feat: 生成语音v2版本，兼容多音字问题

packages/voice-production/data/char/char_introduce_v2.json

@@ -48,7 +48,7 @@
     ],
     "于": ["yú", null, "于，于是的于"],
     "亏": ["kuī", null, "亏，亏损的亏"],
-    "工": ["ɡōnɡ", null, "工，工人的人"],
+    "工": ["ɡōnɡ", null, "工，工人的工"],
     "土": ["tǔ", null, "土，土地的土"],
     "士": ["shì", null, "士，士气的士"],
     "才": ["cái", null, "才，才能的才"],
@@ -957,7 +957,6 @@
     "论": [
         ["lùn", "讨论", "论，讨论的论"],
         ["lún", "经典名", "论，论语的论"]
-        // TODO
     ],
     "讼": ["sònɡ", null, "讼，诉讼的讼"],
     "农": ["nónɡ", null, "农，农业的农"],
@@ -1283,8 +1282,8 @@
     "吹": ["chuī", null, "吹，吹气的吹"],
     "呜": ["wū", null, "呜，呜咽的呜"],
     "吭": [
-        ["háng", "喉咙", "吭，吭声，喉咙的声音"],
-        ["kēng", "出声", "吭，吭声的吭"]
+        ["kēng", "出声", "吭，吭声的吭"],
+        ["háng", "喉咙", "吭，吭声，喉咙的声音"]
     ],
     "吧": [
         ["bā", "酒吧", "吧，吧吧，语气词"],
@@ -1293,8 +1292,8 @@
     "邑": ["yì", null, "邑，城邑，城市"],
     "吼": ["hǒu", null, "吼，吼叫，大声叫喊"],
     "囤": [
-        ["dùn", "储存", "囤，囤积，储存"],
-        ["tún", "存粮的器物", "囤，粮囤的囤"]
+        ["tún", "存粮的器物", "囤，粮囤的囤"],
+        ["dùn", "储存", "囤，囤积，储存"]
     ],
     "别": [
         ["bié", "分别", "别，别扭，不顺畅"],
@@ -1343,8 +1342,8 @@
         ["tián", "耕作", "佃，佃作的佃"]
     ],
     "作": [
-        ["zuō", "手工作坊", "作，作业，作品的作"],
-        ["zuò", "创作", "作，工作的作"]
+        ["zuò", "创作", "作，工作的作"],
+        ["zuō", "手工作坊", "作，作业，作品的作"]
     ],
     "伯": [
         ["bó", "伯母", "伯，伯父，伯母"],
@@ -1403,7 +1402,7 @@
         ["qiū", "古国名", "龟，龟兹的龟"]
     ],
     "甸": [
-        ["diàn", "郊外", "甸，甸甸的甸"],
+        ["diàn", "郊外", "甸，沉甸甸的甸"],
         ["tián", "古同'田'", "甸，甸田的甸"],
         ["shèng", "古读", "甸，甸师的甸"]
     ],
@@ -1692,8 +1691,8 @@
     "抬": ["tái", null, "抬，抬举，抬起"],
     "拇": ["mǔ", null, "拇，拇指，大拇指"],
     "拗": [
-        ["ǎo", "折断", "拗，拗口，弯曲"],
         ["ào", "别扭", "拗，拗口的拗"],
+        ["ǎo", "折断", "拗，拗口，弯曲"],
         ["niù", "固执", "拗，执拗的拗"]
     ],
     "其": [
@@ -1714,7 +1713,7 @@
     ],
     "茂": ["mào", null, "茂，茂盛，繁茂"],
     "苹": [
-        ["píng", "水果", "苹，浮萍，水草"],
+        ["píng", "水果", "苹，苹果的苹"],
         ["pēng", "草名", "苹，苹萦的苹"]
     ],
     "苗": ["miáo", null, "苗，幼苗，禾苗"],
@@ -1726,8 +1725,8 @@
     "直": ["zhí", null, "直，正直的直"],
     "茁": ["zhuó", null, "茁，茁壮的茁"],
     "茄": [
-        ["jiā", "烟草", "茄，茄子的茄"],
-        ["qié", "蔬菜", "茄，茄子的茄"]
+        ["qié", "蔬菜", "茄，茄子的茄"],
+        ["jiā", "烟草", "茄，雪茄的茄"]
     ],
     "茎": ["jīnɡ", null, "茎，茎叶的茎"],
     "苔": [
@@ -1759,8 +1758,8 @@
     "述": ["shù", null, "述，叙述的述"],
     "枕": ["zhěn", null, "枕，枕头的枕"],
     "丧": [
-        ["sāng", "丧葬", "丧，丧事的丧"],
-        ["sàng", "失去", "丧，丧失的丧"]
+        ["sàng", "失去", "丧，丧失的丧"],
+        ["sāng", "丧葬", "丧，丧事的丧"]
     ],
     "或": ["huò", null, "或，或许的或"],
     "画": ["huà", null, "画，画画的画"],
@@ -1869,7 +1868,7 @@
     "岸": ["àn", null, "岸，河岸的岸"],
     "岩": ["yán", null, "岩，岩石，地壳的岩石"],
