前端: 抽公共层(dsp/common.py 共享内存协议)+ 32段EQ/总音量 DSP(dsp/eq.py, scipy sosfilt) + 本地控制面板(web/, 滑块+滚轮捕捉); 安装脚本接 30-collaplex-eq; pyright strict 0 error

README.md

@@ -37,6 +37,17 @@ filters2 = [ { type = bq_highshelf freq = 6000 gain = 0.0 } ]        # 高频架
 ★ 坑: `param_eq` 的端口名**带编号** —— `In 1` / `Out 1`(走 filters1)、`In 2` / `Out 2`(走 filters2)。
 写成 `:In` / `:Out` 会报 `unknown input port` 并让**整张图加载失败**(模块静默降级成直通, 节点都不建)。
 
+## 控制面板
+
+```sh
+bash 脚本/面板.sh            # http://127.0.0.1:8789
+```
+
+- 电平归一化(响度目标)、总音量、32 段 EQ **全部是滑块**:滚轮调节、Shift+滚轮微调(0.1 dB)、双击归零
+- EQ 是 1/3 倍频程 32 段(20 Hz … 25 kHz),每段 ±15 dB,**实时生效**(DSP 从共享内存读参数,不用重装)
+- 参数落点:`/dev/shm/collaplex-eq`(前端写、DSP 读,版本号保证不成半更新);响度目标写在
+  `~/.local/state/cinema-spatial/loudness.json`(归一化 DSP 每秒重读)
+
 ## 实测(2026-09-19)
 
 | 指标 | 数据 |

bin/collaplex-eq

@@ -0,0 +1,8 @@
+#!/bin/sh
+# Collaplex 32 段 EQ + 总音量 —— pipe 插件入口
+#
+# pipe 插件的 command 只支持「程序 + 一个参数」, 所以参数全写在这里;
+# 且路径必须是纯 ASCII(中文路径会静默失效)。
+# 用项目 venv 的解释器(带 scipy/numpy)。
+export CX_RATE=96000
+exec "__PY__" "__DSP__"

dsp/common.py

@@ -0,0 +1,100 @@
+"""Collaplex 效果 DSP 公共部分: 前端 <-> DSP 的共享内存协议。
+
+前端(web/server.py) 与 DSP 进程(pipe 插件载荷)通过这一块共享内存通信:
+
+    参数区: 32 段 EQ 增益(dB) + 总音量(dB)   <- 前端写, DSP 读
+    版本区: 参数版本号                        <- 前端写完参数后 +1
+    状态区: 输入/输出 RMS 与峰值(dB)          <- DSP 写, 前端读
+
+版本号最后写, DSP 只在版本号变化时重读参数, 所以前端一次改多个值也不会让
+DSP 读到半更新的数据。
+"""
+from __future__ import annotations
+
+import mmap
+import os
+import struct
+
+EQ_BANDS = 32
+EQ_PATH = "/dev/shm/collaplex-eq"
+
+# 1/3 倍频程中心频率(20 Hz 起, 32 段, 覆盖到 25 kHz)
+EQ_FREQS: list[float] = [
+    20.0, 25.0, 31.5, 40.0, 50.0, 63.0, 80.0, 100.0,
+    125.0, 160.0, 200.0, 250.0, 315.0, 400.0, 500.0, 630.0,
+    800.0, 1000.0, 1250.0, 1600.0, 2000.0, 2500.0, 3150.0, 4000.0,
+    5000.0, 6300.0, 8000.0, 10000.0, 12500.0, 16000.0, 20000.0, 25000.0,
+]
+
+GAIN_LIMIT = 15.0
+VOLUME_LIMIT = 12.0
+
+_DATA = struct.Struct("<33d")   # 32 段增益 + 总音量
+_VER = struct.Struct("<I")      # 参数版本号
+_STATE = struct.Struct("<6d")   # in_rms in_peak out_rms out_peak 推子(dB) 保留
+
+DATA_OFF = 0
+VER_OFF = _DATA.size
+STATE_OFF = _DATA.size + _VER.size
+SIZE = STATE_OFF + _STATE.size
+
+
+def open_store() -> mmap.mmap:
+    """打开(不存在则创建)共享内存映射。"""
+    fd = os.open(EQ_PATH, os.O_RDWR | os.O_CREAT, 0o666)
+    try:
+        if os.fstat(fd).st_size < SIZE:
+            os.ftruncate(fd, SIZE)
+        return mmap.mmap(fd, SIZE)
+    finally:
+        os.close(fd)
+
+
+def read_params(store: mmap.mmap) -> tuple[list[float], float, int]:
+    """读参数: (32 段增益 dB, 总音量 dB, 版本号)。"""
+    store.seek(DATA_OFF)
+    values = _DATA.unpack(store.read(_DATA.size))
+    store.seek(VER_OFF)
+    version = _VER.unpack(store.read(_VER.size))[0]
+    return list(values[:EQ_BANDS]), float(values[EQ_BANDS]), int(version)
+
+
+def write_params(store: mmap.mmap, gains: list[float], volume_db: float) -> int:
+    """写参数并让版本号 +1, 返回新版本号。"""
+    if len(gains) != EQ_BANDS:
+        raise ValueError("gains 长度必须是 %d" % EQ_BANDS)
+    store.seek(VER_OFF)
+    version = int(_VER.unpack(store.read(_VER.size))[0])
+    store.seek(DATA_OFF)
+    store.write(_DATA.pack(*[float(g) for g in gains], float(volume_db)))
+    version = (version + 1) & 0xFFFFFFFF
+    store.seek(VER_OFF)
+    store.write(_VER.pack(version))
+    store.flush()
+    return version
+
+
+def read_state(store: mmap.mmap) -> dict[str, float]:
+    """读状态(DSP 写的电平)。"""
+    store.seek(STATE_OFF)
+    in_rms, in_peak, out_rms, out_peak, gain_db, _spare = _STATE.unpack(store.read(_STATE.size))
+    return {
+        "in_rms": float(in_rms),
+        "in_peak": float(in_peak),
+        "out_rms": float(out_rms),
+        "out_peak": float(out_peak),
+        "gain_db": float(gain_db),
+    }
+
+
+def write_state(store: mmap.mmap, in_rms: float, in_peak: float, out_rms: float,
+                out_peak: float, gain_db: float) -> None:
+    """写状态(DSP 侧调用)。"""
+    store.seek(STATE_OFF)
+    store.write(_STATE.pack(float(in_rms), float(in_peak), float(out_rms),
+                            float(out_peak), float(gain_db), 0.0))
+
+
+def defaults() -> tuple[list[float], float]:
+    """默认参数: 全平直 + 0 dB 总音量。"""
+    return [0.0] * EQ_BANDS, 0.0

