빠른 시작: 서버 쪽 오디오 스트리밍

Azure Communication Services 오디오 스트리밍 API를 통해 오디오 스트림 사용을 시작합니다. 이 빠른 시작에서는 자동화된 통화 라우팅 솔루션을 빌드하기 위해 통화 자동화 API에 이미 익숙하다고 가정합니다.

이 빠른 시작에서 설명하는 기능은 현재 공개 미리 보기로 제공됩니다.

var mediaStreamingOptions = new MediaStreamingOptions(
  new Uri("wss://YOUR_WEBSOCKET_URL"),
  MediaStreamingContent.Audio,
  MediaStreamingAudioChannel.Mixed,
  startMediaStreaming: true) {
  EnableBidirectional = true,
    AudioFormat = AudioFormat.Pcm24KMono
}
var options = new AnswerCallOptions(incomingCallContext, callbackUri) {
  MediaStreamingOptions = mediaStreamingOptions,
};

AnswerCallResult answerCallResult = await client.AnswerCallAsync(options);

var mediaStreamingOptions = new MediaStreamingOptions(
  new Uri("wss://<YOUR_WEBSOCKET_URL"),
  MediaStreamingContent.Audio,
  MediaStreamingAudioChannel.Mixed,
  startMediaStreaming: false) {
  EnableBidirectional = true,
    AudioFormat = AudioFormat.Pcm24KMono
}
var options = new AnswerCallOptions(incomingCallContext, callbackUri) {
  MediaStreamingOptions = mediaStreamingOptions,
};

AnswerCallResult answerCallResult = await client.AnswerCallAsync(options);

Start media streaming via API call
StartMediaStreamingOptions options = new StartMediaStreamingOptions() {
  OperationContext = "startMediaStreamingContext"
};

await callMedia.StartMediaStreamingAsync();

StopMediaStreamingOptions options = new StopMediaStreamingOptions() {
  OperationContext = "stopMediaStreamingContext"
};

await callMedia.StopMediaStreamingAsync();

private async Task StartReceivingFromAcsMediaWebSocket(Websocket websocket) {

  while (webSocket.State == WebSocketState.Open || webSocket.State == WebSocketState.Closed) {
    byte[] receiveBuffer = new byte[2048];
    WebSocketReceiveResult receiveResult = await webSocket.ReceiveAsync(
      new ArraySegment < byte > (receiveBuffer));

    if (receiveResult.MessageType != WebSocketMessageType.Close) {
      string data = Encoding.UTF8.GetString(receiveBuffer).TrimEnd('\0');
      var input = StreamingData.Parse(data);
      if (input is AudioData audioData) {
        // Add your code here to process the received audio chunk
      }
    }
  }
}

{
  "kind": "AudioMetadata",
  "audioMetadata": {
    "subscriptionId": "89e8cb59-b991-48b0-b154-1db84f16a077",
    "encoding": "PCM",
    "sampleRate": 16000,
    "channels": 1,
    "length": 640
  }
}

{
  "kind": "AudioData",
  "audioData": {
    "timestamp": "2024-11-15T19:16:12.925Z",
    "participantRawID": "8:acs:3d20e1de-0f28-41c5…",
    "data": "5ADwAOMA6AD0A…",
    "silent": false
  }
}

var audioData = OutStreamingData.GetAudioDataForOutbound(audioData)),
byte[] jsonBytes = Encoding.UTF8.GetBytes(audioData);

// Write your logic to send the PCM audio chunk over the WebSocket
// Example of how to send audio data over the WebSocket
await m_webSocket.SendAsync(new ArraySegment < byte > (jsonBytes), WebSocketMessageType.Text, endOfMessage: true, CancellationToken.None);

var stopData = OutStreamingData.GetStopAudioForOutbound();
byte[] jsonBytes = Encoding.UTF8.GetBytes(stopData);

// Write your logic to send stop data to ACS over the WebSocket
// Example of how to send stop data over the WebSocket
await m_webSocket.SendAsync(new ArraySegment < byte > (jsonBytes), WebSocketMessageType.Text, endOfMessage: true, CancellationToken.None);

MediaStreamingOptions mediaStreamingOptions = new MediaStreamingOptions(appConfig.getTransportUrl(), MediaStreamingTransport.WEBSOCKET, MediaStreamingContent.AUDIO, MediaStreamingAudioChannel.MIXED, true).setEnableBidirectional(true).setAudioFormat(AudioFormat.PCM_24K_MONO);
options = new AnswerCallOptions(data.getString(INCOMING_CALL_CONTEXT), callbackUri).setCallIntelligenceOptions(callIntelligenceOptions).setMediaStreamingOptions(mediaStreamingOptions);
Response answerCallResponse = client.answerCallWithResponse(options, Context.NONE);