     "帖": [
-        ["tiè", "法帖", "帖，帖子，网络上的帖子"],
+        ["tiè", "法帖", "帖，字帖的帖"],
         ["tiě", "帖子", "帖，请帖的帖"],
         ["tiē", "顺从", "帖，服帖的帖"]
     ],
@@ -1946,9 +1945,9 @@
     "依": ["yī", null, "依，依靠的依"],
     "卑": ["bēi", null, "卑，卑微的卑"],
     "的": [
+        ["de", "助词", "的，我的的的"],
         ["dí", "确实", "的，的确的的"],
-        ["dì", "靶心", "的，目的的的"],
-        ["de", "助词", "的，我的的的"]
+        ["dì", "靶心", "的，目的的的"]
     ],
     "迫": [
         ["pò", "逼迫", "迫，压迫的迫"],
@@ -1968,8 +1967,8 @@
     ],
     "金": ["jīn", null, "金，金子的金"],
     "刹": [
-        ["chà", "瞬间", "刹，刹车的刹"],
-        ["shā", "制动", "刹，刹车的刹"]
+        ["shā", "制动", "刹，刹车的刹"],
+        ["chà", "瞬间", "刹，刹车的刹"]
     ],
     "命": ["mìnɡ", null, "命，命运的命"],
     "肴": ["yáo", null, "肴，菜肴的肴"],
@@ -2051,8 +2050,8 @@
     "刻": ["kè", null, "刻，雕刻的刻"],
     "育": ["yù", null, "育，教育的育"],
     "氓": [
-        ["méng", "无赖", "氓，流氓的氓"],
-        ["máng", "古代百姓", "氓，氓隶的氓"]
+        ["máng", "无赖", "氓，流氓的氓"],
+        ["méng", "古代百姓", "氓，氓隶的氓"]
     ],
     "闸": ["zhá", null, "闸，闸门的闸"],
     "闹": ["nào", null, "闹，热闹的闹"],
@@ -2622,10 +2621,7 @@
     "贷": ["dài", null, "贷，贷款，借入或借出的款项"],
     "顺": ["shùn", null, "顺，顺利，顺畅"],
     "修": ["xiū", null, "修，修理，修补"],
-    "俏": [
-        ["qiào", "好看", "俏，俏皮，活泼"],
-        ["xiào", "古同'肖'", "俏，俏皮的俏"]
-    ],
+    "俏": [["qiào", "好看", "俏，俏皮，活泼"]],
     "保": ["bǎo", null, "保，保证，担保"],
     "促": ["cù", null, "促，促进，推动"],
     "俄": ["é", null, "俄，俄罗斯，国家名"],
@@ -2634,10 +2630,7 @@
     "俭": ["jiǎn", null, "俭，节俭的俭"],
     "俗": ["sú", null, "俗，风俗的俗"],
     "俘": ["fú", null, "俘，俘虏的俘"],
-    "信": [
-        ["xìn", "书信", "信，信任的信"],
-        ["shēn", "古同'伸'", "信，信阳的信"]
-    ],
+    "信": [["xìn", "书信", "信，信任的信"]],
     "皇": ["huánɡ", null, "皇，皇帝的皇"],
     "泉": ["quán", null, "泉，泉水的泉"],
     "鬼": ["ɡuǐ", null, "鬼，鬼怪的鬼"],
@@ -2647,10 +2640,7 @@
         ["hóu", "封侯", "侯，诸侯的侯"],
         ["hòu", "地名", "侯，闽侯的侯"]
     ],
-    "追": [
-        ["zhuī", "追求", "追，追赶的追"],
-        ["duī", "古同'雕'", "追，追琢的追"]
-    ],
+    "追": [["zhuī", "追求", "追，追赶的追"]],
     "俊": ["jùn", null, "俊，俊杰的俊"],
     "盾": ["dùn", null, "盾，盾牌的盾"],
     "待": [
@@ -2682,8 +2672,8 @@
         ["pàn", "大号的", "胖，胖袄的胖"]
     ],
     "脉": [
-        ["mò", "含情脉脉", "脉，脉搏的脉"],
-        ["mài", "血管", "脉，脉搏的脉"]
+        ["mài", "血管", "脉，脉搏的脉"],
+        ["mò", "含情脉脉", "脉，含情脉脉的脉"]
     ],
     "胎": ["tāi", null, "胎，胎儿的胎"],
     "勉": ["miǎn", null, "勉，勉强的勉"],
@@ -2737,8 +2727,8 @@
     "阀": ["fá", null, "阀，阀门，阀门的阀"],
     "阁": ["ɡé", null, "阁，阁楼，阁楼的阁"],
     "差": [
-        ["chà", "差别", "差，出差，出差的差"],
-        ["chā", "差距", "差，差异的差"],
+        ["chà", "差别", "差，差劲的差"],
+        ["chā", "差距", "差，差距的差"],
         ["chāi", "差事", "差，出差的差"],
         ["cī", "不齐", "差，参差的差"]
     ],
@@ -2876,8 +2866,8 @@
     "院": ["yuàn", null, "院，庭院的院"],
     "娃": ["wá", null, "娃，娃娃的娃"],
     "姥": [
