[asterisk-bugs] [JIRA] (ASTERISK-28303) Interaction between smoother and DTMF can cause out of order timestamps
Torrey Searle (JIRA)
noreply at issues.asterisk.org
Mon Feb 25 09:31:47 CST 2019
[ https://issues.asterisk.org/jira/browse/ASTERISK-28303?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Torrey Searle updated ASTERISK-28303:
-------------------------------------
Attachment: wrong_timestamp.pcap.pcapng
In the attached pcap, the rtp timestamp issue can be found at the following location
[RFC4733_DTMF_VALIDATOR:109]Bad RTP packet,Decreasing packet timestamp, previous packet had timestamp [1196633088]
[RFC4733_DTMF_VALIDATOR:110]Error detected for packet [1551107834.469102 RTP len=172 from=127.0.0.1:10120 v=2 p=0 x=0 cc=0 m=0 pt=8 (PCMA,1,8000) seq=18207 ts=1196249368 ssrc=0x5bbee653
> Interaction between smoother and DTMF can cause out of order timestamps
> -----------------------------------------------------------------------
>
> Key: ASTERISK-28303
> URL: https://issues.asterisk.org/jira/browse/ASTERISK-28303
> Project: Asterisk
> Issue Type: Bug
> Security Level: None
> Components: Resources/res_rtp_asterisk
> Affects Versions: 13.20.0
> Reporter: Torrey Searle
> Attachments: wrong_timestamp.pcap.pcapng
>
>
> When the smoother is non-empty, it infers the delivery time based on the number of samples removed from it.
> However while a dtmf (RFC4744) is being generated, no samples are being removed from the smoother, so it's delivery timestamp falls behind current time.
> When the next dtmf is being generated, this triggers a large jump forward in time in calc_txstamp as the dtmf functions use now() instead of delivery time. This causes a jump forward in rtp timestamps of audio packets w/o timing info (the ones generated by the smoother).
> However once the smoother becomes empty again, then timing will get passed once more from the calling leg to the called leg, and the above issue will manifest itself as a jump backwards in rtp timestamps
--
This message was sent by Atlassian JIRA
(v6.2#6252)
More information about the asterisk-bugs
mailing list