After you enable recording for a live stream, each audio/video frame published from the host’s mobile phone will be relayed to the recording system and written into a recording file.
When a live stream is interrupted, the access layer will immediately request the recording server to end the writing process, save the generated files to VOD, and generate an index for the file. You can then find the recording file in VOD. If you have configured the recording callback, the recording system will send the index ID and playback URL to the server you specified.
If a file is too large, errors may occur while the file is being transferred and processed in the cloud. As a result, to ensure the success of recording, we have capped the duration of a single recording file at 120 minutes. You can record shorter video segments by setting the
RecordInterval parameter to a smaller value.
The recording file is ready when you receive the recording callback ( about 5 minutes after a live stream ends). For more information, see Live Stream Callback.
Recording files are saved to VOD, so you need to activate VOD first. You can get recording files in the following ways:
To migrate your videos, you need to get the download addresses and download your videos.
CSS does not set a limit on the video storage period. You can manage your video files in the console or via RESTful APIs.
RecordIntervalparameter of the CreateLiveRecordTemplate API.
RecordIntervaland is not interrupted, only one recording file will be generated.
RecordInterval, the stream will be recorded into multiple files. This is to reduce the uncertainty of a file’s transfer time in a distributed system.
It depends on how you define a live stream. Assume that a host streamed for 20 minutes, but the stream was interrupted twice, once due to network change and once manually by the host. Does this count as one or three live streams?
In most mobile live streaming scenarios, we consider the period from when a host starts live streaming to when they end it as one live stream.
If you use the above standard, to determine which live stream a recording file belongs to, just search for recording notifications by live stream code and time. Each recording notification carries information including stream ID, start time, and end time.
You can use a cloud API to splice recording files.
This is usually because two recording tasks are initiated for the push domain. Check the following:
If you record the streams of a specific room, add the following parameter to the end of your push URL:
If you want both video and audio-only files, you can record videos first and then transcode the video files to audio files in VOD.
Set the storage period in your recording template to 0. For details, see Live Recording.
No, because the player is not able to identify the beginning and end of a video. Instead, you can use the following methods:
Tencent Cloud does not record your content without your request, so you cannot recover the content. This is the case with most cloud service providers.