Status;
Opened

Date Started;
11/10/2004

Contributions;
Contribution has been decreased, due to resolution/workarround on our NGN platform
Bart Coppens (on behalf of CCA Belgium); 100USD

Contact;
Manager: Bart Coppens, (alias coppens_b), coppens.b@ccafrique.net or coppens_b@hotmail.com

Description;
Currently, Asterisk is using the timing of the input stream to reproduce the output stream. This means that when no RTP streams are being sent from the peer Endpoint/GW, Asterisk is unable not generate audio. This approach/limitation can lead to "one way speech" conditions: 1) Some devices don't generate audio until the answer supervision is received from the called. For all these scenarios, no ringback can be presented to the calling party. 2) In cases where the endpoints are using silence compression, the audio from asterisk is chopped.

Requirements;
To get this solved, Asterisk should get the clocking from an internal source in a way that an ouput stream can be generated without getting any RTP input. The clocking should than be taken from an internal timing mechanism that keeps track of the synchronization. The solution should not require E1/T1 connectivity (no TDM hardware).
It is the intension to solve the "no alerting scenario's (when peer is set in Recvonly mode) and all issues related to the use of silence compression. A configuration option should exist to choose the timing method.