feat(capcut-pipeline): 将 TTS 配音切换为 CosyVoice 并重构动画系统
- 将 TTS 引擎从 Qwen-TTS 切换为阿里云 CosyVoice(DashScope WebSocket) - 输出格式从 WAV(24kHz)改为 MP3 - 重构图片动画分拆逻辑,支持组合动画(如"渐显+放大") - 移除字幕关键词高亮相关字段 - 移除已删除的 `uploadAudioToOSS` 函数,统一使用 `uploadToOSS` - 更新文档和配置默认值以匹配新引擎
This commit is contained in:
@@ -19,7 +19,7 @@ description: 剪映/CapCut 自动化。通过 CapCut Mate API 实现草稿创建
|
||||
|
||||
```
|
||||
1. npm 依赖 → cd .claude/skills/video-from-script/scripts && npm install
|
||||
2. TTS 配音 → 阿里云 Qwen-TTS(config.json 中配置 ttsApiKey)
|
||||
2. TTS 配音 → 阿里云 CosyVoice TTS(config.json 中配置 ttsApiKey)
|
||||
```
|
||||
|
||||
---
|
||||
@@ -81,7 +81,7 @@ digraph capcut_assembly {
|
||||
|
||||
input [label="素材 + manifest.json", shape=folder, fillcolor="#e3f2fd"]
|
||||
|
||||
step1 [label="1. TTS 配音(可选)\nnode qwen-tts.js\n→ WAV + 时长"]
|
||||
step1 [label="1. TTS 配音(可选)\ncosyvoice → MP3 + 时长"]
|
||||
step2 [label="2. 上传图片到 OSS\n本地图片 → 公网 URL"]
|
||||
step3 [label="3. 创建草稿\ncreate_draft\n→ draft_url"]
|
||||
step4 [label="4. 导入素材+音频+字幕+特效\nadd_images / add_videos\nadd_audios / add_captions\nadd_effects"]
|
||||
@@ -146,15 +146,12 @@ digraph capcut_assembly {
|
||||
|
||||
## TTS 配音(成片模式用)
|
||||
|
||||
使用阿里云 Qwen-TTS(Node.js),替代原 Edge-TTS。
|
||||
使用阿里云 CosyVoice TTS(通过 DashScope WebSocket),Pipeline 自动调用。
|
||||
|
||||
```bash
|
||||
# 准备输入
|
||||
echo '{"segments":[{"id":1,"text":"文案"}],"voice":"Cherry","output_dir":"./audio"}' > input.json
|
||||
|
||||
# 生成
|
||||
node .claude/skills/video-from-script/scripts/qwen-tts.js input.json
|
||||
# → stdout: {"segments":[{"id":1,"audio":"./audio/seg_001.wav","duration":3.456}]}
|
||||
```js
|
||||
// 作为模块调用
|
||||
const { synthesize } = require('./qwen-tts')
|
||||
const { filePath, duration } = await synthesize('你好世界', { voice: 'Cherry' })
|
||||
```
|
||||
|
||||
配置在 `skills/config.json`:`ttsApiKey`(必填)、`ttsModel`、`ttsVoice`、`ttsLanguage`。
|
||||
@@ -195,19 +192,19 @@ node .claude/skills/video-from-script/scripts/qwen-tts.js input.json
|
||||
|
||||
## 图片动画预设
|
||||
|
||||
| 动画 | 说明 | 适用 |
|
||||
| 动画名称 | 说明 | 适用 |
|
||||
|------|------|------|
|
||||
| Ken Burns (zoom-in) | 1.0→1.1 缓慢放大 | 默认 |
|
||||
| Ken Burns (pan-left) | 右→左平移 | 风景 |
|
||||
| Ken Burns (pan-right) | 左→右平移 | 风景 |
|
||||
| 缩放弹出 | 0.8→1.0 | 强调 |
|
||||
| 缩放 | 缓慢放大(默认) | 通用 |
|
||||
| 渐显+放大 | 淡入+放大组合 | 强调 |
|
||||
| 左平移 | 右→左平移 | 风景 |
|
||||
| 右平移 | 左→右平移 | 风景 |
|
||||
|
||||
---
|
||||
|
||||
## 质量要求
|
||||
|
||||
- 字幕与文案对应正确,关键词高亮醒目
|
||||
- 图片动画流畅(Ken Burns 幅度 1.0→1.1)
|
||||
- 字幕与文案对应正确
|
||||
- 图片动画流畅
|
||||
- BGM 音量不盖过配音(配音为主、BGM 为辅)
|
||||
- 转场自然(无黑帧、无跳帧)
|
||||
- 底部字幕区不被遮挡
|
||||
|
||||
@@ -12,7 +12,7 @@
|
||||
