[ten9876]

Fixes #6092.

What was wrong

The panadapter SQL line and Auto SQL assume every radio's squelch level means −160 + level dBm. That is Flex's scale. On the HL2 it is wrong in three ways:

• The gate is fixed in raw dBFS. amsq opens at −140 + 0.7·level dBFS (WdspChannel::levelSquelchThresholdDbfsForLevel), so every dB of LNA gain moves the antenna-referred point where squelch opens by a dB. on8st's re-run on a real HL2 (HL2: panadapter SQL line and Auto SQL use the Flex −160+level scale; the HL2 amsq gate is −140+0.7·level dBFS #6092, 2026-10-03) shows the map was fitted at LNA −12 dB. At the shipped +20 dB, level 50 sits 24 dB or more below the no-signal floor: it opens below about level 85 and never closes on a carrier.
• The pan is not raw dBFS. HL2 bins carry offsetDb() = fullScale − LNA (−17 dB at +20). The line was drawn on Flex's scale with neither offset.
• Auto SQL thresholds pan bins against a passband detector. amsq captures after the passband filter and before the AGC (xamsqcap, third_party/wdsp/upstream/RXA.c:650, after xnbp(nbp0) at :647, before xwcpagc at :665). It compares an averaged magnitude against 10^(dB/20) (amsq.c:125-126, :221-224). Pan noise is per bin, so the two differ by an amount that depends on both the filter width and the bin width. That is on8st's point on the issue.

The fix (Jeremy's ruling on #6092: C+R)

1. The HL2 amsq gate is referred to the LNA. gate = −108 + 0.7·L − lnaOffsetDb() dBFS. It equals on8st's map at LNA −12 and tracks the LNA dB for dB, the same way the AGC ceiling already does.
   • WdspChannel stays family-neutral. setSquelch(on, level, levelOffsetDb) adds a caller-supplied offset to the amsq map only, and the channel has no gain of its own to refer to.
   • The HL2 owns the referral in Hl2DbReference::levelSquelchOffsetDb(). That class already "owns everything relating raw dBFS to dBm, plus the AGC setpoint that must move with it" (docs/agents/backends.md: radio-facing logic stays in the family backend).
   • Hl2Backend sends the offset with every squelch push and re-pushes on every LNA change (pushEffectiveLnaGain), including auto RF gain steps. Hl2RxDsp keeps the last finite offset if it is handed a non-finite one, so its retry loop only ever means "the channel is busy".
   • The −12 dB anchor is "supported, not established", in on8st's words. His 27 Sep no-signal and carrier brackets fit the −12 row, and the busy-40 m bracket does not. The 1:1 shift is supported (carrier moved 10.9 dB for 12 and 19.3 for 20, with 2.1 dB fading between repeats), not proven.
   • At the default +20 dB LNA, his rows put the working range at a few steps around level 50: noise opened the gate up to about 49, his carrier up to about 54. radio-capabilities-map.md says so.
2. A new optional capability record, RadioCapabilities::squelchLevelScale (core/backends/SquelchLevelScale.h). It holds offset, dB per step, modes, and whether Auto SQL may use it. Absent means no SQL line and no Auto SQL.
   • Flex: {−160, 1, all modes, Auto}, unchanged behaviour.
   • HL2: −140 + (G + 12) + (3 − G) + 6.02 = −118.98 + 0.7·L on the pan axis, in AM/SAM/DSB/LSB/USB only. The LNA terms cancel, so the record is a constant. FM is excluded: fmsq is a noise-quieting gate with no dB place.
   • Auto SQL is off on the HL2 for the detector-mismatch reason above. If it comes back, the floor should come from the detector side; a pan-derived floor would need both widths applied.
   • Icom, the demo, ANAN and RTL-SDR: legacyDbmSquelchScale(), i.e. Flex's {−160, 1, all modes, Auto} (maintainer decision on this PR). None of them has a measured map of its own, so their SQL line and Auto SQL behave exactly as on main. On main, ANAN and RTL-SDR had both features in the GUI too: the SQL button is live in voice modes, so the line drew and Auto SQL ran. They therefore get the same record, so nothing regresses. Kiwi keeps its own server-relative branch, untouched.
   • The +6.02 dB carrier term is folded in only because it is verified on both sides. Hl2Spectrum divides by half the window sum, so a steady carrier of magnitude A reads 20·log10(A) + 6.02. squelch_level_scale_test pins this against a synthetic carrier. amsq opens on the magnitude itself, which wdsp_channel_test's new carrier case pins: ±2 dB around the gate, at two offsets. A carrier off a bin centre reads up to 1.4 dB lower (Hann scalloping).
   • No ratchet growth. check_capability_records.py --strict stays at the frozen count (70/70 on current main): the record is an optional, and its autoSquelch bool sits inside the nested struct.
3. The GUI reads the record.
   • SpectrumWidget draws the line from thresholdDb(level). Auto SQL computes levelForThresholdDb(floor + margin), which is bit-for-bit the old Flex result across −175..−40 dB, and returns early when the record is absent or has no Auto SQL.
   • MainWindow::syncSquelchScaleToUi() pushes the record for the active slice's mode. It runs on a capability change, an active-slice switch (before RxApplet::setSlice), a slice mode change, and when a pan is added. autoSquelchShouldRunOnSpectrum gates on it too.
   • The SQL button cycle skips Auto where Auto is unavailable (RxApplet; VfoWidget mirrors through it). Every route into Auto, including restore, lands on Manual instead. A slice already in Auto drops to Manual with squelch on at the operator's manual level.
   • The button stays visible and enabled, because Off and Manual still work. The reason goes on its accessibleDescription and tooltip (translated), mirrored on the VFO's SQL button. Before a slice is active the gate stays permissive, as it is while disconnected.

