent.com 删除

本文分享自 作者个人站点/博客 前往查看

如有侵权,请联系 cloudcommunity@tencent.com 删除。

本文参与 腾讯云自媒体分享计划  ,欢迎热爱写作的你一起参与!

评论
登录后参与评论
0 条评论
热度
最新
推荐阅读
相关产品与服务
语音识别
腾讯云语音识别(Automatic Speech Recognition,ASR)是将语音转化成文字的PaaS产品,为企业提供精准而极具性价比的识别服务。被微信、王者荣耀、腾讯视频等大量业务使用,适用于录音质检、会议实时转写、语音输入法等多个场景。
领券
问题归档专栏文章快讯文章归档关键词归档开发者手册归档开发者手册 Section 归档