dsp/eq.py

@@ -0,0 +1,102 @@
+"""32 段 EQ + 总音量 —— pipe 插件载荷(单声道块进, 单声道块出)。
+
+参数实时来自共享内存(dsp/common.py): 32 段增益(dB) + 总音量(dB)。
+做法: 每段一个 peaking biquad, 32 段级联成二阶节(sos), 用 scipy.signal.sosfilt
+逐块滤波(C 实现, 保留滤波状态, 无额外延迟), 末级乘总音量。
+"""
+from __future__ import annotations
+
+import math
+import os
+import sys
+
+import numpy as np
+from scipy.signal import sosfilt
+
+sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
+
+import common  # noqa: E402
+
+RATE = int(os.environ.get("CX_RATE", "96000"))
+BLOCK_BYTES = 4096
+N = BLOCK_BYTES // 4            # 每块样本数(1024)
+Q = 4.318                       # 1/3 倍频程
+
+
+def design_peaking(freq: float, gain_db: float, fs: int, q: float) -> list[float]:
+    """RBJ cookbook 的 peaking biquad, 返回一行 sos [b0 b1 b2 1 a1 a2]。
+
+    gain_db = 0 时结果恰好是单位传递函数, 所以 32 段可以一直挂着(不必增删节点)。
+    """
+    amp = 10.0 ** (gain_db / 40.0)
+    w0 = 2.0 * math.pi * freq / fs
+    alpha = math.sin(w0) / (2.0 * q)
+    cos_w0 = math.cos(w0)
+    b0 = 1.0 + alpha * amp
+    b1 = -2.0 * cos_w0
+    b2 = 1.0 - alpha * amp
+    a0 = 1.0 + alpha / amp
+    a1 = -2.0 * cos_w0
+    a2 = 1.0 - alpha / amp
+    return [b0 / a0, b1 / a0, b2 / a0, 1.0, a1 / a0, a2 / a0]
+
+
+def build_sos(gains: list[float], fs: int) -> np.ndarray:
+    """32 段增益 -> sos 矩阵(32 x 6)。"""
+    return np.array([design_peaking(freq, gain, fs, Q)
+                     for freq, gain in zip(common.EQ_FREQS, gains)], dtype=np.float64)
+
+
+def level_db(block: np.ndarray) -> tuple[float, float]:
+    """返回 (RMS dBFS, 峰值 dBFS)。"""
+    if block.size == 0:
+        return -120.0, -120.0
+    rms = float(np.sqrt(np.mean(block * block)))
+    peak = float(np.max(np.abs(block)))
+    rms_db = 20.0 * math.log10(rms) if rms > 1e-9 else -120.0
+    peak_db = 20.0 * math.log10(peak) if peak > 1e-9 else -120.0
+    return rms_db, peak_db
+
+
+def main() -> None:
+    store = common.open_store()
+    stdin = sys.stdin.buffer
+    stdout = sys.stdout.buffer
+
+    version = -1
+    sos = build_sos(common.defaults()[0], RATE)
+    zi = np.zeros((sos.shape[0], 2), dtype=np.float64)
+    volume = 1.0
+    volume_db = 0.0
+
+    while True:
+        raw = stdin.read(BLOCK_BYTES)
+        if not raw:
+            break
+        if len(raw) % 4:
+            raw = raw[: len(raw) - (len(raw) % 4)]
+
+        gains, want_volume_db, new_version = common.read_params(store)
+        if new_version != version:
+            version = new_version
+            sos = build_sos(gains, RATE)
+            volume_db = want_volume_db
+            volume = 10.0 ** (volume_db / 20.0)
+
+        x = np.frombuffer(raw, dtype="<f4").astype(np.float64)
+        if x.size != N:
+            x = np.resize(x, N)
+
+        y, zi = sosfilt(sos, x, zi=zi)
+        out = y * volume
+
+        in_rms, in_peak = level_db(x)
+        out_rms, out_peak = level_db(np.clip(out, -1.0, 1.0))
+        common.write_state(store, in_rms, in_peak, out_rms, out_peak, volume_db)
+
+        stdout.write(np.clip(out, -1.0, 1.0).astype("<f4").tobytes())
+        stdout.flush()
+
+
+if __name__ == "__main__":
+    main()

