Jigasi. Sound notifications are not working

Developers
#1

Hello. I’m trying to use SoundNotificationManager.java to play notifications in TranscriptionGatewaySession (just to find out how it works).

SoundNotificationManager.injectSoundFile(jvbCall(not null 100%), SoundNotificationManager.PARTICIPANT_JOINED);

During debugging i’ve traced that file was loaded correctly and all packets were sent to MediaStream. But i can’t hear any sounds (Transcriber is joined to the conference)

Can you help me with that? Can i send some sounds in the conference somehow?

#2

Those notifications are working when translator mode of jigasi is turned on.

#3

Hello @damencho. Thanks for the information,
But i still have no sounds.

I’ve noticed that in TranscriptionGatewaySession this method returns false

    @Override
    public boolean isTranslatorSupported()
    {
        return true;
    }

That makes the makes SoundNotificationManage skip injecting sounds in any case

   public static void injectSoundFile(Call call, String fileName)
    {
        MediaStream stream = getMediaStream(call);

        // if there is no stream or the calling account is not using translator
        // or the current call is not using opus
        if (stream == null
            || !call.getProtocolProvider().getAccountID().getAccountPropertyBoolean(
                    ProtocolProviderFactory.USE_TRANSLATOR_IN_CONFERENCE, false)
            || stream.getDynamicRTPPayloadType(Constants.OPUS) == -1
            || fileName == null)
        {
            return;
        }

Is it possible to use SoundNotificationManager without sip configurations? (Because i need transcriber only)

#4

Why would the transcriber push sound notifications … it is using one way audio. This case is something we haven’t tested and used …

#5

I want to get transcriber text and play it in the conference (e.g via google text to speech api). Can you suggest something?

#6

Why would you do that :)? Someone has just said it and you want to play it, there is no point …
Mind that showing the translations is with corrections of previous messages, which you cannot do with audio, so you will need to implement a complicated mechanism of buffering the translations and fixing them before playing them …