Early media: definition and a practical example
Short answer: Early media is audio sent to the caller before the call is answered, usually after a SIP 183 Session Progress response, such as ringback tone or a carrier message.
In Portuguese: Early media
What is early media?
Early media is the audio the caller hears before the call is answered. In SIP, it usually starts with a 183 Session Progress response carrying an SDP: the network opens the RTP stream before the 200 OK and plays something to the caller. The model is described in RFC 3960.
The content can be ringback generated by the carrier, hold music, a notice such as “this call may be recorded” or an error message, such as “the number you have called does not exist” or “this phone is switched off or out of the coverage area”. To the human ear, all of it is sound. To the signaling, the call has not been answered yet.
That sets a trap for any automated test. If the test treats “I heard audio” as success, a carrier message saying the number is out of service passes as a good call. What defines the answer is the 200 OK, and what defines whether the right destination responded is the content of the audio.
In practice
A call to a number whose trunk is down might play out like this:
| Time | Message | What the caller hears |
|---|---|---|
| 0.0 s | INVITE sent | Nothing |
| 0.2 s | 100 Trying | Nothing |
| 1.1 s | 183 Session Progress with SDP | “The number you have called is not available” |
| 9.0 s | 480 Temporarily Unavailable | Call ends |
There was never a 200 OK. A test that only looks for audio would see 8 seconds of speech and mark the line as working. The final response, 480, tells the real story. Some carriers even answer with 200 OK just to play the message, and then only the content of the audio reveals the failure.
How Monitor uses this
For a greeting circuit, a recorded message in place of ringback is usually the carrier speaking for a number it could not reach, and Monitor counts that as a failure. See toll-free and IVR monitoring.
Frequently asked questions
What is the difference between 180 Ringing and 183 Session Progress?
180 Ringing only signals that the destination is ringing; the caller's side generates the ringback tone locally. 183 Session Progress usually carries an SDP and opens the audio before the answer, so the network can play its own tone, an announcement or an error message. Neither one means the call was answered.
Does hearing audio mean the call was answered?
No. With early media, the caller hears sound before any 200 OK. That sound can be ringback, music or a recording such as "the number you have called is not available". The call has only been answered when the 200 OK response to the INVITE arrives.