MediaStreamingOptions mediaStreamingOptions = new MediaStreamingOptions(appConfig.getTransportUrl(), MediaStreamingTransport.WEBSOCKET, MediaStreamingContent.AUDIO, MediaStreamingAudioChannel.MIXED, false)
    .setEnableBidirectional(true)
    .setAudioFormat(AudioFormat.PCM_24K_MONO);

options = new AnswerCallOptions(data.getString(INCOMING_CALL_CONTEXT), callbackUri)
    .setCallIntelligenceOptions(callIntelligenceOptions)
    .setMediaStreamingOptions(mediaStreamingOptions);

Response answerCallResponse = client.answerCallWithResponse(options, Context.NONE);

StartMediaStreamingOptions startMediaStreamingOptions = new StartMediaStreamingOptions()
    .setOperationContext("startMediaStreamingContext");

callConnection.getCallMedia().startMediaStreamingWithResponse(startMediaStreamingOptions, Context.NONE);     

StopMediaStreamingOptions stopMediaStreamingOptions = new StopMediaStreamingOptions()
    .setOperationContext("stopMediaStreamingContext");
callConnection.getCallMedia().stopMediaStreamingWithResponse(stopMediaStreamingOptions, Context.NONE);

@OnMessage
public void onMessage(String message, Session session) {
  System.out.println("Received message: " + message);
  var parsedData = StreamingData.parse(message);
  if (parsedData instanceof AudioData) {
    var audioData = (AudioData) parsedData;
    sendAudioData(session, audioData.getData());
  }
}

{
  "kind": "AudioMetadata",
  "audioMetadata": {
    "subscriptionId": "89e8cb59-b991-48b0-b154-1db84f16a077",
    "encoding": "PCM",
    "sampleRate": 16000,
    "channels": 1,
    "length": 640
  }
}

{
  "kind": "AudioData",
  "audioData": {
    "timestamp": "2024-11-15T19:16:12.925Z",
    "participantRawID": "8:acs:3d20e1de-0f28-41c5…",
    "data": "5ADwAOMA6AD0A…",
    "silent": false
  }
}

private void sendAudioData(Session session, byte[] binaryData) {
    System.out.println("Data buffer---> " + binaryData.getClass().getName());
    if (session.isOpen()) {
        try {
            var serializedData = OutStreamingData.getStreamingDataForOutbound(binaryData);
            session.getAsyncRemote().sendText(serializedData);
        } catch (IOException e) {
            e.printStackTrace();
        }
    }
}

private void stopAudio(Session session) {
    if (session.isOpen()) {
        try {
            var serializedData = OutStreamingData.getStopAudioForOutbound();
            session.getAsyncRemote().sendText(serializedData);
        } catch (IOException e) {
            e.printStackTrace();
        }
    }
}

var mediaStreamingOptions = new MediaStreamingOptions(
	new Uri("wss://YOUR_WEBSOCKET_URL"),
	MediaStreamingContent.Audio,
	MediaStreamingAudioChannel.Mixed,
	startMediaStreaming: true)
{
	EnableBidirectional = true,
		AudioFormat = AudioFormat.Pcm24KMono
}
var options = new AnswerCallOptions(incomingCallContext, callbackUri)
{
	MediaStreamingOptions = mediaStreamingOptions,
};

AnswerCallResult answerCallResult = await client.AnswerCallAsync(options);

const mediaStreamingOptions: MediaStreamingOptions = {
	transportUrl: transportUrl,
	transportType: "websocket",
	contentType: "audio",
	audioChannelType: "unmixed",
	startMediaStreaming: false,
	enableBidirectional: true,
	audioFormat: "Pcm24KMono"
}
const answerCallOptions: AnswerCallOptions = {
	mediaStreamingOptions: mediaStreamingOptions
};

answerCallResult = await acsClient.answerCall(
	incomingCallContext,
	callbackUri,
	answerCallOptions
);

const startMediaStreamingOptions: StartMediaStreamingOptions = {
	operationContext: "startMediaStreaming"
}

await answerCallResult.callConnection.getCallMedia().startMediaStreaming(startMediaStreamingOptions);

const stopMediaStreamingOptions: StopMediaStreamingOptions = {
	operationContext: "stopMediaStreaming"
}
await answerCallResult.callConnection.getCallMedia().stopMediaStreaming(stopMediaStreamingOptions);

wss.on('connection', async (ws: WebSocket) => {
	console.log('Client connected');
	await initWebsocket(ws);
	await startConversation();
	ws.on('message', async (packetData: ArrayBuffer) => {
		try {
			if (ws.readyState === WebSocket.OPEN) {
				await processWebsocketMessageAsync(packetData);
			} else {
				console.warn(`ReadyState: ${ws.readyState}`);
			}
		} catch (error) {
			console.error('Error processing WebSocket message:', error);
		}
	});
	ws.on('close', () => {
		console.log('Client disconnected');
	});
});

async function processWebsocketMessageAsync(receivedBuffer: ArrayBuffer) {
	const result = StreamingData.parse(receivedBuffer);
	const kind = StreamingData.getStreamingKind();

	// Get the streaming data kind  
	if (kind === StreamingDataKind.AudioData) {
		const audioData = (result as AudioData);
		// process your audio data  
	}
}

