Composite Recording

The On-premise Recording SDK supports two recording modes:

• Individual recording mode: This is the default recording mode. The SDK generates one audio and/or video file for each UID.
• Composite recording mode: Generates a single mixed audio and video file for all UIDs in a channel, or mixes the audio of all UIDs in the channel into an audio file and the video of all UIDs into a video file.

This page shows how to record audio and/or video in composite recording mode by the command line.

Before proceeding, ensure that you have compiled the Agora Recorder Demo and know how to record a call by the command line. For more information, see Record by Command Line. You must select either individual recording mode or composite recording mode when you start the recording. You cannot switch between the two modes after a recording starts.

See Differences between individual recording mode and composite recording mode to decide which mode you should use.

For simplicity, we assume that every UID in the channel sends both audio and video streams. If a UID neither sends audio nor video, such as the audience in interactive streaming, there will be no file for this UID.

When starting recording, set the isMixingEnabled parameter as 1 to enable composite recording mode.

The recording files and formats depend on your choice of recording audio, video, or both.

	Setting	Recording files and formats
Record only audio	--isAudioOnly 1 --isMixingEnabled 1	One audio file in AAC format
Record only video (no audio)	--isVideoOnly 1 --isMixingEnabled 1	One video file in MP4 format
Record both audio and video	--isMixingEnabled 1 --mixedVideoAudio 2	One file in MP4 format with both audio and video

• In composite recording mode, if you choose to record both audio and video, we suggest setting the mixedVideoAudio parameter as 2 to mix the audio and video in an MP4 file in real time.
• In composite recording mode, if you choose to record only audio/video, do not set the mixedVideoAudio parameter as 2.

In composite recording mode, you can set the audio and video profile of the recording file.

You can set the audioProfile parameter as the following values:

• 0: (Default) Sample rate of 48 KHz, communication encoding, mono, and a bitrate of up to 48 Kbps.
• 1: Sample rate of 48 KHz, music encoding, mono, and a bitrate of up to 128 Kbps.
• 2: Sample rate of 48 KHz, music encoding, stereo, and a bitrate of up to 192 Kbps.

You can set the mixResolution parameter, including the width, height, frame rate, and bitrate. The default setting is 360,640,15,500. We recommend you set the video profile according to the Video Profile Table.

In composite recording mode, you can use the layoutMode parameter to set the video layout and arrange the size and position of the region for each user. The default value is 0 (Floating Layout). For more information, see Set Video Layout.

The following example shows how to record both audio and video in an interactive live streaming channel in composite recording mode.

_1

./recorder_local --appId <Your App ID> --channel <The name of the channel to be recorded> --channelProfile 1 --uid 0 --appliteDir ~/Agora_Recording_SDK_for_Linux_FULL/bin --isMixingEnabled 1 --mixedVideoAudio 2 --audioProfile 1 --mixResolution 640,480,15,1000 --layoutMode 1

If a user stops sending streams or leave the channel during a recording session, the recording service fills the region of that user with the color of the canvas.