Voximplant. Blog

Audio Recording

Voximplant cloud scenario allows to record audio, video and transcribe voice-to-text for each individual call object. Recorded audio, video or text is stored at Voximplant cloud for 3 months, after that it’s automatically deleted. Recordings can be downloaded if longer storage is required.

In order to start a recording, call a record method of the call object. Depending on parameters Voximplant will start audio, video or text recording. After the recording is started, the call object will receive a CallEvents.RecordStarted event that contains an URL to the file being recorded (that file is constantly updated during a recording). After call is disconnected, the call object will receive a CallEvents.RecordStopped event, which is normally used to pass recording url to backend. Recorded calls are displayed in the “calls” control panel section and can be played from web interface for debug purpose.

If no parameters are specified for a record method call, mono 8kHz 32kbit/s audio in mp3 format will be recorded. Incoming and outgoing audio streams of a call object will be mixed together. If stereo parameter is set to true, both streams will be recorded into separate channels. Audio stream from a call endpoint to Voximplant cloud is recorded into the left channel. Audio stream from Voximplant cloud to a call endpoint is recorded into the right channel.

If hd_audio parameter is set to true, audio is recorded with 48kHz 192kbit/s. See HD Audio Recording for details.

For audio recording only, a Recorder module provides additional functionality: stopping a recording, mixing multiple audio tracks into one and creating secure recordings that require HTTP auth for access.