Hello! I'd like to create little system such that: 1. user from mobile phone calls server 2. server recognizes in real time what user says 3. server answers to the user on mobile phone with mp3 files At this moment my idea is to: 1. have Skype on mobile phone 2. call Skype on server 3. redirect speech from/to Sphinx4 ASR (automatic speech recognition) with the use of C# server application 4. recognize it on Sphinx4 (Java) However, I think it can be done in easier way, i.e. with IVR (interactive voice response). Some of IVRs are based on DTMF (I don't need DTMF) and other on ASR (which I need). And I've got two questions. Are there any freeware, opensource IVR software systems? How can I benefit from using Digium cards or/and Asterisk software? Greetings!
Telephone cards and Asterix PABX software is a combination of hardware and software to transform your CSTN to PSTN inside an Intranet. Several years ago I used Dialogic cards for an IVR application. I would like to give you some guideline. 1. Install a PCI telephone card that comes with drivers for Microsoft TAPI (Telephony Application Programming Interface) The one I used had 4 RT15 ports. There can be cards that support GSM now. Whatever option, make sure the vendor says that it supports TAPI. 2. Write a simple TAPI application to handle the call and record the sound. 3. Use speech recognition techniques to understand. 4. You can play a mp3 file to a call using TAPI according to my memory. If not, you can simply decode mp3 to PCM on the fly and play to call. IVR are based on DTMF. The system plays a menu and user response by DTMF tones. Since your application is based on speech recognition, I'm sure you'll have to handle most of the coding parts.
Thank you very much for your answer! What about such a way that I would have only server with access to internet, withouth any Digium cards, VoIP gates or any other things of this kind? Just computer with relatively fast connection to internet. For testing I would use softphone X-Lite. So I would have: user calling special number from ordinary mobile phone (or do I need special SIP mobile phone?) -> server receiving the call with Asterisk -> ScribbleJ plugin (http://scribblej.com/svn/ , or is it better to use JAsterisk instead of ScribbleJ?) to Asterisk sending the speech to Sphinx4 -> Sphinx4 recognizing the speech. And my question is as follows: What kind of service do I need to have? I guess I need to buy access to some kind of service from SIP provider. Is that right? Are services of SIP providers dependent on where I live (Poland)? What are the companies which I should consider? What should I ask them about? Greetings!