Merge Plan

Thanks for this PR — the DAX edge-sync and SSB delay fixes are well-documented and the engineering report is thorough. CI passes clean. Here's our plan for merging.

Merge approach

We'll rebase onto current main ourselves in a testing branch. Several files have diverged since this PR was opened (MIDI controller support, FlexControl, FreeDV spots, USB cables all landed today in the v0.7.9 cycle), so we expect conflicts in MainWindow.cpp and AudioEngine.cpp.

Conflict resolution plan

File-by-file analysis:

1. AudioEngine.h — moderate conflicts

• PR adds: setDaxTxMode(), setDaxTxUseRadioRoute(), setTransmitting(), clearTxAccumulators(), m_daxPreTxBuffer, m_daxTxUseRadioRoute, PCC_DAX_REDUCED, TX_SAMPLES_PER_PACKET, TX_PCM_BYTES_PER_PACKET
• Our changes: setPcMicGain(), pcMicLevelChanged signal, m_pcMicGain, mic metering members
• Conflict resolution: Take both — PR adds DAX/TX members, ours add mic metering members. No overlap.
• Watch out: PR replaces #include <cmath> with #include <algorithm>. We need BOTH — <cmath> is used by std::log10 in mic metering (added in v0.7.8). Must keep both includes.

2. AudioEngine.cpp — significant conflicts

• PR changes:
  • Expands feedDaxTxAudio() with two routing modes (radio-native vs low-latency)
  • Adds setDaxTxMode(), setTransmitting(), setDaxTxUseRadioRoute() methods
  • Clears TX accumulators on unkey
• Our changes:
  • onTxAudioReady() — removed separate VOX path, added PC mic gain, added mic metering, changed Opus frame size to 240, changed VITA-49 header fields
  • Added processBnr() diagnostic logging (already reverted)
• Conflict resolution: The onTxAudioReady() function will be the main conflict zone. PR's changes are in feedDaxTxAudio() and new methods (after onTxAudioReady), while ours are inside onTxAudioReady(). Should be resolvable by keeping both sets of changes in their respective functions.
• Critical check: PR's setTransmitting() clears accumulators including m_txFloatAccumulator. Our Opus path uses m_opusTxAccumulator — verify PR doesn't clear that, or add it to the clear list.

3. MainWindow.cpp — significant conflicts

• PR changes:
  • New moxChanged handler for immediate SSB TX audio gating
  • New txAudioGateChanged handler (TX-off safety fallback)
  • DAX bridge TX callback mode guard
• Our changes (v0.7.8-v0.7.9):
  • MIDI controller wiring (50+ params, registerMidiParams())
  • FlexControl wiring (tuneSteps, buttonPressed)
  • FreeDV client creation and spot forwarding
  • USB Cables dialog wiring
  • MIDI Mapping menu entry
  • PC mic gain wiring (setPcMicGain, pcMicLevelChanged)
• Conflict resolution: PR touches the TX state handling section (~line 450-500 area). Our changes are mostly in different sections (MIDI at end of file, FlexControl at ~line 1050, FreeDV at ~line 230, menu items at ~line 1530). The moxChanged handler overlap is the key conflict — we need to merge PR's immediate-gating logic with our existing moxChanged handler.

4. MainWindow.h — minor

• PR adds: one member (likely DAX-related)
• Our changes: MidiControlManager, FlexControlManager, FreeDvClient members
• Conflict resolution: Take both.

5. RadioModel.cpp — moderate conflicts

• PR adds: m_txRequested tracking, txAudioGateChanged signal, interlock fallback logic
• Our changes: UsbCableModel status routing, usb_cable handler, commandReady wiring
• Conflict resolution: PR touches the interlock handler section. Our changes are in the status routing section (different area). Should merge cleanly after rebase. Verify the interlock handler doesn't conflict with our FlexControl/MIDI MOX toggle path.

6. RadioModel.h — minor

• PR adds: txAudioGateChanged signal, m_txRequested, m_txAudioGate members
• Our changes: UsbCableModel member and accessor
• Conflict resolution: Take both.

7. TransmitModel.cpp — minor

• PR changes: Adjusts setTransmitting() behavior
• Our changes: None in this file since v0.7.7
• Conflict resolution: Take PR's changes directly.

8. VirtualAudioBridge.cpp — no conflicts

• PR changes: Reduced TX poll interval, backlog clamping
• Our changes: None in this file
• Conflict resolution: Take PR's changes directly.

9. TX_SYNC_FIX_REPORT.md — housekeeping

• Move to docs/TX_SYNC_FIX_REPORT.md instead of repo root
• Or remove from the merge and reference the PR description instead

Testing plan

After rebase, test these scenarios:

1. SSB voice TX (hardware mic) — PTT press should produce immediate audio, no delay. Release should cut cleanly.
2. SSB voice TX (PC mic) — same as above, with Opus encoding on remote_audio_tx
3. DAX TX (WSJT-X/VARA) — digital mode TX via virtual audio should start/stop cleanly at PTT edges
4. VOX — voice-triggered TX should still work (mic audio streams during RX for met_in_rx=1)
5. MIDI MOX toggle — our new MIDI tx.mox parameter should work with the new TX gating
6. FlexControl MOX button — same as above
7. DVK recording — verify recording works with both hardware and PC mic (related to Looks like DVK got broken in 0.7.8 #354)
8. BNR — verify RX audio path unaffected

Timeline

We'll create a test/pr353-dax-ssb-sync branch, rebase, resolve conflicts, and test. Will report back here with results.

73, Jeremy KK7GWY & Claude (AI dev partner)

Merged in 08ae00e. Tested and verified:

• SSB voice TX: immediate edge timing, no delay on key-down, clean release
• WSJT-X FT8 DAX TX: working with radio-native route (dax=1, default)
• VOX: working
• DVK playback: working
• BNR: unaffected

Adjustments made during merge:

• Default DaxTxLowLatency changed from True to False — standard DAX (WSJT-X, VARA, fldigi) works out of the box. FreeDV/RADE users opt in via Settings → "Low-Latency DAX (FreeDV)"
• Menu label shortened to "Low-Latency DAX (FreeDV)"
• TX_SYNC_FIX_REPORT.md moved to docs/

Thanks for the thorough fix and documentation, @pepefrog1234!

73, Jeremy KK7GWY & Claude (AI dev partner)