Behaviour changes to note

• Icom, the demo, ANAN, RTL-SDR: no change. They keep Flex's scale for the line and Auto SQL.
• HL2 FM, CW and data modes: no SQL line. fmsq has no dB place.
• HL2 manual squelch: at the default +20 dB LNA, a given level now gates 32 dB higher (in raw dBFS) than before. That restores on8st's measured calibration at the gain he measured, which is the point of the change.

Validation

All of this was run locally on Arch Linux with an offscreen Qt platform. No HL2 or FLEX hardware was touched.

The unit tests, the full suite and the static gates were re-run at 1c57410, after main (3e4b14e) was merged in with a signed merge. The click table below was recorded at 33711e2. The reconnect check was run at 8bc6ebc.

Unit tests (socket-free, mutation-checked).

• squelch_level_scale_test (new).
  • The gate formula at LNA −12/−4/+8/+20/+48 dB and levels 0..100.
  • The 1:1 LNA shift.
  • At level 50, the gate sits between on8st's no-signal and carrier brackets on all four rows he measured: +20 → −73.0 (noise top −73.5, carrier −70.0); +8 → −85.0; −4 → −97.0; −12 → −105.0.
  • The record each backend publishes: Flex and HL2 their own; Icom, Demo, ANAN and RTL-SDR pinned to {−160, 1, all modes, Auto}.
  • A synthetic carrier through Hl2Spectrum reads 20·log10(A) + 6.02, and a carrier at the amsq gate peaks exactly on the drawn line at LNA −12/+8/+20.
  • Flex's Auto SQL level matches the old −160 + level arithmetic across −175..−40 dB.
  • The mode filter (HL2 FM and CW give no line; absent gives nothing).
• wdsp_channel_test.
  • The offset reaches amsq and only amsq, and a non-finite offset is refused at the channel.
  • New carrier-threshold case: a 50% AM carrier 2 dB above the gate opens amsq (audio equals the squelch-off reference), and 2 dB below stays muted (0). Checked at offset 0 (gate −105 dBFS) and at +32 (gate −73).
• hl2_rxdsp_squelch_test.
  • The offset reaches the channel and survives configure().
  • A NaN offset keeps the last good one and leaves nothing pending.
  • Through Hl2Backend on its real I/O thread: setSliceSquelch lands at the current LNA's offset, setLnaAutoOffsetDb(10) moves the gate −10 dB, and 0 moves it back.
• rx_applet_squelch_reconciliation_test.
  • With Auto unavailable, the cycle runs Off→Manual→Off.
  • The reason is on both the RxApplet and VFO buttons, and both buttons stay visible and enabled.
  • A restored Auto intent lands on Manual.
  • A slice in Auto drops to Manual with squelch on at its manual level.
• Mutation checks. Each mutant was applied, the test target rebuilt, the test run, and the file restored (sha256-checked). All 16 were killed:
  • LNA referral returning 0
  • slope dropped from thresholdDb
  • the +6.02 term dropped
  • FM added to the HL2 modes
  • Flex record removed
  • cycle skip removed
  • the setSqlMode Auto coercion removed (this one survived at first and was killed by the added restore-path test)
  • Auto→Manual drop removed
  • the reason dropped from the description
  • the VFO mirror description removed
  • the channel ignoring the offset
  • the legacy record removed from Icom, from the demo, and from ANAN (three mutants)
  • the NaN-offset guard in Hl2RxDsp::setSquelch
  • the pushSquelch in Hl2Backend::pushEffectiveLnaGain

Full local suite. ctest -j8 ran 563 tests: 561 passed, one skipped by design, and connection_panel_size_test failed, which is known on this host (#5811). Excluded: 68 tests whose sources use a socket API, or whose names match the socket-test families (automation, TCI, rigctl, CAT, discovery, bridge, DAX and so on), classified by grepping each test's sources from tests.cmake.

Agent automation bridge (demo sim, offscreen, own socket, AETHER_AUTOMATION_NO_TX=1).