pipewire/20-collaplex-hrtf.conf

@@ -63,7 +63,7 @@ context.modules = [
                 node.name      = collaplex_hrtf_out
                 audio.channels = 2
                 audio.position = [ FL FR ]
-                node.target    = "__TARGET__"
+                node.target    = "__EQ_IN__"
             }
         }
     }

pipewire/30-collaplex-eq.conf

@@ -0,0 +1,41 @@
+# Collaplex 32 段 EQ + 总音量 —— 本文件由 脚本/安装.sh 生成(占位符 __EQ_WRAPPER__ / __TARGET__)
+# 链路: ... -> HRTF 核心 -> [本图] 32 段 EQ + 总音量(pipe DSP) -> 数字输出
+context.modules = [
+    {   name = libpipewire-module-filter-chain
+        flags = [ nofail ]
+        args = {
+            node.description = "Collaplex EQ + 总音量"
+            media.name       = "Collaplex EQ"
+            filter.graph = {
+                nodes = [
+                    { type = builtin label = copy name = c1L }
+                    { type = builtin label = copy name = c1R }
+                    { type = builtin label = pipe name = eqpL config = { command = "__EQ_WRAPPER__" } }
+                    { type = builtin label = pipe name = eqpR config = { command = "__EQ_WRAPPER__" } }
+                    { type = builtin label = copy name = c2L }
+                    { type = builtin label = copy name = c2R }
+                ]
+                links = [
+                    { output = "c1L:Out" input = "eqpL:In" }
+                    { output = "c1R:Out" input = "eqpR:In" }
+                    { output = "eqpL:Out" input = "c2L:In" }
+                    { output = "eqpR:Out" input = "c2R:In" }
+                ]
+                inputs  = [ "c1L:In" "c1R:In" ]
+                outputs = [ "c2L:Out" "c2R:Out" ]
+            }
+            capture.props = {
+                media.class    = Audio/Sink
+                node.name      = collaplex_eq_in
+                audio.channels = 2
+                audio.position = [ FL FR ]
+            }
+            playback.props = {
+                node.name      = collaplex_eq_out
+                audio.channels = 2
+                audio.position = [ FL FR ]
+                node.target    = "__TARGET__"
+            }
+        }
+    }
+]

pyrightconfig.json

@@ -0,0 +1,13 @@
+{
+    "include": ["dsp", "web"],
+    "exclude": ["dsp/loudness_norm.py", "dsp/状态.py"],
+    "extraPaths": ["dsp"],
+    "venvPath": ".",
+    "venv": ".venv",
+    "typeCheckingMode": "strict",
+    "pythonVersion": "3.11",
+    "reportMissingTypeStubs": false,
+    "reportUnknownMemberType": false,
+    "reportUnknownArgumentType": false,
+    "reportUnknownVariableType": false
+}