-        ["mǔ", "外婆", "姥，姥姥的姥"],
-        ["lǎo", "外祖母", "姥，姥姥的姥"]
+        ["lǎo", "外祖母", "姥，姥姥的姥"],
+        ["mǔ", "外婆", "姥，天姥山的姥"]
     ],
     "姨": ["yí", null, "姨，姨妈的姨"],
     "姻": ["yīn", null, "姻，姻亲的姻"],
@@ -3087,20 +3077,20 @@
     "晓": ["xiǎo", null, "晓，清晨的晓"],
     "哮": ["xiāo", null, "哮，咆哮的哮"],
     "唠": [
-        ["lào", "聊天", "唠，唠叨的唠"],
+        ["lào", "聊天", "唠，唠嗑的唠"],
         ["láo", "说个不停", "唠，唠叨的唠"]
     ],
     "鸭": ["yā", null, "鸭，鸭子的鸭"],
     "晃": [
-        ["huǎng", "晃动（快）", "晃，摇晃的晃"],
-        ["huàng", "摇动", "晃，摇晃的晃"]
+        ["huàng", "摇动", "晃，摇晃的晃"],
+        ["huǎng", "晃动（快）", "晃，晃眼的晃"]
     ],
     "哺": ["bǔ", null, "哺，喂养的哺"],
     "晌": ["shǎnɡ", null, "晌，午后的晌"],
     "剔": ["tī", null, "剔，剔除的剔"],
     "晕": [
-        ["yùn", "头晕", "晕，头晕的晕"],
-        ["yūn", "昏厥", "晕，晕倒的晕"]
+        ["yūn", "昏厥", "晕，晕倒的晕"],
+        ["yùn", "头晕", "晕，光晕的晕"]
     ],
     "蚌": [
         ["bàng", "河蚌", "蚌，蚌壳的蚌"],
@@ -3369,7 +3359,7 @@
     "涉": ["shè", null, "涉，涉及，牵涉到某件事情或某个领域"],
     "消": ["xiāo", null, "消，消除，去除、去掉某物"],
     "涡": [
-        ["wō", "旋涡", "涡，漩涡，水流形成的旋转漏斗状物体"],
+        ["wō", "旋涡", "涡，漩涡的涡"],
         ["guō", "河流名", "涡，涡河的涡"]
     ],
     "浩": ["hào", null, "浩，浩瀚，形容水势、星空等非常广阔"],
@@ -3454,8 +3444,8 @@
     "谈": ["tán", null, "谈，谈笑风生的谈"],
     "谊": ["yì", null, "谊，友谊的谊"],
     "剥": [
-        ["bāo", "口语读音", "剥，剥削的剥"],
-        ["bō", "剥离", "剥，剥削的剥"]
+        ["bō", "剥离", "剥，剥削的剥"],
+        ["bāo", "口语读音", "剥，剥削的剥"]
     ],
     "恳": ["kěn", null, "恳，恳切的恳"],
     "展": ["zhǎn", null, "展，展示的展"],
@@ -3637,8 +3627,8 @@
     "常": ["chánɡ", null, "常，常常的常"],
     "眶": ["kuànɡ", null, "眶，眼眶的眶"],
     "匙": [
-        ["chí", "勺子", "匙，钥匙的匙"],
-        ["shi", "开锁工具", "匙，钥匙的匙"]
+        ["shi", "开锁工具", "匙，钥匙的匙"],
+        ["chí", "勺子", "匙，汤匙的匙"]
     ],
     "晨": ["chén", null, "晨，早晨的晨"],
     "睁": ["zhēnɡ", null, "睁，睁开的睁"],
@@ -3709,8 +3699,8 @@
     ],
     "铐": ["kào", null, "铐，手铐的铐"],
     "铛": [
-        ["chēng", "平底锅", "铛，钟铛的铛"],
-        ["dāng", "拟声词", "铛，铛铛的铛"]
+        ["dāng", "拟声词", "铛，铛铛的铛"],
+        ["chēng", "平底锅", "铛，钟铛的铛"]
     ],
     "铝": ["lǚ", null, "铝，铝制的铝"],
     "铜": ["tónɡ", null, "铜，铜钱的钱"],
@@ -4093,8 +4083,8 @@
     ],
     "韩": ["hán", null, "韩，韩国的韩"],
     "朝": [
-        ["zhāo", "早晨", "朝，朝鲜的朝"],
-        ["cháo", "朝廷", "朝，朝代的朝"]
+        ["cháo", "朝廷", "朝，朝代的朝"],
+        ["zhāo", "早晨", "朝，朝阳的朝"]
     ],
     "辜": ["ɡū", null, "辜，辜月的辜"],
     "葵": ["kuí", null, "葵，向日的葵"],
@@ -4135,23 +4125,23 @@
     "粟": ["sù", null, "粟，粟米，粟的粟"],
     "棘": ["jí", null, "棘，荆棘，棘的棘"],
     "酣": ["hān", null, "酣，酣睡，酣的酣"],
-    "酥": ["sū", null, "酥，指用油类和面粉等制成的软食品，如酥油糕点等"],
-    "厨": ["chú", null, "厨，指负责烹饪食物的厨师"],
+    "酥": ["sū", null, "酥，酥油的酥"],
+    "厨": ["chú", null, "厨，厨房的厨"],
     "厦": [
         ["shà", "高楼", "厦，是厦门市的简称"],
         ["xià", "地名", "厦，厦门的厦"]
     ],
