来自Speech的文本会发生什么，它不能识别为命令，MS Speech SDK 11 [英] What happens to the text from Speech which doesn't recognize as command, MS Speech SDK 11

问题描述

大家好，



我正在使用MS Speech SDK v11处理一些启用语音的应用程序。我想识别我在SRGS文档中指定的命令，同时，如果有人说别的话，我希望将其转换为简单文本并存储。我尝试使用听写模式，这样我就可以继续听写并获取文本。然后使用自定义构建的SRGS解析器确定命令。如果文本不是预定义命令，则将其视为简单文本。现在的问题是AppendDictation因为找不到这个语法所引用的语法而无效。我提到了这个http://stackoverflow.com/questions/9347346/appenddictation-on-microsoft-speech-platform-11-server



我从微软改变了。发言到System.Speech并有趣地发现Microsoft.Speech至少认识到我指定的命令，但System.Speech什么也没做。我很困惑该怎么做。



我重复一遍，我有简单的要求，认识我指定的命令，如果不是只是翻译文本形式。如果这是不可能的，那么我怎样才能为口述模式制作语音API。



也只是一个简单的问题，如前所述，我们必须训练计算机到了解我们的演讲并创建个人资料，当前的演讲引擎是否也必须经过培训？或者经常说话时它会很好地调整自己吗？



提前谢谢你！

Hi All,

I am working on some speech enabled application using MS Speech SDK v11. I want to recognize the commands I specify in my SRGS document, at the same time, if some speak something else, I want that to be converted to simple text and be stored. I tried using Dictation mode, so that I can just continue in Dictation and get the text. Then using custom built SRGS parser, determine the commands. If the text is not predefined command, just treat it as simple text. Now problem is the AppendDictation is not working due to "Cannot find grammar referenced by this grammar." and I referred this http://stackoverflow.com/questions/9347346/appenddictation-on-microsoft-speech-platform-11-server

I changed from Microsoft.Speech to System.Speech and interestingly found that Microsoft.Speech at least recognize the commands I specify but System.Speech is doing nothing. I am confused what to do.

I repeat, I have simple requirement, recognize the commands I specify and if not just translate those in Text form. If this is not possible then how can I make work Speech API for dictation mode only.

Also just a quick question, as previously, we had to train the computer to understand our speech and create profile, does current speech engine also must be trained? Or it fine tunes itself as we speak frequently?

Thank you in advance!

推荐答案

对。您可以从Microsoft下载的识别引擎的质量不足以进行听写。如果你的语法相当小，它的效果相当不错，但即使这样你也可能需要使用更好的发音。 :-)实际上，有人说有更好的商业引擎，包括从微软授权给现在的Windows引擎的公司。我不知道还有什么建议，可能要等到技术改进。 :-)

-SA

Right. The quality of the recognition engine which you can download from Microsoft is not good enough for Dictation. If your grammar is reasonably small, it works reasonably well, but even then you might need to use better pronunciation. :-) Actually, some say there are much better commercial engines, including the one from the company which licensed to Microsoft the present-day engine for Windows. I don't know what else to suggest, probably wait until technology improves. :-)

—SA

这篇关于来自Speech的文本会发生什么，它不能识别为命令，MS Speech SDK 11的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！