web/index.html

@@ -0,0 +1,250 @@
+<!DOCTYPE html>
+<html lang="zh-CN">
+<head>
+<meta charset="utf-8">
+<meta name="viewport" content="width=device-width, initial-scale=1">
+<title>Collaplex 音效</title>
+<style>
+:root{
+  --bg:#0f0f0f; --card:#161616; --card-hi:#1c1c1c; --bd:#262626; --bd-hi:#343434;
+  --fg:#e8e8e8; --mut:#8a8a8a; --dim:#666; --ok:#4ade80; --ok-bg:#12211a; --ok-bd:#166534;
+  --r:12px;
+}
+*{box-sizing:border-box}
+html,body{margin:0;background:var(--bg);color:var(--fg);
+  font-family:"JetBrains Mono",ui-monospace,SFMono-Regular,Menlo,monospace;font-size:13px}
+.wrap{max-width:1280px;margin:0 auto;padding:18px 20px 40px}
+header{display:flex;align-items:baseline;gap:16px;border-bottom:1px solid var(--bd);padding-bottom:12px}
+h1{font-size:15px;font-weight:600;margin:0;letter-spacing:.02em}
+.hint{color:var(--dim);font-size:11px}
+.meter{margin-left:auto;display:flex;gap:18px;align-items:baseline;color:var(--mut);font-size:11px}
+.meter b{color:var(--ok);font-weight:500}
+.rows{display:grid;grid-template-columns:1fr 1fr;gap:14px;margin:16px 0}
+.card{background:var(--card);border:1px solid var(--bd);border-radius:var(--r);padding:14px 16px}
+.card h2{margin:0 0 12px;font-size:12px;font-weight:500;color:var(--mut);letter-spacing:.06em}
+.slider-row{display:flex;align-items:center;gap:14px}
+.val{min-width:78px;text-align:right;color:var(--ok);font-variant-numeric:tabular-nums}
+input[type=range]{-webkit-appearance:none;appearance:none;height:22px;background:transparent;outline:none;flex:1}
+input[type=range]::-webkit-slider-runnable-track{height:4px;background:var(--bd-hi);border-radius:2px}
+input[type=range]::-webkit-slider-thumb{-webkit-appearance:none;width:16px;height:16px;margin-top:-6px;
+  border-radius:50%;background:var(--ok);border:0;cursor:grab;box-shadow:0 0 0 3px rgba(74,222,128,.15)}
+input[type=range]:active::-webkit-slider-thumb{cursor:grabbing}
+input[type=range]::-moz-range-track{height:4px;background:var(--bd-hi);border-radius:2px}
+input[type=range]::-moz-range-thumb{width:16px;height:16px;border:0;border-radius:50%;background:var(--ok)}
+.eqcard{background:var(--card);border:1px solid var(--bd);border-radius:var(--r);padding:16px 10px 10px}
+.eq{display:grid;grid-template-columns:repeat(32,1fr);gap:2px;height:230px}
+.band{display:flex;flex-direction:column;align-items:center;gap:4px}
+.band input[type=range]{writing-mode:vertical-lr;direction:rtl;width:22px;height:190px;flex:none}
+.band .db{font-size:9px;color:var(--dim);font-variant-numeric:tabular-nums;height:11px}
+.freqs{display:grid;grid-template-columns:repeat(32,1fr);gap:2px;margin-top:6px}
+.freqs span{font-size:9px;color:var(--dim);text-align:center;transform:rotate(-60deg);transform-origin:center;
+  white-space:nowrap;height:26px}
+.actions{display:flex;gap:10px;margin-top:14px;align-items:center}
+button{background:var(--card-hi);color:var(--fg);border:1px solid var(--bd-hi);border-radius:8px;
+  padding:7px 14px;font:inherit;font-size:12px;cursor:pointer}
+button:hover{border-color:var(--ok);color:var(--ok)}
+button.lit{background:var(--ok-bg);border-color:var(--ok-bd);color:#bbf7d0}
+.toggle{display:flex;align-items:center;gap:7px;color:var(--mut);font-size:12px;cursor:pointer}
+.state{margin-left:auto;color:var(--dim);font-size:11px}
+</style>
