Hello everyone,
By default, PJSIP appears to enable a jitter buffer with relatively low values, which may not be sufficient in environments where network quality is unstable.
In this context, we are considering strengthening the jitter buffer.
My questions are as follows:
Is the PJSIP jitter buffer actually enabled by default, or does it need to be explicitly configured?
What are the best practices for effectively increasing the jitter buffer (recommended parameters, limits, impacts)?
Is it better to manage this at the level of:
the PJSIP endpoints (pjsip.conf)?
or via the dialplan (extensions_custom.conf) as planned?
Are there any risks or side effects (latency, audio degradation, etc.) to significantly increasing these values?
Thank you for your feedback.
If I understand correctly:
PJSIP does not enable a jitter buffer by default.
And when one is configured, the behavior remains the same whether it is defined on the endpoint side (pjsip.conf) or via the dialplan
Therefore, I would need your insights on best practices for implementation:
In what cases do you actually recommend enabling a jitter buffer on the Asterisk side (PJSIP) rather than leaving the management on the terminal side (phones/IP clients)?
Even if the result is similar, is there an advantage to favoring an endpoint-side configuration (jitter_buffer, jbmaxsize, etc.) rather than via the dialplan (JITTERBUFFER()), particularly in terms of maintainability or performance?
I don’t have any recommendations for enabling it on the Asterisk side. It’s rare for it to be needed.
Those don’t exist in PJSIP. You’re using JITTERBUFFER via set_var on an endpoint, or in the dialplan. There is no performance difference. Maintainability is up to you, but if you’re using FreePBX then you have to work within the confines of what FreePBX allows.