Server-side API connection

Overall connection flow

The overall connection flow for the digital human streaming media service is shown in the following figure. This flow applies to both 3D and 2D digital human streaming media services.

The following sections use Java as an example to demonstrate how to connect using the server-side SDK. For information about other languages, see Digitized Virtual Human Open Platform SDK. For information about connecting using the RTC stream pulling SDK, see Real-time streaming media service connection.

1. Import the API SDK

• For Java applications, you can add the following dependency to your pom.xml file to obtain the Java SDK for the Digitized Virtual Human Open Platform.

<dependency>
  <groupId>com.aliyun</groupId>
  <artifactId>avatar20220130</artifactId>
  <version>${latest_version}</version>
</dependency>

2. Initialize the SDK

The upgraded SDK includes comprehensive logic for handling parameters, return values, and requests. This lets you generate a separate client for each cloud product. Each client contains all the OpenAPI request methods for that product. This design isolates identity and authentication information between cloud products to improve security. The following example shows how to initialize the SDK for the Digitized Virtual Human Open Platform:

public static void main(String[] args) {
    Config config = new com.aliyun.teaopenapi.models.Config();
    config.setAccessKeyId("<accessKeyId>");
    config.setAccessKeySecret("<accessKeySecret>");
    config.setEndpoint("avatar.cn-zhangjiakou.aliyuncs.com");
    Client client = new Client(config);
}

3. Call the OpenAPI

Start the digital human streaming media service

• The following example starts a default digital human streaming media service. This service ingests the rendered video stream to an Alibaba Cloud RTC server. You can then pull the stream using the Alibaba Cloud RTC stream pulling SDK.

public static void main(String[] args) {    
    StartInstanceRequest startInstanceRequest = new StartInstanceRequest();
    // Obtain the tenant ID from the business configuration details on the Digitized Virtual Human Open Platform.
    startInstanceRequest.setTenantId(<tenantId>);
    // Obtain the business AppId from the business configuration details on the Digitized Virtual Human Open Platform.
    startInstanceRequest.setApp(new StartInstanceRequest.StartInstanceRequestApp().setAppId("<appId>"));
    startInstanceRequest.setUser(new StartInstanceRequest.StartInstanceRequestUser().setUserId("<userId>").setUserName("<userName>"));   
    StartInstanceResponse startInstanceResponse = client.startInstance(startInstanceRequest);

    // Get the SessionId after startup. This ID is required for subsequent text-to-speech API calls.
    String sessionId = startInstanceResponse.getBody().getData().getSessionId();
}

• The following example starts a digital human streaming media service that uses RTMP for stream ingest. This service ingests the rendered video stream to a specified RTMP server. You can then pull the stream using the server's protocol. For more information, see 3D stream ingest digital human connection guide.

public static void main(String[] args) {    
    StartInstanceRequest startInstanceRequest = new StartInstanceRequest();
    // Obtain the tenant ID from the business configuration details on the Digitized Virtual Human Open Platform.
    startInstanceRequest.setTenantId(<tenantId>);
    // Obtain the business AppId from the business configuration details on the Digitized Virtual Human Open Platform.
    startInstanceRequest.setApp(new StartInstanceRequest.StartInstanceRequestApp().setAppId("<appId>"));
    startInstanceRequest.setUser(new StartInstanceRequest.StartInstanceRequestUser().setUserId("<userId>").setUserName("<userName>"));
    // Custom RTMP
    String chanelType = "RTMP";
    Map<String, String> reqConfig = new HashMap<>();
    reqConfig.put("url", "<RTMP_URL>");
    reqConfig.put("token", "<RTMP_authentication_token>");
    startInstanceRequest.setChannel(new StartInstanceRequest.StartInstanceRequestChannel().setType(chanelType).setReqConfig(reqConfig));     
    StartInstanceResponse startInstanceResponse = client.startInstance(startInstanceRequest);

    // Get the SessionId after startup. This ID is required for subsequent text-to-speech API calls.
    String sessionId = startInstanceResponse.getBody().getData().getSessionId();
}   

Make the digital human speak text

• After you start the digital human streaming media service, you can make it speak text.

public static void main(String[] args) {    
    SendMessageRequest sendMessageRequest = new SendMessageRequest();
    // Obtain the tenant ID from the business configuration details on the Digitized Virtual Human Open Platform.
    sendMessageRequest.setTenantId(<tenantId>);
    // Pass the SessionId (obtained from the StartInstance return value).
    sendMessageRequest.setSessionId(<sessionId>);
    // Pass the text content to be spoken.
    sendMessageRequest.setTextRequest(new SendMessageRequest.SendMessageRequestTextRequest().setCommandType("START").setSpeechText("Hello, I am the smart digital human Xiaomeng. It's a pleasure to meet you all.")
                .setId(UUID.randomUUID().toString()));
    SendMessageResponse sendMessageResponse = client.sendMessage(sendMessageRequest);
}

Stop the digital human streaming media service

• You can stop the digital human streaming media service.

public static void main(String[] args) {    
    StopInstanceRequest stopInstanceRequest = new StopInstanceRequest();
    // Obtain the tenant ID from the business configuration details on the Digitized Virtual Human Open Platform.
    stopInstanceRequest.setTenantId(<tenantId>);
    // Pass the SessionId (obtained from the StartInstance return value).
    stopInstanceRequest.setSessionId(<sessionId>);
    StopInstanceResponse stopInstanceResponse = client.stopInstance(stopInstanceRequest);
}

For more detailed information about how to use the preceding APIs, see the API Reference.