DVSwitch, DMR and xlxd - one way audio.


 

I've got a weird issue.  Running a fairly complex multi site transcoding
network with USRP between sites.  The issue concerns two sites.  One
site transcodes DMR to PCM over USRP.  There is also an AllStar node at
that site.

The other site has multiple transcoders from PCM/USRP to D-STAR, Fusion
(which is in turn bridged to NXDN), M17, P25 and an ASL node (which has
its own bridge on to an IRLP reflector).

The issue I have concerns the interaction between the D-STAR transcoder
and xlxd, which ircDDBGateway links to.  Until a few days ago, I was
using a patched version of DummyRepeater to do the transcoding.  This
was so I could make use of an old DV Dongle that I had lying around. 
This transcoder worked well, except that there's no audio level
adjustments on the PCM side, and it doesn't pass on metadata (N0CALL
gets around a bit ;) ).  With this transcoder, everything works as
expected, all modes on all sites can hear each other.

During the week, I migrated from DummyRepeater to DVSwitch, using a
ThumbDV (purchased a while ago but finally got it working), with a copy
of Analog_Bridge and MMDVM_Bridge replacing DummyRepeater on the same
USRP and ircDDB ports.  DVSwitch is setup to be a drop in replacement
for DummyRepeater.  Most things work, except as indicated.

D-STAR (on site 2) works to all modes locally, as well as everything on
site 1 (DMR and AllStar), and the level tweaks have improved the audio. 

From site 1 to site 2, AllStar works to all modes on site 2, including
D-STAR.  However, DMR works to everything on both sites, except D-STAR. 
Tracing through the network reveals that DMR audio and data do get to
the USRP backbone between sites, then into Analog_Reflector on site 2,
and into Analog_Bridge and MMDVM_Bridge on the D-STAR branch.  Data also
gets sent to the XLX reflector from MMDVM_Bridge, even when DMR is the
source.  However, the XLX reflector doesn't relay the transcoded signal
to other D-STAR clients.

I did investigate, and the only thing I can see is the metadata from DMR
looks different, containing DMR IDs, instead of callsigns, which every
other mode seems to generate, and this difference propagates through the
entire network.

This is what Analog_Reflector on site 2 sees.

From D-STAR (via DVSwitch)

M: 22-05-21 06:24:18 Begin TX: src=5053029 rpt=505302878 dst=9 slot=2
cc=0 metadata=VK3JED
M: 22-05-21 06:24:20 End TX
M: 22-05-21 06:24:22 Begin TX: src=5053029 rpt=505302878 dst=9 slot=2
cc=0 metadata=VK3JED
M: 22-05-21 06:24:25 End TX
M: 22-05-21 06:24:25 Begin TX: src=5053029 rpt=505302878 dst=9 slot=2
cc=0 metadata=VK3JED
M: 22-05-21 06:24:25 Transmission double, ignoring your stream
M: 22-05-21 06:24:47 Seems I missed the EOT for the last transmission,
resetting...

From DMR

M: 22-05-21 06:24:47 Begin TX: src=5053028 rpt=500000569 dst=969 slot=2
cc=0 metadata=5053028
M: 22-05-21 06:24:54 End TX
M: 22-05-21 06:26:30 Begin TX: src=5053028 rpt=500000569 dst=969 slot=2
cc=0 metadata=5053028
M: 22-05-21 06:27:16 End TX

From D-STAR (via DummyRepeater)

M: 22-05-21 06:27:20 Begin TX: src=5053029 rpt=505302878 dst=9 slot=2
cc=0 metadata=VK3JED
M: 22-05-21 06:27:34 End TX

From Site1 AllStar (same site as DMR)

M: 22-05-21 06:58:36 Begin TX: src=5000005 rpt=0 dst=21992 slot=2 cc=0
metadata=VK3JED
M: 22-05-21 06:58:45 End TX


--
73 de Tony VK3JED/VK3IRL
http://vkradio.com


Mike Zingman - N4IRR
 

Dstar requires callsigns. Your metadata did not resolve to a call. Bad lookup?  Empty database?


 

On 22/5/22 5:18 am, Mike Zingman - N4IRR wrote:
Dstar requires callsigns. Your metadata did not resolve to a call. Bad
lookup?  Empty database?
That thought did cross my mind.  I'll have to check with the operator at
the other end to see how they're doing the transcoding from DMR to USRP
and what databases are in use there. 

--
73 de Tony VK3JED/VK3IRL
http://vkradio.com