Android 2.2.1 Nexus one：语音识别问题
大家好，
我在 Google Nexus 一部手机上的语音识别方面遇到了一个奇怪的问题 有固件：2.2.1.语音识别给出多种解读 当我对语音识别说“你好”时， 收到的结果是“你好，酒店，照片，fomdem，本田” 预计只会出现“你好” 同样的事情在固件 2.1 上运行良好，给出了令人满意的结果。
必须采取什么措施来避免这个问题。任何建议都是有帮助的
最好的问候， 维纳亚克
HI Folks,
I have a strange problem for Voice recognition on Google Nexus one phone which
have Firmware:2.2.1.Voice recognition gives multiple interpretations
of the spoken word When I speak "Hello" to the voice recognition,
the results received is "hello, hotels, photos, fomdem, honda"
which is expected to come only "hello"
The same things works fine on Firmware 2.1 which give satisfactory result.
Whats has to be done to avoid this issue.Any Suggestions are helpful
Best Regards,
Vinayak
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
我无法解释不同版本的不同行为，但是您是否看过 http://developer.android.com/reference/android/speech/RecognizerIntent.html#EXTRA_MAX_RESULTS？
该意图接受一个最大结果参数，该参数告诉识别器有多少候选字符串返回给客户端。通常在语音识别中，客户端可能需要向用户提供消歧步骤（例如“您说的是“你好”还是“酒店”？”）。如果您只想要最有可能的候选者，请将 EXTRA_MAX_RESULTS 设置为 1。
I can't explain the differnet behavior for different versions, but have you looked at http://developer.android.com/reference/android/speech/RecognizerIntent.html#EXTRA_MAX_RESULTS ?
The intent accepts a Max Results parameter which tells the recognizer how many candidate strings to return to the client. Typically in speech recognition, the client may need to provide the user a disambiguation step (like "did you say "hello" or "hotel"?". If you only want the most likely candidate, set EXTRA_MAX_RESULTS to 1.