Speech recognition online (similar to google voice search)

I am been trying to find a mehthod for online speech recognition, for eg: very similar to google voice search, which does not require the user to install any plugin/software/flash. The user just has to plugin the microphone and speak something for the text to get recognised.

I thought of this approach but don't know if this is corrector not. I built a dll which can take an input audio stream and give an output of recognized txt out of audio. I referenced this dll in ASP.NET references, and further thinking to upload an audio file from the user side to the server which will then be used the 'recognizer' dll. I am not sure if this approach is correct or not? Is there any other approach that I can follow?
The main thing is I can't have the user install anything or any dependency for this implementation such as flash/silverlight etc.


If you can specify that your users use Chrome 11, or later, you could use Google's webkit to speech enable your application. Here is a link on how to use webkit for speech. This leverages the audio input capabilities that are available in HTML5. If you take a look at this blog it will explain how it works, because the author reverse engineered it. It is taking the audio input from the user in the browser and sending it to a service for processing, returning the results as a JSON message. You could build your own service on the server side, as you are suggesting, to imitate what Google is doing. Building a scalable service for speech recognition will not be a small feat.

链接地址: http://www.djcxy.com/p/34398.html

上一篇: 未经指导的语音转换为文本

下一篇: 在线语音识别(类似于谷歌语音搜索)