chan_voter: ADPCM not working for clients

[ve7fet]

While testing for Nulaw removal, it appears that ADPCM audio mode doesn't work when using normal voting clients. It appears that it DOES work with mix-mode clients though.

The node will key in ADPCM mode, but the transmitter stays on long after the channel driver says it is unkeyed, and no audio is repeated (however, there are SOME audio artifacts in there). Telemetry tones are also not being sent.

Went back and tried and older version (PR #719) and it was broken back then too... so I suspect it has been broken for a long time.

Will look deeper once the extra Nulaw code is removed. Suspect there may be some bad logic in the voter_xmit function.

[coderabbitai]

📝 CodeRabbit Plan Mode

Generate an implementation plan and prompts that you can use with your favorite coding agent.

• Create Plan

Examples

• Example 1
• Example 2

---

🔗 Similar Issues

Related Issues

• chan_voter: Remove un-implemented "Nulaw" audio code #872
• chan_voter: VOTER (master) client cycles between connected and offline mode if the associated node in app_rpt is broken #845
• duplex=3 is now COMPLETELY BROKEN. #703
• Time out timer local reset functionality broken #754

👤 Suggested Assignees

• ve7fet
• davidgsd
• hwstar

---

🧪 Issue enrichment is currently in open beta.

You can configure auto-planning by selecting labels in the issue_enrichment configuration.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord or schedule a call!

[ve7fet]

Did some more testing, putting clients in ADPCM mode does not work for either voting or mix-mode clients.

There may be issues in chan_voter.c as well as the actual VOTER firmware. I can see inconsistencies in the application of buffer/frame sizes, and possibly transcoding issues in chan_voter.

Tried poking at a few things, but was not successful in making it work.

I'm on the fence about whether time would be better spent fixing a low quality CODEC that no one is using (since it is broken), or removing all the ADPCM code entirely, which would free up valuable space in the VOTER EEPROM for other features.

[ve7fet]

Compiling chan_voter with #define ADPCM_LOOPBACK forces chan_voter to take the incoming packets from the client, and send them right back to the client, with no other processing.

This works. So, the ADPCM implementation in the firmware of the VOTER/RTCM does appear to function correctly.

That should mean that the trouble with ADPCM not working lies in chan_voter (and I can see some issues in there already). At least I think I know where the focus needs to be now.