[asterisk-bugs] [JIRA] (ASTERISK-25866) ChanSpy: allow usage of a long queue to store audio frames, to avoid audio loss

[Rusty Newton (JIRA)]

     [ https://issues.asterisk.org/jira/browse/ASTERISK-25866?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Rusty Newton updated ASTERISK-25866:
------------------------------------

    Status: Open  (was: Triage)

> ChanSpy: allow usage of a long queue to store audio frames, to avoid audio loss
> -------------------------------------------------------------------------------
>
>                 Key: ASTERISK-25866
>                 URL: https://issues.asterisk.org/jira/browse/ASTERISK-25866
>             Project: Asterisk
>          Issue Type: New Feature
>      Security Level: None
>          Components: Applications/app_chanspy
>    Affects Versions: 13.7.2
>            Reporter: Jean Aunis - Prescom
>            Assignee: Unassigned
>            Severity: Minor
>
> We noticed several times that ChanSpy is randomly losing audio frames, which are not transmitted to the spying party. This is not due to a network issue, as we could prove that Asterisk did not even send these lost frames to the spying peer.
> I figured out that these missing frames were actually flushed by the audiohook, because ChanSpy creates its audiohook with the flags AST_AUDIOHOOK_TRIGGER_SYNC and AST_AUDIOHOOK_SMALL_QUEUE.
> I suggest two modifications:
> # do not set the flag AST_AUDIOHOOK_TRIGGER_SYNC if ChanSpy is called with the option "o", i.e. if we just listen to the incoming audio
> # add an option "l" (for Long queue) to ChanSpy: if "l" is set, we do not set the flag AST_AUDIOHOOK_SMALL_QUEUE
> As far as I understand, the philosophy of ChanSpy is to spy a channel with a minimum of delay in the audio, at the cost of audio quality. That is why I suggest to add an option instead of just removing the flag.
> I made and tested this modification on my own, if you find it valuable I would push it to Gerrit for code review.



--
This message was sent by Atlassian JIRA
(v6.2#6252)