From 8f64539328ff06091490f3ea92d24f0047b38f18 Mon Sep 17 00:00:00 2001
From: Matiq <kokagishi@proton.me>
Date: Thu, 20 Aug 2026 16:25:00 +0300
Subject: [PATCH] =?UTF-8?q?P4:=20CRITICAL=20fix=20=E2=80=94=20level=20=3D?=
 =?UTF-8?q?=20am/res=20(not=20am*res);=20fc-resonance=20shape=20now=20corr?=
 =?UTF-8?q?ect?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

res=|2B/A| is minimal at band centre (not maximal). Model uses xv=log10(A_k/res_k)
=> level = am / res. Old am*res inverted the fc-response (cut more off-center).
Fixed; fc-scan (tone1kq, band fc 800..1200 via t1kq_only1_<fc>) now tracks the
reference: scale=42 mean|err| 2.46 dB, intact shape (deepest at fc==tone). Was
flat+inverted before. Remaining under-cut off-center -> FFT-conv smoothing/gaps.
---
 dsp/framed_model.cpp   |  6 ++++--
 handoff/NOTES_LEVEL.md | 18 ++++++++++++++++++
 2 files changed, 22 insertions(+), 2 deletions(-)

diff --git a/dsp/framed_model.cpp b/dsp/framed_model.cpp
index 8b55f6f..af2631f 100644
--- a/dsp/framed_model.cpp
+++ b/dsp/framed_model.cpp
@@ -134,10 +134,12 @@ void FramedDetector::processFrame(const std::complex<double>* spectrum, float* m
     for (size_t k = 0; k <= half; k++) mask[k] = 1.0f;
 
     for (size_t b = 0; b < bands_.size(); b++) {
-        // 1. level = twin |2B/A| x smoothed amp, scaled (0x9be0 buf*=scalar):
+        // 1. level = twin response x smoothed amp, scaled (0x9be0 buf*=scalar):
         //    mask *= (fVar30/0x1a0) * 0x540870 * 0x54088c
+        //    Level uses xv = log10(A_k / res_k): res=|2B/A| is MINIMAL at band
+        //    centre (a notch), so am/res is maximal there -> deepest cut at fc.
         for (size_t k = 0; k <= half; k++) {
-            double level = am_[k] * static_cast<double>(res_[b][k]) *
+            double level = am_[k] / std::max(static_cast<double>(res_[b][k]), 1e-12) *
                            static_cast<double>(bands_[b].level_scale);
             scratch[k] = static_cast<float>(level * (fVar30 / C_0x1a0) *
                                             C_0x540870 * C_0x54088c);
diff --git a/handoff/NOTES_LEVEL.md b/handoff/NOTES_LEVEL.md
index 29c8dc7..91c1c4f 100644
--- a/handoff/NOTES_LEVEL.md
+++ b/handoff/NOTES_LEVEL.md
@@ -905,3 +905,21 @@ Verified constants at raw=RVA: 0x24c3c58=0.001, 0x24c3e28=0.8, 0x24c4674=-0.7,
 
 ### framed_model.cpp: scale coeff now computed via prng_fvar30() (prng_state_=112),
    not a hardcoded constant. t1kq still -0.43 dB at level_scale=600 (unchanged).
+
+## ============ UPDATE 2026-08-20o: CRITICAL — level = am / res (not am * res) ============
+Found + fixed a sign/direction bug in framed_model.cpp level computation.
+
+res = |2B/A| (twin) is MINIMAL at band centre (0.117), NOT maximal. The model uses
+xv = log10(A_k / res_k), i.e. LEVEL = am / res. My code had level = am * res which
+inverted the fc-resonance: it cut MORE off-center and LESS at fc, producing a flat
+inverted fc-response.
+
+Fix: level = am_ / res_[b][k] (with 1e-12 floor). Now the fc-scan (tone1kq, band fc
+swept 800..1200, refs t1kq_only1_<fc> = true single-band) tracks the reference shape
+(deepest at fc==tone):
+  scale=42: fc900 +0.3, fc950 +1.3, fc1000 +0.9, fc1100 +0.2; mean|err| 2.46 dB,
+  max 5.3 dB (worst at edges fc800 +3.9, fc1200 +5.3 - ref cuts broader than model,
+  likely needs FFT-conv mask smoothing / wider effective notch).
+Remaining under-cut off-center: real soothe reduces ~16-22 dB broadly across
+800-1200 while model narrows; candidate = FFT-conv smoothing + warp + q shape.
+DEFAULT level_scale for am/res path ~= 35-45 (was 600 for the wrong am*res path).