• Connected serial DEMO-0001; get radio returns AetherSDR Demo.
• The demo now publishes Flex's scale. dumpTree shows "Squelch mode" with the description "Cycle squelch through Off, Manual, and Auto modes".
• Three invoke RxApplet/Squelch mode click calls, reading the button and get slice after each:

Click	Button	Slice squelch
start	SQL	off, level 20
1	SQL	on, 20
2	AUTO	on, 66 (Auto SQL picked floor + 10 + 160 from the pan)
3	SQL	off, 20

That is main's behaviour, unchanged.

• Reconnect with Auto on (at 8bc6ebc): disconnect and then connect … DEMO-0001 again, which goes through the connected-but-no-active-slice window. Afterwards the button still reads AUTO (slice on, 66), with the Auto-available description, so it is not demoted to Manual.
• The Auto-unavailable path: only the HL2 publishes it now, so the demo cannot reach it. It is proven in rx_applet_squelch_reconciliation_test: the cycle skips Auto, the reason reaches both buttons, a restored Auto intent lands on Manual, and a slice in Auto drops to Manual.

Not verified:

• No HL2 hardware run. The HL2 path is proven against on8st's measured rows and synthetic IQ only. Two things still need an on-air check: the anchor (−12 dB) and the 1:1 shift.
• No pan screenshot. Offscreen, the spectrum renderer is unavailable on this host, with or without AETHER_NO_GPU=1, so the line could not be grabbed.

Static gates (merge base vs head)

Run on the merge base and on the head; findings diffed per file.

• check_engine_boundary.py --strict: 102 → 102 findings. The only diffs are line shifts in MainWindow_Wiring.cpp from one added include.
• gen_touchpoint_manifest.py --check: up to date after regeneration. One new row, core/backends/SquelchLevelScale.h, tagged universal, with 2 includers (SpectrumWidget.h, MainWindow_Wiring.cpp).
  • Maintainer decision (Jeremy, on this PR): the new touchpoint is accepted — core/backends/SquelchLevelScale.h, universal, 2 GUI includers. It is the record type and its pure mode/Auto predicates. It follows the NoiseFloorAutoAdjustGate.h pattern, so the widget, MainWindow and the test read one function rather than copies.
• check_capability_records.py --strict: 70 → 70 booleans. The record is a std::optional, and its autoSquelch bool is nested.
• check_command_plane.py --strict: 400 → 400 (main's current baseline).
• Also clean: check_test_registration, check_ci_test_gate, check_qsettings, gen_seam_probe_table --check. check_a11y reports 16 findings on the 8 touched GUI files, the same as main.

Thanks to @on8st for the hardware sweeps (the original map on #5982 and the LNA-pinned re-run on #6092), and for the RXA.c/amsq.c analysis of what amsq actually reads.

🤖 Generated with Claude Code

https://claude.ai/code/session_01AADGw3HVn5FMMVrx2RQZ9g

[github-actions]

Accessibility check findings

This PR touches src/gui/ files with 16 accessibility issue(s).
Full details appear as inline annotations on the diff above.

Files scanned : 8
Files with findings: 4
Total findings : 16

See docs/a11y.md for the canonical
Qt patterns (setAccessibleName, QAccessibleValueChangeEvent,
QAccessibleInterface subclasses, etc.). Structural remediation
(keyboard navigation, custom QAccessibleInterface subclasses)
is tracked in #3288.

Findings are warnings only — the build is not blocked.

[aethersdr-agent]

Issue fit

Yes. #6092's three defects map cleanly onto the diff: the −160 + level line scale (SpectrumWidget::drawSquelchLine now reads thresholdDb()), the Auto SQL conversion (levelForThresholdDb() + the autoSquelchAvailable() early return), and the missing capability gate (autoSquelchShouldRunOnSpectrum + syncSquelchScaleToUi). The shape is the one the maintainer ruled on in the thread ("The plan (my decision): 1. a backend-published squelch-scale record… 2. refer the HL2 amsq map to the LNA… 3. Auto SQL stays hidden on the HL2"), with on8st's 2026-10-03 re-anchoring at −12 dB folded in. I checked the formula against his correction: −108 + 0.7L − lnaOffsetDb() and the implemented −140 + 0.7L + (G + 12) are the same function (−128 + 0.7L + G), since kReferenceLnaGainDb == kDefaultLnaGainDb == 20. The test asserts the first form and the code computes the second, which is the right way round.

Coverage is at the right layer — capabilities and DSP mapping, socket-free CTest. No socket test is added or modified; nothing in the diff touches QTcpServer/QUdpSocket/QLocalServer/bind/listen/connectToHost, and no fake radio peer appears.

