Morten Tryfoss

I've not looked into the source code of OpenCore AMR yet. Asterisk started the call. It's still Asterisk 13.8 and chan_sip. I've already debugged that `attr` is null in that...

I've looked a bit at pjsip's implementation of this. Check out [https://github.com/chebur/pjsip/blob/27dc7a969434c80c95a5c86fb3a36d07bce4df74/build/pjproject/src/pjmedia/include/pjmedia-codec/amr_helper.h](url) They have a predecode function which handles the SID frame specially on line 812: `} else if (in_info->frame_type...

I'll try to dig a little more. The setup is like this (all UDP): `Cisco SPA504 (any phone would probably do) Asterisk 13.8 Kamailio in front of the two MSC's.`...

One thing I notice in Wireshark. Every AMR packet coming from the MSC has CMR field set to 2, even though I'm now (for debugging purposes) forcing mode 2 that...

Just updating this ticket with things I find. I've debugged opencore decoder and I see that all SID frames are being decoded as SID_FIRST, instead of SID_FIRST and subsequent SID_UPDATE....

Here's the logic for decoding (mime_io.cpp): ``` case MRSID: if (quality) { if (temp & 0x80) { *frame_type = RX_SID_UPDATE; } else { *frame_type = RX_SID_FIRST; } } ```

The generation of comfort noise seems to stop way too early, and this is causing the clicking (generator stops between each SID). Earlier you mention that a packet should arrive...

Thanks for the reply. I'm only answering what I've been able to check right now. > Here in this repository, my module expects a frame every 20 ms. Consequently, VAD/DTX/CN...

Thanks for you reply! I've tried this patch, and now the frames seems to be decoded correctly. However, I'm still experiencing this clicking sound. I've been hacking a lot in...

I do not think this is related to the client. Here's a recording done at the G.722-side of Asterisk. Adjust the volume up a little bit. I know Asterisk do...