+</head>
+<body>
+<div class="wrap">
+  <header>
+    <h1>Collaplex 音效</h1>
+    <span class="hint">滚轮调滑块 · Shift+滚轮微调 · 双击归零</span>
+    <div class="meter">
+      <span>入 <b id="lvl-in">-∞</b> dBFS</span>
+      <span>出 <b id="lvl-out">-∞</b> dBFS</span>
+    </div>
+  </header>
+
+  <div class="rows">
+    <div class="card">
+      <h2>电平归一化（响度目标）</h2>
+      <div class="slider-row">
+        <input type="range" id="loud" min="-40" max="-5" step="0.5" value="-16">
+        <span class="val" id="loud-val">-16.0 dB</span>
+      </div>
+      <div class="actions">
+        <label class="toggle"><input type="checkbox" id="loud-on" checked> 启用</label>
+      </div>
+    </div>
+
+    <div class="card">
+      <h2>总音量放大</h2>
+      <div class="slider-row">
+        <input type="range" id="vol" min="-12" max="12" step="0.5" value="6">
+        <span class="val" id="vol-val">+6.0 dB</span>
+      </div>
+      <div class="actions">
+        <button id="reset-all">全部归零</button>
+        <button id="reset-eq">EQ 归零</button>
+        <span class="state" id="state">—</span>
+      </div>
+    </div>
+  </div>
+
+  <div class="eqcard">
+    <div class="eq" id="eq"></div>
+    <div class="freqs" id="freqs"></div>
+  </div>
+</div>
+
+<script>
+const $ = (id) => document.getElementById(id);
+let gains = new Array(32).fill(0);
+let freqs = [];
+let pending = null;
+
+function fmt(v, plus) {
+  const s = (v > 0 && plus ? "+" : "") + v.toFixed(1);
+  return s + " dB";
+}
+
+// 滑块 + 滚轮捕捉
+function bindWheel(el, onChange) {
+  el.addEventListener("wheel", (e) => {
+    e.preventDefault();
+    const step = e.shiftKey ? 0.1 : 0.5;
+    const dir = e.deltaY < 0 ? 1 : -1;
+    const next = parseFloat(el.value) + dir * step;
+    el.value = Math.min(parseFloat(el.max), Math.max(parseFloat(el.min), next));
+    onChange(parseFloat(el.value));
+  }, { passive: false });
+}
+
+function throttled(fn, ms) {
+  let last = 0, timer = null, lastArgs = null;
+  return function (...args) {
+    lastArgs = args;
+    const now = Date.now();
+    const run = () => { last = Date.now(); timer = null; fn(...lastArgs); };
+    if (now - last >= ms) run();
+    else if (!timer) timer = setTimeout(run, ms - (now - last));
+  };
+}
+
+const pushEq = throttled(async () => {
+  setState("写入 EQ…");
+  try {
+    const r = await fetch("/api/eq", {
+      method: "POST", headers: { "Content-Type": "application/json" },
+      body: JSON.stringify({ gains })
+    });
+    setState(r.ok ? "EQ 已生效" : "EQ 写入失败");
+  } catch (err) { setState("EQ 写入失败"); }
+}, 90);
+
+const pushVol = throttled(async (db) => {
+  setState("写入总音量…");
+  try {
+    const r = await fetch("/api/volume", {
+      method: "POST", headers: { "Content-Type": "application/json" },
+      body: JSON.stringify({ db })
+    });
+    setState(r.ok ? "总音量已生效" : "写入失败");
+  } catch (err) { setState("写入失败"); }
+}, 90);
+
+const pushLoud = throttled(async (db, on) => {
+  setState("写入归一化…");
+  try {
+    const r = await fetch("/api/loudness", {
+      method: "POST", headers: { "Content-Type": "application/json" },
+      body: JSON.stringify({ target_db: db, on })
+    });
+    setState(r.ok ? "归一化已生效" : "写入失败");
+  } catch (err) { setState("写入失败"); }
+}, 90);
+
+function setState(text) { $("state").textContent = text; }
+