Scope

File / group	What it changes	Claimed?	Verdict
core/backends/SquelchLevelScale.h (new)	the record + squelchScaleForMode / autoSquelchAvailable predicates	yes	in scope; new public surface, pre-ruled in the thread
RadioCapabilities.h	std::optional<SquelchLevelScale> squelchLevelScale	yes	in scope
flex/FlexBackend.cpp, hl2/Hl2Backend.cpp	publish the record	yes	in scope
hl2/Hl2DbReference.h, Hl2RxDsp.{h,cpp}, dsp/WdspChannel.{h,cpp}	the LNA referral and its plumbing	yes	in scope
hl2/Hl2Spectrum.h	kToneGainDb constant	yes	in scope — names existing behaviour, no code change
gui/SpectrumWidget.{h,cpp}, MainWindow*.{h,cpp}, RxApplet.{h,cpp}, VfoWidget.cpp	read the record; skip/drop Auto; a11y text	yes	in scope
docs/architecture/* (3 files)	caps-map row, touchpoint manifest + tags	yes	in scope; manifest is generated, 233→234 is consistent
tests/* (5 files) + tests.cmake	new target, signature updates	yes	in scope

Nothing in the diff is unexplained by #6092 or its thread. No CHANGELOG.md entry (correct). No deleted guard — the only - lines of consequence are the removed kSqlMinDbm constants and the inlined SQL tooltip, both replaced. The three disclosed behaviour changes (Icom and Demo lose the line and Auto; HL2 manual squelch gates 32 dB higher at the default LNA) are all consequences of the ruled shape, not smuggled preference: the issue and the triage comment both establish that Icom's 0..255 register has no published dB mapping, so the line there was drawn on a scale that never applied.

Governance. New capability field arrives with the full trio (RadioCapabilities + radio-capabilities-map.md row + squelch_level_scale_test gating test), per AGENTS.md. New universal touchpoint header follows the NoiseFloorAutoAdjustGate.h precedent, and the justification is in the PR body. I re-ran the gates on the head checkout myself: check_capability_records.py --strict → 71 boolean(s), at the frozen 71 — ok, gen_touchpoint_manifest.py --check → manifest up to date, check_engine_boundary.py --strict → 102 findings / 0 blocking, check_test_registration.py → OK. The body's numbers reproduce.

Blockers

None.

Nits (all non-blocking)

• A non-finite offset is refused, and the refusal is retried forever. WdspChannel::setSquelch returns false for !std::isfinite(levelOffsetDb), but Hl2RxDsp::pushSquelchToChannel cannot tell that apart from a transient control-op refusal: it sets m_squelchPending = true, and processIqBlock() then retries once per block for the rest of the session. A permanently-invalid argument and "busy, try later" want different answers. Inline comment on WdspChannel.cpp:602.
• The unavailability reason can be wrong for a radio that does publish Auto. With no active slice yet (activeSlice() == nullptr) but isConnected() true, activeSliceSquelchScale() returns nullopt, and the reason selector falls through !published (false on Flex) and s && … (false) into the detector message — so a connecting Flex can briefly announce "this radio's squelch detector does not read the panadapter's noise floor". Harmless in practice: setActiveSliceInternal re-syncs before RxApplet::setSlice restores intent, which is exactly what the ordering comment protects. Treating !s as permissive (like disconnected) would close it. Inline on MainWindow_Wiring.cpp:1389.
• At the shipped LNA default the usable slider span is ~5 steps. From on8st's +20 row and the implemented map, level 50 gates at −73.0 dBFS with his noise top at −73.5 and the carrier above −70 — the test's own 0.5 dB window. Everything below level 50 is under the floor (always open) and above ~55 is above any signal he measured. That is what the measurement says, and it is better than the pre-change behaviour (open below ~85), but the caps-map paragraph would be more useful to an operator if it said where the live range sits at +20. Inline on Hl2DbReference.h:105.
• pushEffectiveLnaGain() now issues a second beginControlOperation()-taking push per LNA step (including every auto-RF-gain step), and pushSquelch sends even when squelch is off. Same cost class as the adjacent setAgc push that was already there, so probably fine — flagging only so it is a decision rather than an accident.
• recordsPerBackend pins the HL2 mode list in publishedModeStrings() order, so a harmless reorder there fails a test that only cares about set membership. Inline on tests/squelch_level_scale_test.cpp:126.
• Mixed i18n on the same user-visible sentence: MainWindow builds the reason with tr(), RxApplet embeds it in QStringLiteral scaffolding, so a translated reason lands inside an untranslated frame. Pre-existing for the surrounding strings.