-    "硬": ["yìnɡ", null, "硬，形容物体的质地非常坚硬"],
-    "硝": ["xiāo", null, "硝，指一种化学物质，常用于制作火药等"],
-    "确": ["què", null, "确，表示肯定、确定的意思"],
-    "硫": ["liú", null, "硫，是一种化学元素，常用于制作肥料等"],
-    "雁": ["yàn", null, "雁，一种候鸟，属于鸭科"],
+    "硬": ["yìnɡ", null, "硬，坚硬的硬"],
+    "硝": ["xiāo", null, "硝，硝酸的硝"],
+    "确": ["què", null, "确，正确的确"],
+    "硫": ["liú", null, "硫，硫磺的硫"],
+    "雁": ["yàn", null, "雁，大雁的雁"],
     "殖": [
-        ["zhí", "殖民", "殖，指生物体通过生殖繁衍后代的过程"],
+        ["zhí", "殖民", "殖，殖民的殖"],
         ["shi", "骨头", "殖，骨殖的殖"]
     ],
     "裂": [
-        ["liè", "分裂", "裂，指物体表面出现的裂缝或分裂"],
+        ["liè", "分裂", "裂，裂缝的裂"],
         ["liě", "口语", "裂，裂开的裂"]
     ],
     "雄": ["xiónɡ", null, "雄，形容事物雄伟、壮观"],
@@ -4365,8 +4355,8 @@
         ["suì", "于是", "遂，遂愿的遂"]
     ],
     "曾": [
-        ["zēng", "隔代", "曾，曾经的曾"],
-        ["céng", "曾经", "曾，曾经的曾"]
+        ["céng", "曾经", "曾，曾经的曾"],
+        ["zēng", "隔代", "曾，曾孙的曾"]
     ],
     "焰": ["yàn", null, "焰，火焰的焰"],
     "港": [
@@ -4666,7 +4656,7 @@
         ["xiè", "姓氏", "解，解姓的解"]
     ],
     "煞": [
-        ["shā", "极；很", "煞，煞费苦心煞"],
+        ["shā", "极；很", "煞，煞费苦心的煞"],
         ["shà", "凶气", "煞，煞气的煞"]
     ],
     "雏": ["chú", null, "雏，雏菊雏"],

packages/voice-production/data/temp/voice_generated_v2.json

@@ -0,0 +1 @@
+{"generated":["一_yi1","乙_yi3","二_er4","十_shi2","丁_ding1","厂_chang3","七_qi1","卜_bu3","八_ba1","人_ren2"],"updatedAt":"2026-02-11T03:30:00.988Z"}

packages/voice-production/package.json

@@ -7,6 +7,7 @@
         "test": "echo \"Error: no test specified\" && exit 1",
         "generate-introduce": "cross-env node scripts/generate_char_introduce.js",
         "generate-voice": "node scripts/generate_char_voice.js",
+        "generate-voice-v2": "node scripts/generate_char_voice_v2.js",
         "deploy-assets": "node scripts/deploy-assets.js",
         "deploy-assets:local": "node scripts/deploy-assets.js --local",
         "deploy-assets:server": "node scripts/deploy-assets.js --server"

packages/voice-production/scripts/generate_char_voice_v2.js

@@ -0,0 +1,397 @@
+/**
+ * 使用腾讯云语音合成 TTS 为汉字介绍生成语音（v2）
+ * 读取 char_introduce_v2.json，按单音字/多音字取第一条 [pinyin, meaning, introduce]，
+ * 用 SSML <phoneme> 仅标注当前字的读音，生成语音保存到 data/voice，命名：汉字_拼音.mp3
+ */
+const fs = require('fs');
+const path = require('path');
+const crypto = require('crypto');
+const https = require('https');
+require('dotenv').config({ path: path.join(__dirname, '../.env') });
+
+const SECRET_ID = process.env.TENCENT_SECRET_ID;
+const SECRET_KEY = process.env.TENCENT_SECRET_KEY;
+
+const HOST = 'tts.tencentcloudapi.com';
+const SERVICE = 'tts';
+const REGION = 'ap-guangzhou';
+const ACTION = 'TextToVoice';
+const VERSION = '2019-08-23';
+
+const VOICE_TYPE = 101015;
+const CODEC = 'mp3';
+const SAMPLE_RATE = 16000;
+
+/** 只处理前 N 个汉字，便于调试；设为 0 或 null 表示不限制 */
+const MAX_CHAR_COUNT = 10;
+
+/** 已生成记录文件（存放在 data/temp） */
