Genesys Multicloud CX on-premises (formerly Engage on-premises)

Discussion Thread View
Expand all | Collapse all

VoiceBot Development With IBM TTS/ASR via WebSocket APIs

  • 1.  VoiceBot Development With IBM TTS/ASR via WebSocket APIs

    Posted 07-24-2021 12:11
    Dears,

    Do anybody know the best practice to develop a voicebot that integrates with IBM TTS/ASR knowing that in normal TTS/ASR integrations the GVP is involved to integrate via MRCPv1/MRCPv2 from MCP but here the challenge is that the TTS/ASR engines are providing integration methods via WebSocket which I can't find any clue for this in Genesys docs.

    I'm asking if there is a way to do this from Composer/GIA how it could be achieved? If we decided to make a custom application that acts as a middleware between the MCP and IBM TTS/ASR there will be two challenges here which are how could we integrate with MCP to retrieve the RTP customer voice of the current call and send it to IBM ASR and how to send back the TTS RTP/Stream to MCP so it can play it back to the customer.

    Appreciate your help in advance.
    #GenesysEngageDev
    #Implementation
    #Integrations
    #Self-Service

    ------------------------------
    Mostafa Awad
    IST Integration Services And Technologies Co.
    ------------------------------


  • 2.  RE: VoiceBot Development With IBM TTS/ASR via WebSocket APIs

    Posted 07-25-2021 15:50
    Mostafa,
    you're right that the real challenge is to forward RTP stream to ASR engine. TTS is not so complicated because you could synthesize a text, trasforming it in a .wav and play back it. It is not very efficient but, in many circumstances, it could work. The very big problem is related to ASR because, if you save the speech utterance by the customer in a file to process later by ASR, you could get a definitely bad experience.
    The only alternative you could have is to develop a MRCP interface that could invoke ASR (and TTS) services through the interface you have (websocket). It's not an easy task but you can do it, even because there should be some free MRCP code available on internet.

    But, at the end, the most important thing: you're talking about IBM ASR/TTS. As far as I know, they should provide a MRCP interface and, maybe, also a gRPC interface (you need to upgrade MCP to v9 to use that). We should start a trial with them very soon but they told me they already integrated GVP through a MRCP interface. I would suggest to double check this with them, because it could be the easier solution.

    Hope this helps you a bit.

    ------------------------------
    Giuliano Ferri
    Indra Italia spa
    ------------------------------



  • 3.  RE: VoiceBot Development With IBM TTS/ASR via WebSocket APIs

    Posted 07-26-2021 06:45
    Thanks Ferri for your reply.
    If you allow me to intervene here. I agree with you that developing MRCP interface is the hard way alternatively you could use UniMRCP open project which has a ready interface with IBM Speech Services. I believe Awad ask here if there is any other method directly from Composer/G-IA (but it seems there isn't any).

    As for the part you mentioned for IBM ASR/TTS that they provide MRCP/gRPC interfaces, As far as I know, gRPC is not an option at all in IBM case (correct me please).
    Also for MRCP, yes IBM has MRCP interface but they use it internally with their IBM VG to connect to IBM SST/TTS services. However external IVR parties (like GVP) can't use it directly as the standard integration should be through SIP connectivity to IBM Voice Gateway (which is not serving the business use case).

    ------------------------------
    Mostafa Rabie
    IST Integration Services And Technologies Co.
    ------------------------------



  • 4.  RE: VoiceBot Development With IBM TTS/ASR via WebSocket APIs

    Posted 07-26-2021 18:34
    Hello Rabie,
    I agree with you: UniMRCP is another option even if, personally, I don't know how they interface exactly IBM Speech Services (Awad was telling about websocket).
    As far as gRPC/MRCP is concerned: as I said in my post, it comes from some talk we're having about a trial we're setting up. In a few weeks we'll start the trial and I will be able to be more precise. By now, they told me they have a gRPC interface but they did not confirm me to have used it with Genesys. But, about MRCP, they said me that a customer in Germany already uses this integration from within GIA. And I don't think they were talking about SIP connectivity to IBM Voice Gateway because, with that customer, we are already using that kind of integration (and the trial is aimed at verify the feasibility of a different interface).
    So, I can't confirm you anything about gRPC but MRCP interface, according to IBM talks, should be available.

    Best regards

    ------------------------------
    Giuliano Ferri
    Indra Italia spa
    ------------------------------



  • 5.  RE: VoiceBot Development With IBM TTS/ASR via WebSocket APIs

    Posted 07-27-2021 08:36
    Hey Ferri,
    Thanks for the clarifications. I wish you all the best on your coming integrations trials with IBM Speech Services also please share with us whatever -shareable- results you have reached.

    As for the German customer that already did MRCP integration within GIA, Can we get more info about this? I believe having any contact from Genesys/IBM/Customer that has some knowledge about it would be more than great....

    Thanks again and have a good day

    ------------------------------
    Mostafa Rabie
    IST Integration Services And Technologies Co.
    ------------------------------



  • 6.  RE: VoiceBot Development With IBM TTS/ASR via WebSocket APIs

    Posted 07-27-2021 12:35
    Hello Rabie,
    I'll certainly share the result of our trial, as soon as we'll get something useful.
    As far as the German customer is concerned: I'm sorry but I don't know if they allowed the reference, so I'd suggest you to contact some Genesys sales account or your channel partner and ask him some more detail. I can just tell you that it is an international telco company... :-)
    Best regards

    ------------------------------
    Giuliano Ferri
    Indra Italia spa
    ------------------------------



  • 7.  RE: VoiceBot Development With IBM TTS/ASR via WebSocket APIs

    Posted 07-28-2021 11:12
    Thanks, Ferri ...this is really helpful.

    According to Genesys SC and the regional architect, the only way to integrate the existing Engage premise with IBM Speech Services is through MRCP interface in the middle (either to develop something on your own as you rightly mentioned or to bring 3rd party like Unimrcp). so it seems that the German customer has taken the way of developing his own MRCP interface.

    It is worth adding here that Genesys is working on having direct integration with MS & google Speech Services (not IBM) but even this would happen in the coming multi-cloud Engage (but not the Engage premise one).



    ------------------------------
    Mostafa Rabie
    IST Integration Services And Technologies Co.
    ------------------------------



  • 8.  RE: VoiceBot Development With IBM TTS/ASR via WebSocket APIs

    Posted 07-29-2021 05:07
    Hello Rabie,
    very interesting because they told me that IBM exposes directly a MRCP interface but, according to your description, this seems not exactly true because an external MRCP interface seems to be required.
    I will have to investigate better on this point because it is very important for our trial.
    Best regards

    ------------------------------
    Giuliano Ferri
    Indra Italia spa
    ------------------------------