This document describes how to publish (relay) audio/video streams in TRTC to CDNs so that viewers can watch the streams using standard live streaming players.
Publishing the Local User’s Stream to CDNs
Prerequisites
1. Go to activate the Tencent Cloud CSS service (please note that your account has been bound to the card and and has no outstanding charges, otherwise, the CSS service cannot be activated). Configuring a playback domain is mandatory for CSS. For detailed instructions, please refer to Adding Your Own Domain.
2. Select Domain Management in the left navigation bar, and you will see a new push domain name added to your domain list, formatted as xxxxx.livepush.myqcloud.com, where xxxxx represents a number called bizid.
3. Click Add Domain, enter the playback domain name you have registered, select the domain name type as Playback Domain, select the acceleration region, and click Confirm.
4. After the domain name is added successfully, the system will automatically assign you a CNAME domain name (ending with .liveplay.myqcloud.com). The CNAME domain name cannot be accessed directly. You need to complete the CNAME configuration with your domain name service provider. After the configuration takes effect, you can enjoy the CSS service. For detailed instructions, please refer to CNAME Configuration.
Note：
You do not need to add a push domain name. After enabling the relayed live streaming feature in Step 1, Tencent Cloud will add a push domain name formatted as xxxxx.livepush.myqcloud.com to your CSS console by default. This domain name serves as a default push domain name agreed upon between CSS service and TRTC service.
Description
You can use the startPublishMediaStream API of TRTCCloud to publish the audio/video streams of local users to live streaming CDNs (known in TRTC as "relay to CDN").
The TRTC server will send the audio/video data directly to the CDN server. Because the data is not transcoded, the cost is relatively low.
However, if there are multiple users publishing audio/video streams in a room, there will be a CDN stream for each user. Multiple players are needed to play the streams, and they may not play in sync.
Directions
Follow the steps below to publish the local user’s stream to CDNs.
1. Create a TRTCPublishTarget object and set mode in the object to TRTCPublishBigStreamToCdn or TRTCPublishSubStreamToCdn. The former is used to publish the user’s primary stream (usually the camera), and the latter is used to publish the user’s substream (usually the screen).
2. Set cdnUrlList in the TRTCPublishTarget object to one or multiple CDN addresses (which usually starts with rtmp://). If you publish to the Tencent Cloud CDN, set isInternalLine to true; otherwise, set it to false.
3. Because the data is not transcoded, leave TRTCStreamEncoderParam and TRTCStreamMixingConfig empty.
4. Call startPublishMediaStream. If the taskId parameter returned by the onStartPublishMediaStream callback is not empty, the local API call is successful.
5. To stop publishing, call stopPublishMediaStream, passing in the taskId returned by onStartPublishMediaStream.
Sample code
The code below publishes the local user’s stream to a live streaming CDN.
You can call startPublishMediaStream to mix the streams of multiple users in a TRTC room into one stream and publish the stream to a CDN. The TRTCStreamEncoderParam and TRTCStreamMixingConfig parameters of the API allow you to determine the details of stream mixing and transcoding.
The streams will be decoded on the cloud first, mixed, and then re-encoded according to the stream mixing parameters (TRTCStreamMixingConfig) and transcoding parameters (TRTCStreamEncoderParam) you specify. Afterward, they will be published to CDNs. In this mode, additional transcoding fees are charged.
Directions
Follow the steps below to mix the streams of multiple users in a room and publish the mixed stream to CDNs.
1. Create a TRTCPublishTarget object and set mode in the object to TRTCPublishMixStreamToCdn.
2. Set cdnUrlList in the TRTCPublishTarget object to one or multiple CDN addresses (which usually start with rtmp://). If you publish to the Tencent Cloud CDN, set isInternalLine to true; otherwise, set it to false.
3. Set the encoding parameters (TRTCStreamEncoderParam):
Video encoding parameters: Specify the resolution, frame rate (15 fps is recommended), bitrate, and GOP (3 seconds is recommended). Bitrate and resolution work in correlation with each other. The table below lists some recommended resolution and bitrate settings.