+const VOICE_RECORD_FILE_V2 = 'voice_generated_v2.json';
+
+const SKIP_KEYS = new Set(['_schema', '_verified_before']);
+
+// 带声调字母 -> 无声调字母
+const TONE_TO_VOWEL = {
+    ā: 'a',
+    á: 'a',
+    ǎ: 'a',
+    à: 'a',
+    ē: 'e',
+    é: 'e',
+    ě: 'e',
+    è: 'e',
+    ī: 'i',
+    í: 'i',
+    ǐ: 'i',
+    ì: 'i',
+    ō: 'o',
+    ó: 'o',
+    ǒ: 'o',
+    ò: 'o',
+    ū: 'u',
+    ú: 'u',
+    ǔ: 'u',
+    ù: 'u',
+    ǖ: 'v',
+    ǘ: 'v',
+    ǚ: 'v',
+    ǜ: 'v',
+};
+// 带声调字母 -> 声调数字 1-4
+const TONE_TO_NUM = {
+    ā: 1,
+    á: 2,
+    ǎ: 3,
+    à: 4,
+    ē: 1,
+    é: 2,
+    ě: 3,
+    è: 4,
+    ī: 1,
+    í: 2,
+    ǐ: 3,
+    ì: 4,
+    ō: 1,
+    ó: 2,
+    ǒ: 3,
+    ò: 4,
+    ū: 1,
+    ú: 2,
+    ǔ: 3,
+    ù: 4,
+    ǖ: 1,
+    ǘ: 2,
+    ǚ: 3,
+    ǜ: 4,
+};
+
+/**
+ * 将带声调符号的拼音转为腾讯 SSML 用的格式：无声调 + 数字声调（如 chǎng -> chang3）
+ * 无声调拼音（如 le、bo）末尾补 5（轻声）
+ * 数据中可能含 Unicode 印刷体 ɡ (U+0261)，需规范为 ASCII g，否则 TTS 报 invalid words
+ */
+function pinyinToPhoneme(pinyin) {
+    if (!pinyin || typeof pinyin !== 'string') return 'unknown5';
+    // 规范为 TTS 可识别的拼音字符：ɡ(U+0261)->g, ɑ(U+0251)->a
+    let normalized = pinyin.replace(/\u0261/g, 'g').replace(/\u0251/g, 'a');
+    let tone = 5;
+    let base = '';
+    for (const c of normalized) {
+        if (TONE_TO_VOWEL[c] !== undefined) {
+            base += TONE_TO_VOWEL[c];
+            tone = TONE_TO_NUM[c];
+        } else {
+            base += c;
+        }
+    }
+    // ü 在 SSML 中常用 v 表示
+    base = base.replace(/ü/g, 'v');
+    return base + tone;
+}
+
+/**
+ * 将 introduce 中所有「当前字」用 <phoneme> 包裹，只标该字读音
+ */
+function buildSSML(introduce, char, pinyinPhoneme) {
+    const tag = `<phoneme alphabet="py" ph="${pinyinPhoneme}">${escapeXml(char)}</phoneme>`;
+    const parts = introduce.split(char);
+    const ssmlBody = parts.join(tag);
+    return `<speak>${ssmlBody}</speak>`;
+}
+
+function escapeXml(s) {
+    return s
+        .replace(/&/g, '&amp;')
+        .replace(/</g, '&lt;')
+        .replace(/>/g, '&gt;')
+        .replace(/"/g, '&quot;')
+        .replace(/'/g, '&apos;');
+}
+
+function sha256(message, secret = '', encoding = 'hex') {
+    const hmac = secret ? crypto.createHmac('sha256', secret) : crypto.createHash('sha256');
+    return hmac.update(message).digest(encoding);
+}
+
+function getDate(timestamp) {
+    const date = new Date(timestamp * 1000);
+    const year = date.getUTCFullYear();
+    const month = ('0' + (date.getUTCMonth() + 1)).slice(-2);
+    const day = ('0' + date.getUTCDate()).slice(-2);
+    return `${year}-${month}-${day}`;
+}
+
+function generateSignature(secretId, secretKey, host, payload, timestamp) {
+    const date = getDate(timestamp);
+    const httpRequestMethod = 'POST';
+    const canonicalUri = '/';
+    const canonicalQueryString = '';
+    const contentType = 'application/json';