What I tried to break

• The arithmetic. Expanded the HL2 record offset by hand: −140 + (G + 12) + (3 − G) + 6.02 → −128.98 + 6.02, LNA-free, matching the test's −125 + kToneGainDb and the body's claim. Checked that a +10 dB LNA move leaves the displayed line fixed (gate moves +10 dBFS, offsetDb() moves −10) — the cancellation holds in both the gate and the pan direction, so the line does not walk with RF gain. Recomputed level 50 against all four of on8st's rows (+20 → −73.0, +8 → −85.0, −4 → −97.0, −12 → −105.0) and confirmed each sits inside his brackets, using his lower-bound −70 at +20 as the only thing that row can support.
• Flex bit-for-bit equivalence. Walked levelForThresholdDb against the deleted clamp(int(target + 160 + 0.5f), 1, 100) at both ends of the test's sweep, including the negative-steps case where static_cast<int> truncates toward zero rather than rounding up as the comment says — both land on 1, so the comment is loose but the result is identical.
• The Auto-drop as a Flex self-race (the attack that would have been a blocker): if isConnected() could be true while squelchLevelScale was absent on a Flex, every connect would silently drop a restored Auto to Manual and saveClientSquelchIntent() would persist autoEnabled = false with no path back. RadioModel::backendCapabilities() is a live call-through to m_backend->capabilities(), not a cached copy, and syncSquelchScaleToUi runs after m_activeSliceId is assigned and before m_appletPanel->setSlice(s). No window I can find.
• Stale string-based invokers. The Q_INVOKABLE arity changed, which a wrong-arity invokeMethod would fail at runtime rather than at compile time. grep -rn '"setSquelch"' src tests on the head checkout returns exactly one hit, the updated Hl2Backend::pushSquelch. Confirmed m_dbRef.setLnaGainDb() runs before the push loop in pushEffectiveLnaGain.
• The Kiwi paths. Checked that the new guards sit after the Kiwi branches in drawSquelchLine and updateAutoSquelchFromBins, that autoSquelchShouldRunOnSpectrum still returns on its Kiwi early-returns before reaching the new &&, and that RxApplet::autoSqlAvailable() ORs in usingExternalReceiveSquelch() so Kiwi keeps its own Auto. Kiwi is untouched.
• The truncated diff. The wdsp_channel_test.cpp hunk is cut off above, so I read the file: runSquelchCarrierThresholdTest exists, is registered at line 3957, and does pin the ±2 dB carrier case at offsets 0 and +32 plus the non-finite refusal at :3607. The "verified on both sides" claim for the +6.02 dB term checks out.

Not verified, and it matters: everything above is reasoned from the code and from on8st's posted brackets — I ran no build, no test and no GUI. CI green means "it compiles." Every ctest step in ci.yml is -R-filtered and none of the filters names a squelch, wdsp-channel or rx-applet test, so the author's local run and the 11 mutation kills are the only evidence those tests pass. And as the body says honestly, the −12 dB anchor and the 1:1 LNA shift are supported, not established — one on-air sweep on an HL2 is what would settle the thing this PR is built on.

Recommendation

Approve with nits — with a maintainer note. The code is careful, the seam is the one that was ruled on, the gates reproduce locally, and the LNA referral is algebraically consistent in both directions. None of the nits break a user. What a maintainer should weigh is not the code: it is that HL2 manual squelch moves 32 dB at the shipped default gain on the strength of bracket fits from one afternoon, with the author saying so plainly. Worth landing and then getting an on-air confirmation of the anchor from @on8st, rather than holding the GUI fix hostage to it. Concrete next step: the isfinite-refusal nit is a two-line change and worth taking before merge; the rest can ride.

Nice work — in particular the mutation table, the honest "supported, not established" framing, and excluding Auto SQL on the HL2 instead of shipping a plausible-looking pan-derived floor.

---

_{🤖 aethersdr-agent · cost: $8.2055 · model: claude-opus-5}

[on8st]

@ten9876 Checked 790ef15 against our rows by reading the source; nothing was built or run. Bad news first.

• The +20 row carries the anchor, and it is our weakest leg. Its floor bracket is one coarse pass (the fine passes opened once in 30), on a frequency that had a signal on it 40 minutes later; the carrier there is only a lower bound. squelch_level_scale_test.cpp:85-90 accepts an anchor from −15 to −11.5 dB with that row, and −21.4 to −8.2 without it.
• The line is exact for a steady carrier only (SquelchLevelScale.h:3-4). On noise alone the gate opens while the line is still 10·log10(passband / 1.5 bins) − 1 dB above the pan floor: 19.5 dB at 48 kHz in AM 8 kHz, 5.8 dB at 384 kHz in 2.7 kHz. Arithmetic, not measured.
• I found no test that reaches the re-push on an LNA change (Hl2Backend.cpp:7043, :3293).

Level 50 from Hl2DbReference.h:105-112 and WdspChannel.cpp:651, raw dBFS:

LNA	gate	no signal	carrier
+20	−73.0	−80.5..−73.5	above −70
+8	−85.0	−95.9..−93.8	−75.6..−71.4
−4	−97.0	−103.6..−100.8	−84.7..−82.6
−12	−105.0	−113.4..−109.9	−93.1..−91.0

Inside all four. The record includes the pan offset (Hl2Backend.cpp:1892-1893, :858), so the line holds across gain. FM is untouched (WdspChannel.cpp:677).