1. CapCut Mate API 可达 → curl {config.capcutMateApiBase}/../docs
|
||||
- 部署在 capcut.muyetools.cn(配置在 skills/config.json)
|
||||
2. npm 依赖 → cd scripts && npm install
|
||||
3. TTS 配音 → 阿里云 Qwen-TTS(配置在 config.json 的 ttsApiKey)
|
||||
3. TTS 配音 → 阿里云 CosyVoice TTS(配置在 config.json 的 ttsApiKey)
|
||||
4. 同步到本地剪映 → 纯 Node.js(sync-to-jianying.js),无需 Python/uv
|
||||
```
|
||||
|
||||
@@ -57,7 +57,7 @@ digraph assembly_flow {
|
||||
node [shape=box, style=filled, fillcolor="#f5f5f5", fontsize=11]
|
||||
|
||||
input [label="素材 + manifest.json", shape=folder, fillcolor="#e3f2fd"]
|
||||
step1 [label="1. TTS 配音(可选)\nnode qwen-tts.js\n→ WAV + 时长"]
|
||||
step1 [label="1. TTS 配音(可选)\ncosyvoice → MP3 + 时长"]
|
||||
step2 [label="2. 上传图片到 OSS\n本地图片 → 公网 URL"]
|
||||
step3 [label="3. 创建草稿\ncreate_draft → draft_url"]
|
||||
step4 [label="4. 导入素材+音频+字幕+特效"]
|
||||
@@ -75,15 +75,12 @@ digraph assembly_flow {
|
||||
|
||||
### 1. TTS 配音(可选)
|
||||
|
||||
使用阿里云 Qwen-TTS 进行语音合成(Node.js,无需 Python)。
|
||||
使用阿里云 CosyVoice TTS 进行语音合成(通过 DashScope WebSocket,Node.js)。
|
||||
|
||||
```bash
|
||||
# 准备输入 JSON
|
||||
echo '{"segments":[{"id":1,"text":"第一段文案"},{"id":2,"text":"第二段文案"}],"voice":"Cherry","output_dir":"./audio"}' > input.json
|
||||
|
||||
# 批量生成
|
||||
node scripts/qwen-tts.js input.json
|
||||
# → stdout: {"segments":[{"id":1,"text":"...","audio":"./audio/seg_001.wav","duration":3.456}]}
|
||||
```js
|
||||
const { synthesize } = require('./qwen-tts')
|
||||
const { filePath, duration } = await synthesize('文案', { voice: 'Neil', outputDir: './audio' })
|
||||
// → ./audio/seg_001.mp3, duration: 3.456
|
||||
```
|
||||
|
||||
配置在 `skills/config.json`:
|
||||
@@ -91,8 +88,8 @@ node scripts/qwen-tts.js input.json
|
||||
| 字段 | 说明 | 默认值 |
|
||||
|------|------|--------|
|
||||
| `ttsApiKey` | 阿里云百炼 API Key | (必填) |
|
||||
| `ttsModel` | 模型名称 | `qwen-tts` |
|
||||
| `ttsVoice` | 音色名称 | `Cherry` |
|
||||
| `ttsModel` | 模型名称 | `cosyvoice-v3.5-plus` |
|
||||
| `ttsVoice` | 音色名称 | 账号配置 |
|
||||
| `ttsLanguage` | 语言类型 | `Chinese` |
|
||||
|
||||
推荐音色:
|
||||
@@ -108,7 +105,7 @@ node scripts/qwen-tts.js input.json
|
||||
| `Neil` | 阿闻 | 新闻主持人 | 新闻、财经 |
|
||||
| `Bellona` | 燕铮莺 | 洪亮有力女声 | 热血、武侠 |
|
||||
|
||||
所有音色均支持中英文,输出 WAV 格式(24kHz),URL 有效期 24 小时。
|
||||
所有音色均支持中英文,输出 MP3 格式(24kHz)。
|
||||
|
||||
**作为模块调用**:
|
||||
|
||||
@@ -151,7 +148,7 @@ POST /create_draft { width: 1080, height: 1920 }
|
||||
|
||||
```
|
||||
POST /add_images
|
||||
每张图片 3-5 秒,附带 Ken Burns 动画(缩放 1.0→1.1)
|
||||
每张图片 3-5 秒,附带动画(默认缩放)
|
||||
```
|
||||
|
||||
**视频模式** (`--mode videos`):
|
||||
@@ -173,8 +170,7 @@ POST /add_audios
|
||||
|
||||
```
|
||||
POST /add_captions
|