+    const canonicalHeaders = `content-type:${contentType}\nhost:${host}\nx-tc-action:${ACTION.toLowerCase()}\n`;
+    const signedHeaders = 'content-type;host;x-tc-action';
+    const hashedRequestPayload = sha256(payload);
+    const canonicalRequest = `${httpRequestMethod}\n${canonicalUri}\n${canonicalQueryString}\n${canonicalHeaders}\n${signedHeaders}\n${hashedRequestPayload}`;
+
+    const algorithm = 'TC3-HMAC-SHA256';
+    const credentialScope = `${date}/${SERVICE}/tc3_request`;
+    const hashedCanonicalRequest = sha256(canonicalRequest);
+    const stringToSign = `${algorithm}\n${timestamp}\n${credentialScope}\n${hashedCanonicalRequest}`;
+
+    const secretDate = sha256(date, 'TC3' + secretKey, 'buffer');
+    const secretService = sha256(SERVICE, secretDate, 'buffer');
+    const secretSigning = sha256('tc3_request', secretService, 'buffer');
+    const signature = sha256(stringToSign, secretSigning);
+    const authorization = `${algorithm} Credential=${secretId}/${credentialScope}, SignedHeaders=${signedHeaders}, Signature=${signature}`;
+
+    return authorization;
+}
+
+function textToVoice(text) {
+    return new Promise((resolve, reject) => {
+        const timestamp = Math.floor(Date.now() / 1000);
+        const sessionId = `voice-v2-${timestamp}-${Math.random().toString(36).slice(2, 10)}`;
+
+        const payload = JSON.stringify({
+            Text: text,
+            SessionId: sessionId,
+            VoiceType: VOICE_TYPE,
+            PrimaryLanguage: 1,
+            SampleRate: SAMPLE_RATE,
+            Codec: CODEC,
+            Speed: 0,
+        });
+
+        const authorization = generateSignature(SECRET_ID, SECRET_KEY, HOST, payload, timestamp);
+
+        const options = {
+            hostname: HOST,
+            port: 443,
+            path: '/',
+            method: 'POST',
+            headers: {
+                'Content-Type': 'application/json',
+                Host: HOST,
+                'X-TC-Action': ACTION,
+                'X-TC-Version': VERSION,
+                'X-TC-Timestamp': timestamp.toString(),
+                'X-TC-Region': REGION,
+                Authorization: authorization,
+            },
+        };
+
+        const req = https.request(options, (res) => {
+            let data = '';
+            res.on('data', (chunk) => (data += chunk));
+            res.on('end', () => {
+                try {
+                    const result = JSON.parse(data);
+                    if (result.Response && result.Response.Audio) {
+                        resolve(Buffer.from(result.Response.Audio, 'base64'));
+                    } else if (result.Response && result.Response.Error) {
+                        reject(new Error(result.Response.Error.Message));
+                    } else {
+                        reject(new Error('Unknown API response: ' + data));
+                    }