73, ON8ST

[github-actions]

Accessibility check findings

This PR touches src/gui/ files with 16 accessibility issue(s).
Full details appear as inline annotations on the diff above.

Files scanned : 8
Files with findings: 4
Total findings : 16

See docs/a11y.md for the canonical
Qt patterns (setAccessibleName, QAccessibleValueChangeEvent,
QAccessibleInterface subclasses, etc.). Structural remediation
(keyboard navigation, custom QAccessibleInterface subclasses)
is tracked in #3288.

Findings are warnings only — the build is not blocked.

[github-actions]

Accessibility check findings

This PR touches src/gui/ files with 16 accessibility issue(s).
Full details appear as inline annotations on the diff above.

Files scanned : 8
Files with findings: 4
Total findings : 16

See docs/a11y.md for the canonical
Qt patterns (setAccessibleName, QAccessibleValueChangeEvent,
QAccessibleInterface subclasses, etc.). Structural remediation
(keyboard navigation, custom QAccessibleInterface subclasses)
is tracked in #3288.

Findings are warnings only — the build is not blocked.

[github-actions]

Accessibility check findings

This PR touches src/gui/ files with 16 accessibility issue(s).
Full details appear as inline annotations on the diff above.

Files scanned : 8
Files with findings: 4
Total findings : 16

See docs/a11y.md for the canonical
Qt patterns (setAccessibleName, QAccessibleValueChangeEvent,
QAccessibleInterface subclasses, etc.). Structural remediation
(keyboard navigation, custom QAccessibleInterface subclasses)
is tracked in #3288.

Findings are warnings only — the build is not blocked.

[ten9876]

Maintainer pass on 8bc6ebc. This covers the points in the bot's review body and in @on8st's 19:49 check that have no inline thread. All four inline threads are answered and resolved.

on8st, untested re-push on an LNA change (Hl2Backend.cpp:7043, :3293): now pinned. In 8bc6ebc, hl2_rxdsp_squelch_test drives Hl2Backend with a configured chain on its real I/O thread:

• setSliceSquelch(0, true, 50) lands amsq at the current LNA's offset.
• setLnaAutoOffsetDb(10) moves the gate −10 dB.
• Setting the offset back to 0 restores it.

The values are read on the chain's own thread. If the pushSquelch(r) in pushEffectiveLnaGain is deleted, the test fails with FAIL: an LNA change re-pushes the squelch: the gate moves with it. It passed 20 of 20 repeat runs.

on8st, the +20 row carries the anchor. The maintainer took −12 dB as supported, not established, and that is how the PR body and Hl2DbReference.h state it. The caps map now also states the working range at +20 dB (thread on Hl2DbReference.h:105). An on-air sweep stays the thing that would settle it.

on8st, the line is exact for a steady carrier only. That is the record's stated contract (SquelchLevelScale.h:3-4, "read as the pan peak of a steady carrier"). It is also the reason Auto SQL stays off on the HL2: amsq reads passband-limited magnitude while the pan reads per bin. No code change.

Bot body: extra beginControlOperation() push per LNA step, and pushSquelch sends while off. Kept as is. The push costs the same as the adjacent setAgc push that was already there. Sending while off keeps Hl2RxDsp's held offset in step with the LNA, so a later "on", or the re-application in installChannel() after a rebuild, never uses a stale offset. pushNoiseBlanker follows the same rule.

Bot body: mixed i18n. In f77795a, the SQL button's tooltip and accessible-description frames in RxApplet::applySqlButtonDescription / sqlButtonAccessibleDescription go through tr(), matching the tr() reason that MainWindow puts inside them.

Bot review predates 33711e2. Icom, the demo, ANAN and RTL-SDR now keep Flex's {−160, 1, all modes, Auto}, so they no longer lose the line or Auto SQL. squelch_level_scale_test pins this for every backend.

Local proof at 8bc6ebc: RelWithDebInfo build is clean. squelch_level_scale_test, hl2_rxdsp_squelch_test, wdsp_channel_test, rx_applet_squelch_reconciliation_test, hl2_apf_agc_off_test and the other wdsp/hl2_rxdsp tests pass. The full socket-free ctest passes 558/559; the one failure is connection_panel_size_test, which is known on this host (#5811). The static gates match main: engine-boundary 102 findings with 0 blocking, capability records at the frozen 70, command plane 401/32, manifest up to date. The only new touchpoint row is the accepted SquelchLevelScale.h.

[github-actions]

Accessibility check findings

This PR touches src/gui/ files with 16 accessibility issue(s).
Full details appear as inline annotations on the diff above.

Files scanned : 8
Files with findings: 4
Total findings : 16

See docs/a11y.md for the canonical
Qt patterns (setAccessibleName, QAccessibleValueChangeEvent,
QAccessibleInterface subclasses, etc.). Structural remediation
(keyboard navigation, custom QAccessibleInterface subclasses)
is tracked in #3288.

