Announcing Dual-channel Call Recordings by Default

May 25, 2022
Written by

Programmable Voice Dual-channel by default header

We’re excited to announce that voice recordings are now recorded and stored at Twilio in dual-channel by default at no additional cost, starting today for new accounts and being rolled out for all accounts.

With this change, two-party audio files of call recordings will now have two channels and contain the audio from each call leg in a separate channel, enabling you to process the individual participants of the call recorded for much better speaker accuracy. Previously, you needed to specify dual-channel to record each participant in a separate channel, while all audio mixed in a single channel was the default behavior.

As part of this release, dual-channel is now the same price as mono-channel storage because we want all of our voice customers to benefit from dual-channel recordings.

The value of dual-channel recordings

Call recordings are often used to understand the customer journey and satisfaction, collect evidence and identify areas of improvement. For example, call centers often use conversation analysis for service quality assurance and agent training.

Organizations have traditionally relied on people listening to recordings to review agent quality scores or compliance which becomes expensive at scale and has driven organizations only to sample a very small number of calls. Listening to calls may also involve security and privacy issues. With advancements in artificial intelligence, it is now possible to automatically transcribe recordings, then apply machine learning algorithms for sentiment analysis, call scoring, or other means of analysis in a more scalable, consistent, and secure manner.

With this release, the media associated with the recording resource of a two-party call will have two channels and contain the audio from each call leg in a separate channel, automatically increasing the accuracy of transcriptions with each channel representing a different participant.

How to enable dual-channel voice recordings

Starting today, all of your two-party calls will be recorded and stored in dual-channel without any change required in your Twilio application. You can continue downloading files in mono or dual channel as desired.

It’s still possible to continue enabling single-party call recordings using the recordingTrack parameter to select whether a call’s inbound or outbound audio tracks should be recorded. In this case, the recording file is always mono.

Please note dual-channel is not supported in TwiML <conference> noun and <record> verb. The media associated with the recording resources requested from these methods, will be mixed in a single channel file as usual.

Download the dual-channel media file

Call recordings are now stored at Twilio in dual-channel by default. The media file associated with the recording resource of a two-party call has two channels and contains the audio from each call leg in a separate channel. A new RequestedChannels query parameter can be used to specify whether the file should be downmixed to a single-channel by Twilio before downloading or downloading a dual-channel file as-is.

  • For backward compatibility, when the RequestedChannels query parameter is not specified, the default download will be either mono or dual, depending on what was requested for the recording.
  • Voice recordings where dual-channel is not supported, like for the TwiML <conference> noun and <record> verb, store all audio in a single channel file, and Not Found is returned if a dual-channel recording is explicitly requested for download.

Example: download MP3 media in dual-channel


Unlock the power of voice

Twilio call recording helps you get the most out of your conversations, driving compliance, quality assurance, and a better understanding of your customers through conversation analytics, all while ensuring you have full control of your data.

Stay tuned for more to come, and we can’t wait to see what you build with Twilio!