+                } catch (e) {
+                    reject(e);
+                }
+            });
+        });
+
+        req.on('error', reject);
+        req.write(payload);
+        req.end();
+    });
+}
+
+/**
+ * 文件名：汉字_拼音数字.mp3，拼音中的特殊字符（如 ɡ）规范为 ASCII
+ */
+function getVoiceFilename(char, pinyinPhoneme) {
+    const safePinyin = pinyinPhoneme.replace(/\s/g, '').replace(/[^\w\-]/g, '_');
+    return `${char}_${safePinyin}.${CODEC}`;
+}
+
+/** 生成记录用的 key（与文件名不含扩展名一致） */
+function getRecordKey(char, pinyinPhoneme) {
+    const safePinyin = pinyinPhoneme.replace(/\s/g, '').replace(/[^\w\-]/g, '_');
+    return `${char}_${safePinyin}`;
+}
+
+/**
+ * 从 data/temp 目录加载已生成记录
+ */
+function loadVoiceRecordV2(tempDir) {
+    const recordPath = path.join(tempDir, VOICE_RECORD_FILE_V2);
+    let generated = [];
+    if (fs.existsSync(recordPath)) {
+        try {
+            const data = JSON.parse(fs.readFileSync(recordPath, 'utf-8'));
+            generated = data.generated || [];
+        } catch (e) {
+            generated = [];
+        }
+    }
+    return new Set(generated);
+}
+
+/**
+ * 将已生成记录写入 data/temp
+ */
+function saveVoiceRecordV2(tempDir, generatedList) {
+    const recordPath = path.join(tempDir, VOICE_RECORD_FILE_V2);
+    fs.writeFileSync(
+        recordPath,
+        JSON.stringify({
+            generated: generatedList,
+            updatedAt: new Date().toISOString(),
+        }),
+        'utf-8'
+    );
+}
+
+function delay(ms) {
+    return new Promise((resolve) => setTimeout(resolve, ms));
+}
+
+/**
+ * 从 char_introduce_v2 的一条 value 解析出 [pinyin, meaning, introduce]
+ * 规则：
+ * - 单音字：一维数组 [pinyin, meaning, introduce]，index0=拼音，index2=introduce
+ * - 多音字：二维数组 [[pinyin, meaning, introduce], ...]，用第一个元素是 string 还是 array 判断；
+ *   多音字只读第一个子数组，index0=拼音，index2=introduce
+ */
+function parseEntry(value) {
+    if (!Array.isArray(value) || value.length === 0) return null;
+    const first = value[0];
+    // 单音字：首元素是 string
+    if (typeof first === 'string') {
+        if (value.length < 3) return null;
+        return [value[0], value[1], value[2]];
+    }
+    // 多音字：首元素是 array，取第一个子数组
+    if (Array.isArray(first) && first.length >= 3) {
+        return [first[0], first[1], first[2]];
+    }
+    return null;
+}
+
+/**
+ * 从 JSON 文件内容中按「出现顺序」提取 key 列表，避免依赖 Object 遍历顺序。
+ * 只匹配行首的 "key":，避免误匹配到值字符串内的引号。
+ */
+function getKeyOrderFromJsonText(text) {
+    const keys = [];
+    const re = /^\s*"([^"]+)":\s*(?:\[|")/gm;
+    let m;
+    while ((m = re.exec(text)) !== null) {
+        const key = m[1];
+        if (!SKIP_KEYS.has(key)) {
+            keys.push(key);
+        }
+    }
+    return keys;
+}
+
+async function main() {
+    if (!SECRET_ID || !SECRET_KEY) {
+        console.error('请在 .env 文件中配置 TENCENT_SECRET_ID 和 TENCENT_SECRET_KEY');
+        process.exit(1);
+    }
+
