search

qasim9872 / ibm-voice-gateway-docker-compose

Docker Compose Configurations for running IBM Voice Gateway.
0 stars 1 forks source link

Calling Watson Assistant via Linphone #1

Open Kaustubh0699 opened 2 years ago

Kaustubh0699 commented 2 years ago

Hi @qasim9872 , Thank you for great repository but can you elaborate some points:

  1. the EXTERNAL_IP is my system IP or docker's IP.
  2. How to setup or call using Linphone.
  3. How can I connect this via Twilio SIP trunk.
  4. What if I want to add my STT and TTS of some other vendor than IBM or Google.
qasim9872 commented 2 years ago

Hi @Kaustubh0699

Sure, happy to help. This repository is built as part of this article I wrote, it might be helpful to go through it first as it should answer some of your questions.

The EXTERNAL_IP should be the ip address on which the services are running. For example, if you are running this on your local machine, it will be the ip address of your machine. Alternatively, if you are running this on the cloud, it will be the public ip address of the server.

To configure Twilio sip trunk, you can follow this guide.

If you want to configure Ibm voice gateway to work with stt or tts from a different vendor, you can create a custom adaptor which will communicate with ibm voice gateway. You can find a sample in this repository

let me know if this helps? And if you have any more questions

Kaustubh0699 commented 2 years ago

Hi @qasim9872 , Thank you for responding back. As per the latest update I referred your medium blog and am able to setup IBM voice gateway using Docker. Now I am able to communicate with Watson Assistant using Linphone which is internally using IBM Watson STT and TTS services.

But the observation is IBM Watson STT/TTS does not work well with Arabic dialects and thus I want to replace Watson STT/TTS with custom speech engine. I have referred this repo but over here they are just referring to change the Watson STT Engine class to custom STT engine class. The documentation over here is not explanatory enough and I cannot find and relatable source over on Internet to achieve the same . Can you share some sample document or any reference which might help me to solve this. Thanks in advance for you help

qasim9872 commented 2 years ago

Hi @Kaustubh0699

Yes, it might be worth trying with Google STT as that will be the easiest. In the case of Google, IBM voice gateway provides a service that can be used as an adaptor for communicating with Google STT service. You can find more information about it here.

Alternatively, you will have to develop a custom service that will work similarly to the Google speech-to-text adaptor provided by IBM. In this case, the service will receive a WebSocket connection from the media-relay service for each call. This is where this repo will be much more useful as behind the scenes, the google STT adaptor is doing the same. The biggest difference is that instead of this file, it will have a file that will instead communicate with google.

Hope that helps. Let me know if you have any more questions