audio: per-slot postGain for parallel-branch loudness leveling (#58)

Adds an optional per-slot output gain (ProcessorSlot.postGain, default 1.0 = no-op) applied in SignalChain::runSlot after processBlock + pan, plumbed through setPostGain / IPC / preload / loadPreset / getChainState (mirroring setPan). Gives each parallel branch its own loudness trim.

Review hardening (multi-angle + Codex): savePreset() now serializes postGain (it was read back but never written → save/load dropped it); setPostGain() rejects non-finite input (NaN would poison the buffer); new signalchain_postgain_test.cpp asserts gain scaling, NaN rejection, and serialization.

Verified locally: build:audio clean; signalchain_postgain_test 5/5 pass. (Org CI runners failed to start — infra/billing, unrelated to the change; changes are platform-neutral C++.)

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

src/audio/SignalChain.cpp

@@ -1,5 +1,6 @@
 #include "SignalChain.h"
 #include "Sandbox/SandboxedProcessor.h"
+#include <cmath>   // std::isfinite (postGain sanitisation)
 
 #if ! JUCE_WINDOWS
  #include <csetjmp>
@@ -287,6 +288,8 @@ void SignalChain::process(juce::AudioBuffer<float>& buffer, juce::MidiBuffer& mi
                 slotMidi.addEvent(drained[i].msg, 0);
         invokePlugin(*slot, [&](juce::AudioProcessor& p) { p.processBlock(buf, slotMidi); });
         applyPan(buf, numSamples, slot->pan);
+        if (slot->postGain != 1.0f)
+            buf.applyGain(0, numSamples, slot->postGain);
     };
 
     // Fast path: no parallel branch → plain serial chain. Behaviour is unchanged
@@ -548,6 +551,18 @@ void SignalChain::setPan(int slotId, float pan)
     if (idx >= 0) slots[idx]->pan = juce::jlimit(-1.0f, 1.0f, pan);
 }
 
+void SignalChain::setPostGain(int slotId, float gain)
+{
+    // Reject non-finite input: juce::jlimit passes NaN through unchanged (both
+    // of its comparisons are false for NaN), and a NaN gain would then multiply
+    // the slot buffer to NaN and poison the whole chain until the slot rebuilds.
+    if (! std::isfinite(gain)) return;
+    const juce::ScopedLock sl(lock);
+    int idx = findSlotIndex(slotId);
+    // Clamp to [0, 16] — a 0..+24 dB linear ceiling for the per-slot trim.
+    if (idx >= 0) slots[idx]->postGain = juce::jlimit(0.0f, 16.0f, gain);
+}
+
 void SignalChain::setBranch(int slotId, int branch)
 {
     const juce::ScopedLock sl(lock);
@@ -674,6 +689,9 @@ juce::String SignalChain::savePreset() const
         if (slot->pan != 0.0f)       slotObj->setProperty("pan", slot->pan);
         if (slot->branch != 0)       slotObj->setProperty("branch", slot->branch);
         if (slot->branchSrc != 0)    slotObj->setProperty("branchSrc", slot->branchSrc);
+        // Per-slot output trim (loudness leveling). LoadPresetWorker reads this
+        // back, so it must be written here or a save/load round-trip drops it.
+        if (slot->postGain != 1.0f)  slotObj->setProperty("postGain", slot->postGain);
 
         // Save processor state as base64
         auto state = slot->getState();