+    const charDir = path.join(__dirname, '../data/char');
+    const introducePath = path.join(charDir, 'char_introduce_v2.json');
+    const voiceDir = path.join(__dirname, '../data/voice');
+    const tempDir = path.join(__dirname, '../data/temp');
+
+    if (!fs.existsSync(introducePath)) {
+        console.error('未找到 char_introduce_v2.json');
+        process.exit(1);
+    }
+
+    if (!fs.existsSync(voiceDir)) {
+        fs.mkdirSync(voiceDir, { recursive: true });
+    }
+    if (!fs.existsSync(tempDir)) {
+        fs.mkdirSync(tempDir, { recursive: true });
+    }
+
+    const text = fs.readFileSync(introducePath, 'utf-8');
+    const raw = JSON.parse(text);
+    // 按 JSON 文件中 key 的出现顺序构建列表，不依赖 Object 遍历顺序
+    const keyOrder = getKeyOrderFromJsonText(text);
+    const entries = keyOrder
+        .map((char) => {
+            const value = raw[char];
+            const parsed = parseEntry(value);
+            return parsed
+                ? { char, pinyin: parsed[0], meaning: parsed[1], introduce: parsed[2] }
+                : null;
+        })
+        .filter(Boolean);
+
+    const list = MAX_CHAR_COUNT ? entries.slice(0, MAX_CHAR_COUNT) : entries;
+    const generatedSet = loadVoiceRecordV2(tempDir);
+    const toGenerate = list.filter((item) => {
+        const key = getRecordKey(item.char, pinyinToPhoneme(item.pinyin));
+        return !generatedSet.has(key);
+    });
+    const generatedList = Array.from(generatedSet);
+
+    console.log(
+        `共 ${entries.length} 个汉字，本次处理前 ${list.length} 个${MAX_CHAR_COUNT ? `（MAX_CHAR_COUNT=${MAX_CHAR_COUNT}）` : ''}；已生成 ${generatedList.length} 个，待生成 ${toGenerate.length} 个`
+    );
+
+    const DELAY_MS = 200;
+
+    for (let i = 0; i < toGenerate.length; i++) {
+        const { char, pinyin, introduce } = toGenerate[i];
+        const pinyinPhoneme = pinyinToPhoneme(pinyin);
+        const recordKey = getRecordKey(char, pinyinPhoneme);
+        const ssml = buildSSML(introduce, char, pinyinPhoneme);
+        const filename = getVoiceFilename(char, pinyinPhoneme);
+        const filepath = path.join(voiceDir, filename);
+
+        try {
+            const audioBuffer = await textToVoice(ssml);
+            fs.writeFileSync(filepath, audioBuffer);
+            generatedList.push(recordKey);
+            saveVoiceRecordV2(tempDir, generatedList);
+            console.log(
+                `[${i + 1}/${toGenerate.length}] ${char} (${pinyin} -> ${pinyinPhoneme}) -> ${filename}`
+            );
+        } catch (error) {
+            console.error(`[${i + 1}/${toGenerate.length}] 失败 ${char}: ${error.message}`);
+        }
+
+        if (i < toGenerate.length - 1) {
+            await delay(DELAY_MS);
+        }
+    }
+
+    console.log(
+        `\n语音已保存到: ${voiceDir}，已生成记录: ${path.join(tempDir, VOICE_RECORD_FILE_V2)}`
+    );
+}
+
+main().catch(console.error);