{
  "kind": "AudioMetadata",
  "audioMetadata": {
    "subscriptionId": "89e8cb59-b991-48b0-b154-1db84f16a077",
    "encoding": "PCM",
    "sampleRate": 16000,
    "channels": 1,
    "length": 640
  }
}

{
  "kind": "AudioData",
  "audioData": {
    "timestamp": "2024-11-15T19:16:12.925Z",
    "participantRawID": "8:acs:3d20e1de-0f28-41c5…",
    "data": "5ADwAOMA6AD0A…",
    "silent": false
  }
}

async function receiveAudioForOutbound(data: string) { 
    try {
        const jsonData = OutStreamingData.getStreamingDataForOutbound(data);
        if (ws.readyState === WebSocket.OPEN) {
            ws.send(jsonData);
        } else {
            console.log("socket connection is not open.");
        }
    } catch (e) {
        console.log(e);
    }
}

async function stopAudio() {
	try {
		const jsonData = OutStreamingData.getStopAudioForOutbound();
		if (ws.readyState === WebSocket.OPEN) {
			ws.send(jsonData);
		} else {
			console.log("socket connection is not open.");
		}
	} catch (e) {
		console.log(e);
	}
}

media_streaming_configuration = MediaStreamingOptions(
    transport_url=TRANSPORT_URL,
    transport_type=MediaStreamingTransportType.WEBSOCKET,
    content_type=MediaStreamingContentType.AUDIO,
    audio_channel_type=MediaStreamingAudioChannelType.MIXED,
    start_media_streaming=True,
    enable_bidirectional=True,
    audio_format=AudioFormat.PCM24_K_MONO,
)
answer_call_result = call_automation_client.answer_call(
    incoming_call_context=incoming_call_context,
    media_streaming=media_streaming_configuration,
    callback_url=callback_uri,
)

media_streaming_configuration = MediaStreamingOptions(
    transport_url=TRANSPORT_URL,
    transport_type=MediaStreamingTransportType.WEBSOCKET,
    content_type=MediaStreamingContentType.AUDIO,
    audio_channel_type=MediaStreamingAudioChannelType.MIXED,
    start_media_streaming=False,
    enable_bidirectional=True,
    audio_format=AudioFormat.PCM24_K_MONO
)

answer_call_result = call_automation_client.answer_call(
    incoming_call_context=incoming_call_context,
    media_streaming=media_streaming_configuration,
    callback_url=callback_uri
)

call_automation_client.get_call_connection(call_connection_id).start_media_streaming(
    operation_context="startMediaStreamingContext"
)

call_automation_client.get_call_connection(call_connection_id).stop_media_streaming(operation_context="stopMediaStreamingContext") 

async def handle_client(websocket):
    print("Client connected")
    try:
        async for message in websocket:
            json_object = json.loads(message)
            kind = json_object["kind"]
            if kind == "AudioData":
                audio_data = json_object["audioData"]["data"]
                # process your audio data
    except websockets.exceptions.ConnectionClosedOK:
        print("Client disconnected")
    except websockets.exceptions.ConnectionClosedError as e:
        print(f"Connection closed with error: {e}")
    except Exception as e:
        print(f"Unexpected error: {e}")

{
  "kind": "AudioMetadata",
  "audioMetadata": {
    "subscriptionId": "89e8cb59-b991-48b0-b154-1db84f16a077",
    "encoding": "PCM",
    "sampleRate": 16000,
    "channels": 1,
    "length": 640
  }
}

{
  "kind": "AudioData",
  "audioData": {
    "timestamp": "2024-11-15T19:16:12.925Z",
    "participantRawID": "8:acs:3d20e1de-0f28-41c5…",
    "data": "5ADwAOMA6AD0A…",
    "silent": false
  }
}

async def send_data(websocket, buffer):
    if websocket.open:
        data = {
            "Kind": "AudioData",
            "AudioData": {
                "Data": buffer
            },
            "StopAudio": None
        }
        # Serialize the server streaming data
        serialized_data = json.dumps(data)
        print(f"Out Streaming Data ---> {serialized_data}")
        # Send the chunk over the WebSocket
        await websocket.send(serialized_data)

async def stop_audio(websocket):
    if websocket.open:
        data = {
            "Kind": "StopAudio",
            "AudioData": None,
            "StopAudio": {}
        }
        # Serialize the server streaming data
        serialized_data = json.dumps(data)
        print(f"Out Streaming Data ---> {serialized_data}")
        # Send the chunk over the WebSocket
        await websocket.send(serialized_data)

리소스 정리

Communication Services 구독을 정리하고 제거하려면 리소스 또는 리소스 그룹을 삭제하면 됩니다. 리소스 그룹을 삭제하면 해당 리소스 그룹에 연결된 다른 모든 리소스가 함께 삭제됩니다. 리소스 정리에 대해 자세히 알아보세요.

다음 단계

• 오디오 스트리밍에 대해 자세히 알아봅니다.
• 통화 자동화 및 해당 기능에 대해 자세히 알아봅니다.
• 작업 재생에 대해 자세히 알아봅니다.
• 인식 작업에 대해 자세히 알아봅니다.