task-39: point at the eight unmerged mic branches before theorising again

None reached blueline. The new measurement rules out the decimator and the
front-end mixer, which favours the sysclk/SLIM-lifecycle direction of mic-fix
over the pad-configuration branches.

docs/tasks/39-audio-levels-as-state.md

@@ -211,3 +211,25 @@ of the loudspeaker with no way back"), and the hook is enabled, active and
 unshadowed (`FragmentPath=/usr/lib/systemd/system/...`). So something else put
 the card in a speaker profile before or during that call — unreproduced, and it
 needs a call to chase.
+
+### Prior art: this has been attacked before — read the branches first
+
+`linux-blueline` carries at least eight unmerged microphone branches, none of
+which reached `blueline`. Read them before forming a new theory; the cheap
+hypotheses are already spent:
+
+- `mic-dmic-pads` — *program all DMIC interface pads*, *fix SLIM shared channel
+  numbers*, *replay complete blueline recorder state*
+- `mic-android-tx7-setup` — *apply blueline Android TX7 capture state*, *use
+  blueline handset mic topology*, *opt in blueline capture ADM identity*
+- `mic-dmic3-pad-config-g56` — *initialize DMIC3 pad function*, *prepare SLIM
+  stream before PCM start*, *encode DEF_ACT channel client field*
+- `mic-fix` — *wcd934x sysclk/sido lifecycle instrumentation*, *SLIMbus
+  CHAN_CTRL remove on stream disable*, *QCOM_PD_MAPPER built in*
+- also `mic-slim-prepare-order`, `mic-slim-prepare-order-g56`, `mic-unpin`
+
+The 2026-08-02 measurement adds one hard constraint to that pile: the decimator
+and the front-end mixer are **ruled out** (TX5 and TX7 both give peak `0`, and
+`MultiMedia2 Mixer SLIMBUS_0_TX` is on in both profiles). What separates working
+from silent is the Q6 voice session, which matches the sysclk/SLIM-lifecycle
+direction of `mic-fix` rather than the pad-configuration direction.