Low Volume and "Network problems between PBX and SipTrunk.com"

[William H.]

Forgive me for posting a redundant question, but I'm having issues piecing the puzzle pieces together.
SETUP
* Country US
* 3CX Hosted Enterprise Version 20.5
* 13 users configured, let's say 30 SIP devices/apps/softphones behind the SBC
* a Few Yealink T57W all up to date,
* tech's using Softphone on Windows 11 kept updated.
* Tested on PWA client
* SBC: Debian local Virtual (8 GRAM, 2 proc, 50G HDD) with 3CX SBC same version as PBX.

ISSUE: Users have complained about mostly low call volume levels during a call for a period of time, and then all the time on some calls.

TROUBLESHOOTING
I found a configuration issue where G.729 was configured and SipTrunk.com using PCMU. I have since changed it to PCMU across the board on the 3CX.
my call monitors on all were showing:
* "Network problems between PBX and [number] SipTrunk.com"
* Mode Transcoding
* Codec PCMU - G729, PCMU - PCMU
I found the article about what causes transcoding, and I was using "audio through PBX," as I thought that would be easier to keep security tight. As I understand it, the SBC would route SIP and RTP through it to the PBX and vice-versa. I turned this option off on my extension and made a test call, but the same issue arose on the call monitor, although I had PCMU on both legs and, audio through PBX, off in my extension, it was still transcoding. The call monitor shows the same and no change when the volume issue occurs and does not.

The low volume occurs on all devices; phone, softphone, PWA. It is intermittent, on ext to ext, and external calls. All call monitors show issues with network between SipTrunk.com and 3cx hosted PBX. I have had a ticket open with SipTrunk.com since last week.

QUESTIONS
1. I wouldn't think that "network issues" would cause an issue with the volume level of a call as my understanding is this is set and "inside" the packet. However, I would like to see a clean report so that is not an issue anywhere else.
2. I wouldn't think the transcoding could cause this issue as the codec is decompressed, modified, and reassembled as the other codec. But is 3cx really transcoding from PCMU to PCMU, in other words, opening the packet and re-assembling it?
3. Where should I look next, and should I include support now for at least the Network issues between the PBX and SipTrunk.com, or both?

 

[ivank]

1. The reports are good, "problems" are from the fact that SIP trunk doesn't send RTCP reports or they don't get thru to PBX.
Even if there were real problems like packet loss, they wouldn't change the volume of the call.
2. Transcoding in general does not change the amplitude of the signal. But yes, the process goes as you described even if codecs match.
3. You mentioned ext 2 ext calls, try disabling everything that forces transcoding for all of them (call recording, SRTP, PBX Delivers Audio) and make more tests to see if there's improvement.

 

[William H.]

I tried a few things suggested and I thought the issue was gone. I am attempting to keep a capture running to catch any reported issues. I have a PC connected to a monitor port on the WAN side of the firewall. However, I was wondering if I could get some advice.

(environment: hosted 3cx v20, SBC VM; Yealink, windows softphone, cell app used; windows 11 pc running Wireshark/TShark; everything on current versions and patches)

I am trying to automate the start of captures at 08:00am when we open and before my shift. I keep having issues getting the capture filter correct for TShark. I also don't think I will get much because of devices using the tunnel to connect to the hosted 3cx server. Any suggestions are appreciated. Here is my current TShark command script for task manager to start.
cd "c:\program files\wireshark"
: Set interface name -i
: Set packet filter in libpcap filter syntax -f
: Set Stop duration in seconds -a
: set Capture output buffer duration in seconds -b
: -V add packet tree details
: TimeStamp
set TimeStamp=%date:~4,2%%date:~7,2%%date:~10,4%_%time:~0,2%%time:~3,2%%time:~6,2%
: set Capture File -w
tshark.exe (port 5060 or port 5090 or udp portrange 9000-10999) -P -i 9 -a duration:32400 -b duration:3600 -V -w C:\Digit3\WireShark\capture_%TimeStamp%.pcap

I also tried -f (port 5060 or port 5090 or udp portrange 9000-10999)
and variations of those using opening/ending quotes ".."
yes, all those constants are marked out as I was attempting to simplify but wanted a reminder of what they were.

 

[William H.]

1. The reports are good, "problems" are from the fact that SIP trunk doesn't send RTCP reports or they don't get thru to PBX.
Even if there were real problems like packet loss, they wouldn't change the volume of the call.
2. Transcoding in general does not change the amplitude of the signal. But yes, the process goes as you described even if codecs match.
3. You mentioned ext 2 ext calls, try disabling everything that forces transcoding for all of them (call recording, SRTP, PBX Delivers Audio) and make more tests to see if there's improvement.

I tried a few things suggested and I thought the issue was gone. I am attempting to keep a capture running to catch any reported issues. I have a PC connected to a monitor port on the WAN side of the firewall. However, I was wondering if I could get some advice.

(environment: hosted 3cx v20, SBC VM; Yealink, windows softphone, cell app used; windows 11 pc running Wireshark/TShark; everything on current versions and patches)

I am trying to automate the start of captures at 08:00am when we open and before my shift. I keep having issues getting the capture filter correct for TShark. I also don't think I will get much because of devices using the tunnel to connect to the hosted 3cx server. Any suggestions are appreciated. Here is my current TShark command script for task manager to start.
cd "c:\program files\wireshark"
: Set interface name -i
: Set packet filter in libpcap filter syntax -f
: Set Stop duration in seconds -a
: set Capture output buffer duration in seconds -b
: -V add packet tree details
: TimeStamp
set TimeStamp=%date:~4,2%%date:~7,2%%date:~10,4%_%time:~0,2%%time:~3,2%%time:~6,2%
: set Capture File -w
tshark.exe (port 5060 or port 5090 or udp portrange 9000-10999) -P -i 9 -a duration:32400 -b duration:3600 -V -w C:\Digit3\WireShark\capture_%TimeStamp%.pcap

I also tried -f (port 5060 or port 5090 or udp portrange 9000-10999)
and variations of those using opening/ending quotes ".."
yes, all those constants are marked out as I was attempting to simplify but wanted a reminder of what they were.