Findings are warnings only — the build is not blocked.

[github-actions]

Accessibility check findings

This PR touches src/gui/ files with 16 accessibility issue(s).
Full details appear as inline annotations on the diff above.

Files scanned : 8
Files with findings: 4
Total findings : 16

See docs/a11y.md for the canonical
Qt patterns (setAccessibleName, QAccessibleValueChangeEvent,
QAccessibleInterface subclasses, etc.). Structural remediation
(keyboard navigation, custom QAccessibleInterface subclasses)
is tracked in #3288.

Findings are warnings only — the build is not blocked.

[github-actions]

Accessibility check findings

This PR touches src/gui/ files with 16 accessibility issue(s).
Full details appear as inline annotations on the diff above.

Files scanned : 8
Files with findings: 4
Total findings : 16

See docs/a11y.md for the canonical
Qt patterns (setAccessibleName, QAccessibleValueChangeEvent,
QAccessibleInterface subclasses, etc.). Structural remediation
(keyboard navigation, custom QAccessibleInterface subclasses)
is tracked in #3288.

Findings are warnings only — the build is not blocked.

[github-actions]

Accessibility check findings

This PR touches src/gui/ files with 16 accessibility issue(s).
Full details appear as inline annotations on the diff above.

Files scanned : 8
Files with findings: 4
Total findings : 16

See docs/a11y.md for the canonical
Qt patterns (setAccessibleName, QAccessibleValueChangeEvent,
QAccessibleInterface subclasses, etc.). Structural remediation
(keyboard navigation, custom QAccessibleInterface subclasses)
is tracked in #3288.

Findings are warnings only — the build is not blocked.

[github-actions]

Accessibility check findings

This PR touches src/gui/ files with 16 accessibility issue(s).
Full details appear as inline annotations on the diff above.

Files scanned : 8
Files with findings: 4
Total findings : 16

See docs/a11y.md for the canonical
Qt patterns (setAccessibleName, QAccessibleValueChangeEvent,
QAccessibleInterface subclasses, etc.). Structural remediation
(keyboard navigation, custom QAccessibleInterface subclasses)
is tracked in #3288.

Findings are warnings only — the build is not blocked.

[ten9876]

Admin-merged on the maintainer's approval (self-authored PR). Head 1c574109:

• HL2 amsq gate referred to the LNA, anchored at −12 per @on8st's re-measured sweep. That anchor is "supported, not established" and still needs an on-air check.
• squelchLevelScale record. Icom, the demo, ANAN and RTL keep the Flex scale; HL2 has its own line and no Auto SQL.
• Touchpoint core/backends/SquelchLevelScale.h accepted.
• All bot threads answered and resolved. Required checks green, from a run created after main's tip; no drift.

[ten9876]

PR #6127 — post-merge review of 1c574109331ed0855331d2eb3c822813533a54ea (merged while this review was running).

Issue fit: The implementation addresses #6092's HL2 scale mismatch: the backend publishes the line's scale, the DSP gate follows LNA gain, and Auto SQL is unavailable where a passband detector cannot use the per-bin pan floor. The issue thread explicitly authorizes this shape; the later maintainer PR comment accepts the new touchpoint and retaining legacy mappings on Icom/demo/ANAN/RTL. Those families' calibration is intentionally not solved here. Hardware confirmation of the −12 dB anchor remains outstanding and is disclosed.

Scope

File/group	Change	Claimed?	Verdict
RadioCapabilities.h, new SquelchLevelScale.h	Optional record, forward/inverse mapping, mode/Auto predicates	Yes	In scope; new surface has maintainer ruling
Flex/Icom/sim/ANAN/RTL backend declarations	Preserve existing line and Auto behavior	Yes	In scope; accepted compatibility choice
hl2/Hl2Backend.{h,cpp}, Hl2DbReference.h, Hl2RxDsp.{h,cpp}	LNA referral, queued offset propagation, gain-change re-push, test access	Yes	In scope
dsp/WdspChannel.{h,cpp}, hl2/Hl2Spectrum.h	Offset only on level squelch; named existing carrier normalization	Yes	In scope
MainWindow.{h,cpp}, MainWindow_Wiring.cpp, SpectrumWidget.{h,cpp}	Publish active-mode scale to rendering and Auto algorithm	Yes	In scope
RxApplet.{h,cpp}, VfoWidget.cpp	Skip unavailable Auto, coerce restored Auto, explain availability	Yes	In scope; issue-authorized correction
Four test sources and tests/tests.cmake	Mapping, DSP, lifecycle and UI regressions	Yes	In scope
Three docs/architecture/ files	Capability map and generated touchpoint/tag entry	Yes	In scope; new universal include explicitly accepted

All 31 changed files are explained by the issue and accepted follow-up decisions. Commit author dates and messages reveal no unrelated bundled work. No CHANGELOG entry or ratchet weakening. The boundary audit reproduced 102→102 findings (only line-number changes), 70→70 capability booleans and 400→400 above-seam commands; both manifests passed. The new universal touchpoint is the documented 236→237 growth, not a vendor reclassification. No new family-string branch, concrete-backend cast, wire command above the seam, thread or TX path was introduced.