+// 32 段 EQ 滑块
+function buildEq() {
+  const box = $("eq"), fbox = $("freqs");
+  box.innerHTML = ""; fbox.innerHTML = "";
+  for (let i = 0; i < 32; i++) {
+    const band = document.createElement("div");
+    band.className = "band";
+    const db = document.createElement("div");
+    db.className = "db"; db.id = "db" + i; db.textContent = "0.0";
+    const sl = document.createElement("input");
+    sl.type = "range"; sl.min = "-15"; sl.max = "15"; sl.step = "0.5"; sl.value = "0";
+    sl.dataset.index = i;
+    sl.addEventListener("input", () => {
+      gains[i] = parseFloat(sl.value);
+      $("db" + i).textContent = gains[i].toFixed(1);
+      pushEq();
+    });
+    bindWheel(sl, (v) => {
+      gains[i] = v; $("db" + i).textContent = v.toFixed(1); pushEq();
+    });
+    sl.addEventListener("dblclick", () => {
+      sl.value = "0"; gains[i] = 0; $("db" + i).textContent = "0.0"; pushEq();
+    });
+    band.appendChild(db); band.appendChild(sl);
+    box.appendChild(band);
+    const f = document.createElement("span");
+    f.textContent = freqs[i] ? (freqs[i] >= 1000 ? (freqs[i] / 1000) + "k" : String(freqs[i])) : "";
+    fbox.appendChild(f);
+  }
+}
+
+// 同步 DSP 实时电平
+async function poll() {
+  try {
+    const s = await (await fetch("/api/state")).json();
+    $("lvl-in").textContent = (s.levels.in_rms <= -120 ? "-∞" : s.levels.in_rms.toFixed(1));
+    $("lvl-out").textContent = (s.levels.out_rms <= -120 ? "-∞" : s.levels.out_rms.toFixed(1));
+  } catch (err) { /* 忽略 */ }
+}
+
+async function init() {
+  const s = await (await fetch("/api/state")).json();
+  freqs = s.freqs;
+  gains = s.gains.slice();
+  buildEq();
+  for (let i = 0; i < 32; i++) {
+    const sl = document.querySelector('.band input[data-index="' + i + '"]');
+    sl.value = String(gains[i]);
+    $("db" + i).textContent = gains[i].toFixed(1);
+  }
+  const loud = s.loudness || { target_db: -16, on: true };
+  $("loud").value = loud.target_db; $("loud-val").textContent = fmt(loud.target_db, false);
+  $("loud-on").checked = loud.on !== false;
+  $("vol").value = s.volume_db; $("vol-val").textContent = fmt(s.volume_db, true);
+
+  $("loud").addEventListener("input", (e) => {
+    const v = parseFloat(e.target.value);
+    $("loud-val").textContent = fmt(v, false);
+    pushLoud(v, $("loud-on").checked);
+  });
+  bindWheel($("loud"), (v) => { $("loud-val").textContent = fmt(v, false); pushLoud(v, $("loud-on").checked); });
+  $("loud-on").addEventListener("change", (e) => pushLoud(parseFloat($("loud").value), e.target.checked));
+
+  $("vol").addEventListener("input", (e) => {
+    const v = parseFloat(e.target.value);
+    $("vol-val").textContent = fmt(v, true);
+    pushVol(v);
+  });
+  bindWheel($("vol"), (v) => { $("vol-val").textContent = fmt(v, true); pushVol(v); });
+
+  $("reset-eq").addEventListener("click", () => {
+    gains = new Array(32).fill(0);
+    for (let i = 0; i < 32; i++) { document.querySelector('.band input[data-index="' + i + '"]').value = "0"; $("db" + i).textContent = "0.0"; }
+    pushEq();
+  });
+  $("reset-all").addEventListener("click", () => {
+    $("reset-eq").click();
+    $("vol").value = "0"; $("vol-val").textContent = fmt(0, true); pushVol(0);
+  });
+
+  setInterval(poll, 700);
+  poll();
+}
+
+init();
+</script>
+</body>
+</html>

