Livestreams

Preparations

• If the preset virtual avatar image does not meet the requirements, customize a virtual avatar image in advance by referring to Image Modeling.
• Access the MetaStudio console from the Chrome browser.

Constraints

• The preset virtual avatar images on the MetaStudio console are valid for three years (till the the middle of 2027). The notice of bringing a preset image offline will be released on the live network three months in advance.
• See Constraints on virtual avatar livestreaming.

Procedure

1. Log in to the MetaStudio console.

2. Choose Workbench in the left navigation pane. Click Create under Livestreaming.

   The Virtual Avatar Livestreaming home page is displayed, as shown in Figure 1.

   Figure 1 Virtual Avatar Livestreaming home page
   

3. On the Livestreams tab, click the Create a live room area.

   The page displayed is shown in About the GUI. The operations are described in Operations of Livestreaming Materials.

   • In the material area on the left, there are preset video materials. You can customize a virtual avatar by importing backgrounds, stickers, and videos from the PC. You can also create a product and enter text.
   • Stickers, videos, and virtual avatar added to the preview area can be managed layer by layer, for example, moving up (), moving down (), pinning to top (), pinning to bottom (), applying globally (), and deleting (). You can also set the size and drag these materials to change their positions.

4. After a live video feed is created, select a script control method as needed.

   There are three script control methods:

   • Text control

     Text control enables virtual avatars to speak. You can add text in either of the following ways:

     • Enter text in the paragraph box of each scene. You can click to add paragraphs.
     • Click Import in the upper right corner to import a TXT file of at least 200 characters from the PC. If there are multiple paragraphs, you can use line feeds. The input text will be automatically parsed.

     Example of manually entering text:

     These examples are for reference only. A single scene must contain at least 200 characters, and a single paragraph must contain at most 2,000 characters.

     • Paragraph 1.1: Hello, everyone. Welcome to the meeting room. The topic of today is how to learn and work better.
     • Paragraph 1.2: All of us will inevitably encounter some problems when learning and working. What can we do to solve these problems?
       Figure 2 Text control
       

     Table 1 describes the operations of text control.

     Table 1 Operations

     Operation	Description
     Insert Pause	Insert a pause in the paragraph for breathing or reading pause.
     Multi-pronunciation Word	Select the text and set the correct pronunciations of multi-pronunciation words.
     Global Speaking Speed	Sets the global speaking speed, which ranges from 0.5x to 2x. When the speed is set to 0.5x, the voice is played at 0.5x speed. When the speed is set to 2x, the voice is played at 2x speed.
     Co-streamer	Select a piece of text for it to be read by the co-streamer.
     Global voice (streamer)	Global voice of the streamer, which is applicable to all scenes in the live room. If there is a specific voice for a scene, this voice will be prioritized. There is a default voice. You can also select other voices as needed. If you want to use a preset voice marked with Third-party, perform operations in Purchasing a DupDub Voice Package.
     Global voice (co-streamer)	Global voice of the co-streamer, which is applicable to all scenes in the live room. The co-streamer's voice cannot be set separately for a single scene. Figure 3 shows an ongoing livestream. If you want to interact, click to switch to the co-streamer's voice. Enter and send the interactive content. The streamer pauses the reading and the co-streamer reads the sent interactive content.
     Import	Text files can be imported from the PC and parsed to generate paragraphs. A file to be imported must: Be a TXT file of at least 200 characters. Be automatically parsed into paragraphs based on the line feed.
     Audition	Click Audition to listen to the current live room script. By default, the global voice of the streamer is used. If there is a specific voice for a scene, this voice will be prioritized. Co-streamer's content that is separately marked will be read by the co-streamer. You can also click in a specific paragraph area to listen to the content of that paragraph.
     Scene	You can perform the following operations on a scene: : changing the scene name, which defaults to Scene 1 : setting the prioritized voice of the current scene : copying a scene to add a same one : deleting a scene : adjusting the position of a scene : adding a blank paragraph to the current scene
     Paragraph	You can perform the following operations on a paragraph: : changing the paragraph name, which defaults to Paragraph 1.1 : listening to the reading of the current paragraph : copying a paragraph to add a same one : deleting a paragraph : adjusting the position of a paragraph in the scene : adding a blank paragraph

     Figure 3 Ongoing livestream
     
   • Audio control
     1. Click Upload Audio to upload a recorded audio file from the PC, as shown in Figure 4.

        WAV, M4A, and MP3 files are supported. The size of each audio file cannot exceed 100 MB. The total audio duration in each scene cannot be less than 1 minute.

        After a livestream starts, the origin voice in the audio is directly played. You can also set the voices of the streamer and co-streamer to read interactive content.

        Figure 4 Audio control
        
     2. After the audio file is uploaded, click Audition to listen to the audio file.
        After the livestream starts, you can click shown in Figure 5 to switch to the global voice of the streamer or the co-streamer's voice. The sent interactive content will be read with the voice you choose.

        Figure 5 Ongoing livestream
        

   • Improvising

     A livestream starts without script, as shown in Figure 6. After the livestream starts, the virtual streamer remains silent, and will not speak until interactive content is input or the livestream is taken over by a human.