Audio encoding parameters: Specify the codec, bitrate, sample rate, and sound channels according to the AudioQuality value you pass in when calling startLocalAudio.
videoEncodedWidth
videoEncodedHeight
videoEncodedFPS
videoEncodedGOP
videoEncodedKbps
640
360
15
3
800 Kbps
960
540
15
3
1200 Kbps
1280
720
15
3
1500 Kbps
1920
1080
15
3
2500 Kbps
TRTCAudioQuality
audioEncodedSampleRate
audioEncodedChannelNum
audioEncodedKbps
audioEncodedCodecType
TRTCAudioQualitySpeech
48000
1
50
0
TRTCAudioQualityDefault
48000
1
50
0
TRTCAudioQualityMusic
48000
2
60
2
4. Set the parameters for audio mixing and video layout (TRTCStreamMixingConfig):
Audio mixing parameters (audioMixUserList): You can leave this parameter empty to mix all audios in a room, or you can set it to the IDs of users whose audios you want to mix.
Video layout parameters (videoLayoutList): Video layout is determined by an array. Each TRTCVideoLayout element in the array determines the position, dimensions, and background color of a video window. If you specify fixedVideoUser, the window defined by the TRTCVideoLayout element will display the video of a specific user. If you set fixedVideoUser to null, the TRTC server will determine whose video to display in the window.
Example 1: Mix four users’ streams and use an image as the background.
layout1 specifies the position (upper half of the canvas) and dimensions (640 x 480) of the camera video of user jerry.
Because no user IDs are specified for layout2, layout3, and layout4, TRTC will display the videos of the other three users in the windows based on its own rule.
Example 2: Mix the camera video and screen of one user plus the camera videos of three other users.
layout1 specifies the position (left) and dimensions (1280 x 720) of user jerry’s screen. The rendering mode used is aspect fit (Fit), and the background color is black.
layout2 specifies the position (top right) and dimensions (300 x 200) of user jerry’s camera video. The rendering mode used is aspect fill (Fill).
Because no user IDs are specified for layout3, layout4, and layout5, TRTC will display the videos of the other three users in the windows based on its own rule.
5. Call startPublishMediaStream. If the taskId parameter returned by the onStartPublishMediaStream callback is not empty, the local API call is successful.
6. To change the stream mixing parameters (for example, the video layout), call updatePublishMediaStream, passing in the taskId returned in step 6 as well as the new TRTCStreamMixingConfig parameters. We recommend you do not change TRTCStreamEncoderParam during relay because doing so will affect the stability of CDN playback.
7. To stop publishing, call stopPublishMediaStream, passing in the taskId returned by onStartPublishMediaStream.
Sample code
The code below mixes the streams of multiple users in a room and publishes the result to a CDN.
1. Can I listen for the status of CDN streams? What should I do if an error occurs?
You can listen for the onCdnStreamStateChanged callback event to get the latest status of a relay to CDN task.
2. How do I switch from publishing a single stream to publishing mixed streams? Do I need to stop publishing first and create a new relay task?
To switch from publishing a single stream to publishing mixed streams, just call updatePublishMediaStream, passing in the taskid of the current task. Note that, in order to ensure the stability of the publishing process, you cannot switch from publishing a single stream to mixing and publishing only audios or only videos. By default, both audio and video data are published when you publish a single stream. If you switch to publishing mixed streams, you must also publish both audios and videos.
3. How can I mix only videos (without audio)?
Do not set the audio parameters in TRTCStreamEncodeParam and leave audioMixUserList of TRTCStreamMixingConfig empty.
4. Can I add watermarks to mixed streams?
Yes, you can use watermarkList of TRTCStreamMixingConfig to set watermarks.
5. In online learning scenarios, can I mix the screen shared by the teacher?
Yes, you can. We recommend you publish the screen as the substream and mix the teacher’s camera video and screen. When specifying the stream mixing parameters, set fixedVideoStreamType of TRTCVideoLayout to TRTCVideoStreamTypeSub.
6. When a preset layout is used, how are audios mixed?
When a preset layout is used, TRTC will mix the audios of up to 16 users in the room.
Billing
Cost calculation
Relay to CDN fees are charged based on the peak bandwidth used each month. If you mix streams, the MCU cluster will decode and re-encode the streams in the cloud, so an additional transcoding fee will be charged. Transcoding fees vary with the resolutions of the streams transcoded and the transcoding duration. For details, see Billing of MixTranscoding and Relay to CDN.
Cost control
If you use client-side APIs, stream mixing stops at the backend when either of the following conditions is met.
The anchor who called startPublishMediaStream to mix streams left the room.
The anchor called stopPublishMediaStream to stop mixing the streams.
In all other cases, TRTC will continue to mix streams in the cloud. Therefore, to reduce costs, when you no longer want to mix streams, please stop it using either of the above methods.
Yes
No
Was this page helpful?