web/server.py

@@ -0,0 +1,154 @@
+"""Collaplex 音效控制面板 —— 本地服务(纯标准库, 无第三方依赖)。
+
+    python3 web/server.py [端口]        默认 8789
+
+页面:      /
+接口:
+    GET  /api/state              参数 + DSP 实时电平
+    POST /api/eq                 {"gains": [32 个 dB]}
+    POST /api/volume             {"db": 数值}
+    POST /api/loudness           {"target_db": 数值, "on": true/false}
+"""
+from __future__ import annotations
+
+import json
+import os
+import sys
+from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
+from typing import cast
+
+HERE = os.path.dirname(os.path.abspath(__file__))
+sys.path.insert(0, os.path.join(HERE, "..", "dsp"))
+
+import common  # noqa: E402
+
+LOUDNESS_JSON = os.path.expanduser("~/.local/state/cinema-spatial/loudness.json")
+UI = os.path.join(HERE, "index.html")
+DEFAULT_PORT = 8789
+
+
+def _as_float(value: object, fallback: float = 0.0) -> float:
+    """把 JSON 取到的值转 float(来源类型不安全, 这里只做值校验)。"""
+    if isinstance(value, bool):
+        return fallback
+    if isinstance(value, (int, float)):
+        return float(value)
+    if isinstance(value, str):
+        try:
+            return float(value)
+        except ValueError:
+            return fallback
+    return fallback
+
+
+def read_loudness() -> dict[str, object]:
+    """读响度归一化设置(归一化 DSP 每秒重读这个文件)。"""
+    try:
+        with open(LOUDNESS_JSON, "r", encoding="utf-8") as fh:
+            data = json.load(fh)
+        if isinstance(data, dict):
+            return cast(dict[str, object], data)
+    except (OSError, ValueError):
+        pass
+    return {"on": True, "target_db": -16.0}
+
+
+def write_loudness(target_db: float, on: bool) -> None:
+    os.makedirs(os.path.dirname(LOUDNESS_JSON), exist_ok=True)
+    with open(LOUDNESS_JSON, "w", encoding="utf-8") as fh:
+        json.dump({"on": bool(on), "target_db": float(target_db)}, fh)
+
+
+class Handler(BaseHTTPRequestHandler):
+    store = common.open_store()
+    protocol_version = "HTTP/1.1"
+
+    def log_message(self, format: str, *args: object) -> None:
+        pass
+
+    def _send(self, body: bytes, ctype: str, code: int = 200) -> None:
+        self.send_response(code)
+        self.send_header("Content-Type", ctype)
+        self.send_header("Content-Length", str(len(body)))
+        self.send_header("Cache-Control", "no-store")
+        self.end_headers()
+        self.wfile.write(body)
+
+    def _send_json(self, obj: object, code: int = 200) -> None:
+        self._send(json.dumps(obj, ensure_ascii=False).encode("utf-8"),
+                   "application/json; charset=utf-8", code)
+
+    def _read_json(self) -> dict[str, object]:
+        length = int(self.headers.get("Content-Length") or 0)
+        raw = self.rfile.read(length) if length else b"{}"
+        try:
+            data = json.loads(raw.decode("utf-8"))
+        except ValueError:
+            return {}
+        return cast(dict[str, object], data) if isinstance(data, dict) else {}
+
+    def do_GET(self) -> None:
+        if self.path in ("/", "/index.html"):
+            try:
+                with open(UI, "rb") as fh:
+                    self._send(fh.read(), "text/html; charset=utf-8")
+            except OSError:
+                self._send("面板文件缺失".encode("utf-8"), "text/plain; charset=utf-8", 500)
+            return
+        if self.path == "/api/state":
+            gains, volume_db, _ver = common.read_params(self.store)
+            loudness = read_loudness()
+            self._send_json({
+                "freqs": common.EQ_FREQS,
+                "gains": gains,
+                "volume_db": volume_db,
+                "gain_limit": common.GAIN_LIMIT,
+                "volume_limit": common.VOLUME_LIMIT,
+                "loudness": {
+                    "target_db": _as_float(loudness.get("target_db", -16.0), -16.0),
+                    "on": bool(loudness.get("on", True)),
+                },
+                "levels": common.read_state(self.store),
+            })
+            return
+        self._send_json({"error": "not found"}, 404)
+
+    def do_POST(self) -> None:
+        data = self._read_json()
+        if self.path == "/api/eq":
+            raw_gains = data.get("gains")
+            if not isinstance(raw_gains, list) or len(raw_gains) != common.EQ_BANDS:
+                self._send_json({"error": "gains 必须是 %d 个数" % common.EQ_BANDS}, 400)
+                return
+            items = cast(list[object], raw_gains)
+            gains = [max(-common.GAIN_LIMIT, min(common.GAIN_LIMIT, _as_float(item))) for item in items]
+            _g, volume_db, _v = common.read_params(self.store)
+            version = common.write_params(self.store, gains, volume_db)
+            self._send_json({"ok": True, "version": version})
+            return
+        if self.path == "/api/volume":
+            volume_db = max(-common.VOLUME_LIMIT,
+                            min(common.VOLUME_LIMIT, _as_float(data.get("db", 0.0))))
+            gains, _old, _v = common.read_params(self.store)
+            version = common.write_params(self.store, gains, volume_db)
+            self._send_json({"ok": True, "version": version, "volume_db": volume_db})
+            return
+        if self.path == "/api/loudness":
+            current = read_loudness()
+            target_db = _as_float(data.get("target_db", current.get("target_db", -16.0)), -16.0)
+            on = bool(data.get("on", current.get("on", True)))
+            write_loudness(max(-40.0, min(-5.0, target_db)), on)
+            self._send_json({"ok": True, "loudness": read_loudness()})
+            return
+        self._send_json({"error": "not found"}, 404)
+
+
+def main() -> None:
+    port = int(sys.argv[1]) if len(sys.argv) > 1 else DEFAULT_PORT
+    server = ThreadingHTTPServer(("127.0.0.1", port), Handler)
+    print("Collaplex 音效控制面板: http://127.0.0.1:%d" % port, flush=True)
+    server.serve_forever()
+
+
+if __name__ == "__main__":
+    main()