     Figure 6 Improvising
     

5. If you need to create multiple scenes, click the + icon under Scene 1 on the right of the livestream preview window to add a scene. Then set the layout and script of the new scene based on the previous scene.
   Figure 7 Live room page
   

6. After setting the image layout and script of all scenes, click the Interaction Settings tab, as shown in Figure 8.

   There are four types of interaction settings: On-screen Comments, Joining Live Rooms, Likes, and Gifts. You can import the interaction rule library set in Interactions, or set interaction rules one by one in the current live room. A maximum of 100 interaction rules can be added for all types of interaction settings.

   The operation for adding interaction rules is similar for these four types. The following uses the operation for on-screen comments as an example.

   Figure 8 Interaction settings
   
   1. Click Add Rule in Figure 8. The page shown in Figure 9 is displayed.

      You can set the rule in the lower part of the page.

      Set the keywords of the on-screen comment triggering rule. If there are multiple keywords, separate them with vertical bars (|), for example, link|model|discount. When a keyword is identified in a comment, the preset reply of this keyword will be automatically read.

      Select Streamer or Co-streamer to reply to the user's question. The reply can be set in one the following 3 ways:

      • Text: Set the reply as text. A maximum of five replies can be added. After a livestream starts, the replies are matched in sequence based on the number of triggering times. Only one reply will be read each time.

        Example: For details about the link, model, and discount of the product, see the description in the upper part of the page.

      • Audio: Upload an audio file in MP4, WAV, or M4A format from the PC. When the rule is triggered, the audio file is used for reply.

      Images or videos can be displayed in the live feed to make a reply more informative. The method of adding and setting images is the same as that of adding and setting videos. The following uses the operation of images as an example.

      1. Click , select an image or video from the local PC, and add it to the + area.

         After the image is added, the thumbnail of the image is displayed in the area, as shown in .

      2. Move the pointer to the area. The area changes to .
      3. Click Preview. The image is added to the livestream preview window. You can drag the image or adjust the image size.
         Figure 9 On-screen comment settings
         
   2. Click Add to add on-screen comment reply rules by referring to 6.a.
   3. After all on-screen comment reply rules are added, you can click Add General Reply to add a general reply. The setting method is the same as that in 6.a.

      Only one general reply can be set. The general reply is triggered when a user question does not trigger any of the preset reply rules.

   4. You can perform the following operations on a rule:
      • Batch Manage: You can batch set the application scenarios of selected rules or delete selected rules.
      • Viewer Mention: After this function is enabled, you can see a text message "Replying to [name]: [content]" on the live feed.
      • Enable All: You can enable all rules.
      • : Set the application scenario of a single rule.
      • : Enable or disable the current rule.
      • : Delete a rule.

7. Click the Live Settings tab, as shown in Figure 10.
   See Table 2.

   Figure 10 My videos
   

   Table 2 Live settings

   Area	Parameter	Description
   Playback Startup	Playback Startup Mode (Either)	Select a livestream startup mode as needed and perform operations as prompted.
   Output	Livestreaming Image	Select the definition of the livestreaming image. Options: Smooth HD UHD Default value: HD
   	Playback Times	Set the number of times that a livestream should be played in a cyclical manner, or select Infinite loop. By default, the livestream is played once.
   	Live Window	Size of the live window. The window can be adjusted to vertical orientation. Options: Portrait Full screen (recommended): The captured frame is larger and needs to be rotated in the third-party live studio. The default value is Portrait.
   More	Event Callback Address	Fields of scene and paragraph events can be returned during livestreaming. This parameter is applicable to API calling or secondary development.
   	Live Room ID	You do not need to set it. It is automatically displayed after the current live room is saved.

8. After the preceding operations are complete, click in the upper right corner to save the streaming content.
9. Click Prepare Livestreaming in the lower right corner of the page.

   Automatic livestreaming after loading is selected by default. Click Prepare Livestreaming. The livestream starts automatically after the loading is complete.

   If you want to delay the livestream, deselect Automatic livestreaming after loading and click Prepare Livestreaming. To start the livestream, click Start Streaming at the same position.

   To start a livestream with improvising, you only need to click Prepare Livestreaming.

10. For text control and audio control, you need to confirm the risks. You are advised to complete the optimization as instructed before starting a livestream.
    1. The Warning dialog box is displayed, as shown in Figure 11.
       You need to check the warning items one by one to optimize the livestreaming content.

       Figure 11 Warning
       
    2. You can start livestreaming after the risks are eliminated, or select Skip to Stream.

11. After the livestream is started, the livestream startup progress is displayed in the livestream preview window.

    Figure 12 shows an ongoing livestream. Table 3 describes the operations in the lower right corner of the page.

    Figure 12 Ongoing livestream
    

    Table 3 About the GUI

    GUI Element	Description
    00:01:20	Livestreaming timing. For example, 00:01:20 indicates that the livestream has lasted one minute and 20 seconds.
    	Switches the voice of reading interactive content. The streamer's voice is used by default. You can switch to the co-streamer's voice.
    	Click this icon to select and send the preset interactive content for the streamer to read.
    Entering interactive content and pressing Enter to send the content	After the interactive content is entered and sent, the streamer or co-streamer reads the newly inserted content. After reading the newly inserted content, the streamer or co-streamer continues reading the original content.
    Real Person Takeover	Click Real Person Takeover. In this case, the virtual streamer stops voicing and only acts as programmed, and a real person voices for the virtual streamer. After the real person stops speaking, click Cancel Real Person Takeover to let the virtual streamer continue reading the original content.
    Pause	Click Pause to pause the livestream. Click Continue to resume the livestream.
    Stop	Click Stop to stop the livestream.