Blockers

None found.

Nits — non-blocking

1. Isolate the newly settings-touching HL2 test. tests/hl2_rxdsp_squelch_test.cpp:89 starts QCoreApplication without TestSettingsProfile, and the new backend at :194 calls AutomationBridgeSettings::txAllowed() during construction. That instantiates AppSettings, whose constructor creates the normal configuration directory and runs the platform path-migration hook. The generated CTest environment only supplies the WDSP planner/wisdom overrides. With AETHER_SETTINGS_DIR unset and XDG_CONFIG_HOME redirected to a fresh scratch directory, I observed the head test create AetherSDR/; the identical probe against the built merge-base test did not create it. This was an isolated probe, not an experiment on the operator's profile; I did not observe settings-value modification on Linux. Add the same TestSettingsProfile used by the new scale test before QCoreApplication, and check isValid(). This follows the project test-isolation convention and keeps the backend fixture self-contained.

   Add #include "TestSettingsProfile.h", then start main with:

       TestSettingsProfile settingsProfile(QStringLiteral("hl2-rxdsp-squelch-test"));
       if (!settingsProfile.isValid()) {
           return 1;
       }
       QCoreApplication app(argc, argv);

What I tried to break

• Built the exact PR head in an isolated worktree. All four focused socket-free tests passed: squelch_level_scale_test, hl2_rxdsp_squelch_test, wdsp_channel_test, rx_applet_squelch_reconciliation_test. The GUI test initially hit the host's inherited GTK3 platform theme/display dependency; it passed with offscreen generic/Fusion styling. The preflight covered full test sources and the cold backend constructors, including Flex's inert socket allocation; no test opens/binds/connects a socket or runs a synthetic firmware peer.
• Mutation-checked the actual test sources using separate compiled objects so the production build remained unchanged: dropping the line slope caused ten scale-test failures; omitting the WDSP offset failed both WDSP and HL2 tests, including reconfigure and LNA-change assertions; removing the Auto coercion failed restoredAutoIntentLandsOnManualWhenUnavailable (Actual: 2, Expected Manual: 1). All mutants exited 1. The unmodified DSP test also exercised the ±2 dB AM carrier gate at offsets 0 and 32.
• Drove the head's own offscreen simulator over /tmp/6127-demo-head/bridge.sock, isolated settings, AETHER_AUTOMATION_NO_TX=1. get radio confirmed "model":"AetherSDR Demo","serial":"DEMO-0001","transmitting":false. Clicking the real RxApplet button produced (squelch,level) false/20 → true/20 → true/66 → false/20; manual level stayed 20. Auto remained on at level 66 after disconnect/reconnect and USB→AM→USB. The accessible description retained all three modes. Only this owned instance was driven and it was closed.
• Built exact merge base 3e4b14e2e and ran the identical demo script on its own /tmp/6127-demo-base/bridge.sock. All eight state snapshots matched the head, including manual-cache preservation, reconnect and the mode round trip. Re-launched both builds with their respective existing isolated profiles: each starts Off/20 and then produces the same sequence. Thus this demonstrates unchanged restart behavior, not cross-process Auto persistence. All four runs were explicitly compared by assertions. Both owned demo instances were closed after every run; the merge-base HL2 and RxApplet tests passed as well.
• Read the failure paths and sibling surfaces: the non-finite-offset guard prevents permanent retries; the queued invoker uses the new arity; gain is updated before re-pushing; FM's detector ignores the added level offset; Kiwi retains its early-return path and availability override. The existing review's stale no-active-slice and non-finite-offset concerns are fixed in this head. No independent automated-review skill was available.

Evidence limits: The simulator cannot verify the HL2 hardware anchor or RF-gain transfer, nor its Auto-unavailable mode through a real HL2 session; the latter is covered by the injected UI and DSP tests. The screenshot request returned {"error":"grab produced an empty image for SpectrumWidget","ok":false}; no pixel-level line-placement proof is claimed. Synthetic carrier FFT arithmetic and the line mapping were tested, including all gain/level combinations in a separate header harness. No live radio was used. The head's CI and Static Checks were successful, created at 2026-10-04 01:28:12 UTC, after its included main tip's 01:13:20 UTC commit; this is not a claim that all tests were CI-gating or that the later squash has been independently certified.

Recommendation

Approve with nits, as a post-merge assessment rather than a pending merge action. No user-facing regression was found in the tested paths. Add test-profile isolation in a small follow-up and obtain the disclosed HL2 hardware confirmation of the anchor when feasible; this review does not substitute simulator evidence for RF measurements.