脚本/安装.sh

@@ -19,6 +19,7 @@ WET="${WET:-0.3}"
 VOL_DB="${VOL_DB:-6}"
 VOL="$(/usr/bin/python3 -c "print(f'{10**(${VOL_DB}/20):.4f}')")"
 HRTF_IN="collaplex_hrtf_in"
+EQ_IN="collaplex_eq_in"
 
 # ---- 1. 探测数字输出 ----
 TARGET=$(pw-dump | /usr/bin/python3 -c '
@@ -39,23 +40,28 @@ echo "数字输出: $TARGET"
 # ---- 2. 部署 pipe 插件入口(必须 ASCII 路径) ----
 mkdir -p "$(dirname "$WRAPPER")"
 sed -e "s|__DSP__|$ROOT/dsp/loudness_norm.py|" "$ROOT/bin/collaplex-loudness-norm" > "$WRAPPER"
-chmod 755 "$WRAPPER"
-echo "已装入口: $WRAPPER"
+sed -e "s|__PY__|$ROOT/.venv/bin/python|" -e "s|__DSP__|$ROOT/dsp/eq.py|" \
+    "$ROOT/bin/collaplex-eq" > "$HOME/.local/bin/collaplex-eq"
+chmod 755 "$WRAPPER" "$HOME/.local/bin/collaplex-eq"
+echo "已装入口: $WRAPPER / $HOME/.local/bin/collaplex-eq"
 
 # ---- 3. 渲染两份配置 ----
 mkdir -p "$CONF_DIR"
 sed -e "s|__WRAPPER__|$WRAPPER|" -e "s|__HRTF_IN__|$HRTF_IN|" \
     "$ROOT/pipewire/10-collaplex-vsink.conf" > "$CONF_DIR/10-collaplex-vsink.conf"
 sed -e "s|__HRIR__|$ROOT/hrir|" -e "s|__REVERB__|$ROOT/reverb/房间混响IR-96k.wav|" \
-    -e "s|__TARGET__|$TARGET|" -e "s|__WET__|$WET|" -e "s|__VOL__|$VOL|" \
+    -e "s|__EQ_IN__|$EQ_IN|" -e "s|__WET__|$WET|" -e "s|__VOL__|$VOL|" \
     "$ROOT/pipewire/20-collaplex-hrtf.conf" > "$CONF_DIR/20-collaplex-hrtf.conf"
-echo "已装配置: 10-collaplex-vsink.conf / 20-collaplex-hrtf.conf(混响湿量 $WET, 总音量 +${VOL_DB}dB)"
+sed -e "s|__EQ_WRAPPER__|$HOME/.local/bin/collaplex-eq|" -e "s|__TARGET__|$TARGET|" \
+    "$ROOT/pipewire/30-collaplex-eq.conf" > "$CONF_DIR/30-collaplex-eq.conf"
+echo "已装配置: 10-collaplex-vsink.conf / 20-collaplex-hrtf.conf(湿量 $WET, 总音量 +${VOL_DB}dB) / 30-collaplex-eq.conf"
 
 # ---- 4. 重启前清掉残留 DSP(占住共享槽位会让新实例起不来) ----
 for p in /proc/[0-9]*; do
     c=$(tr '\0' ' ' < "$p/cmdline" 2>/dev/null) || continue
     case "$c" in
         "/usr/bin/python3 "*dsp/loudness_norm.py*) kill -9 "${p#/proc/}" 2>/dev/null || true;;
+        *dsp/eq.py*) kill -9 "${p#/proc/}" 2>/dev/null || true;;
     esac
 done
 sleep 1
@@ -72,5 +78,11 @@ for p in /proc/[0-9]*; do
     c=$(tr '\0' ' ' < "$p/cmdline" 2>/dev/null) || continue
     case "$c" in "/usr/bin/python3 "*dsp/loudness_norm.py*) n=$((n+1));; esac
 done
-echo "DSP 进程数: $n (应为 2)"
+echo "DSP 进程数: $n (归一化, 应为 2)"
+m=0
+for p in /proc/[0-9]*; do
+    c=$(tr '\0' ' ' < "$p/cmdline" 2>/dev/null) || continue
+    case "$c" in */dsp/eq.py*) m=$((m+1));; esac
+done
+echo "EQ 进程数: $m (应为 2)"
 echo -n "默认输出: "; wpctl inspect @DEFAULT_SINK@ 2>/dev/null | grep -m1 "node.name" | sed 's/.*= //'

脚本/面板.sh

@@ -0,0 +1,4 @@
+#!/bin/bash
+# Collaplex 音效控制面板 —— 起本地服务(默认 8789)
+cd "$(dirname "$0")/.." || exit 1
+exec .venv/bin/python web/server.py "${1:-8789}"