||||
- 文案来自 manifest.json
|
||||
- 关键词高亮(account.json 中 subtitleStyle.highlightColor)
|
||||
- 文案来自 manifest.json(TTS 分句时按 segment 精确对齐)
|
||||
- 字体大小、颜色从账号配置读取
|
||||
```
|
||||
|
||||
@@ -212,19 +208,18 @@ add_videos 提交 9+ 视频时可能触发网关 504。脚本自动降级:
|
||||
|
||||
## 图片动画预设
|
||||
|
||||
| 动画类型 | 说明 | 适用场景 |
|
||||
| 动画名称 | 说明 | 适用场景 |
|
||||
|---------|------|---------|
|
||||
| Ken Burns (zoom-in) | 从 1.0 缓慢放大到 1.1 | 默认,适合大多数场景 |
|
||||
| Ken Burns (pan-left) | 画面从右向左平移 | 风景、全景 |
|
||||
| Ken Burns (pan-right) | 画面从左向右平移 | 风景、全景 |
|
||||
| 缩放弹出 | 从 0.8 弹到 1.0 | 强调、冲击感 |
|
||||
| 缩放 | 缓慢放大 | 默认,适合大多数场景 |
|
||||
| 渐显+放大 | 淡入+放大组合 | 强调、冲击感 |
|
||||
| 左平移 | 右→左平移 | 风景、全景 |
|
||||
| 右平移 | 左→右平移 | 风景、全景 |
|
||||
|
||||
---
|
||||
|
||||
## 质量检查
|
||||
|
||||
- [ ] 字幕与文案对应正确
|
||||
- [ ] 关键词高亮颜色醒目
|
||||
- [ ] 图片动画流畅(无卡顿)
|
||||
- [ ] BGM 音量与配音平衡
|
||||
- [ ] 转场自然(无黑帧)
|
||||
|
||||
@@ -334,10 +334,9 @@ node kling-video-generator.js --image <url> --prompt <prompt> -o ./videos
|
||||
```
|
||||
output/{name}_{YYYYMMDD}_{NNN}/
|
||||
├── manifest.json # 主清单(贯穿全流程)
|
||||
├── prompts.txt # 原始提示词存档
|
||||
├── images/ # scene_{NN}_{slug}.jpeg(slug 从 script/shotDesc 派生,首尾帧加 _last 后缀)
|
||||
├── videos/ # scene_{NN}_{slug}.mp4(与图片对应)
|
||||
└── urls.json # OSS 公网 URL 映射
|
||||
└── audio/ # seg_001.mp3(TTS 分句音频,多句时 seg_{id}_{j}.mp3)
|
||||
```
|
||||
|
||||
**命名对应关系**:图片 `scene_01_觉醒.jpeg` → 视频 `scene_01_觉醒.mp4`;首尾帧尾帧 `scene_01_觉醒_last.jpeg`;MJ 候选 `scene_01_觉醒_cand1.jpeg`
|
||||
@@ -396,7 +395,7 @@ output/{name}_{YYYYMMDD}_{NNN}/
|
||||
|
||||
所有子技能共享以下资源(位于本目录):
|
||||
|
||||
- `scripts/` — 共享脚本(gemini-image-generator.js, mj-image-generator.js, grok-video-generator.js, veo-video-generator.js, capcut_assemble.js, sync-to-jianying.js, oss-upload.js)
|
||||
- `scripts/` — 共享脚本(gemini-image-generator.js, mj-image-generator.js, grok-video-generator.js, veo-video-generator.js, kling-video-generator.js, qwen-tts.js, capcut_assemble.js, sync-to-jianying.js, oss-upload.js)
|
||||
- `accounts/` — 账号配置(项目根目录,详见 [account-system.md](references/account-system.md))
|
||||
- `references/account-system.md` — 账号系统说明
|
||||
|
||||
|
||||
@@ -218,8 +218,7 @@ async function assemble(args) {
|
||||
format = '9:16',
|
||||
apiKey = '',
|
||||
duration = '4',
|
||||
animation = '缩放',
|
||||
localAudio = 'true',
|
||||
animation = '渐显+放大',
|
||||
} = args
|
||||
|
||||
if (!input) throw new Error('缺少 --input 参数')
|
||||
@@ -352,12 +351,11 @@ async function assemble(args) {
|
||||
// Step 2: 上传(已调速的)视频到 OSS
|
||||
const missingUrl = items.filter(it => it.video && !it.videoUrl)
|
||||
if (missingUrl.length > 0) {
|
||||
const { uploadFile } = require('./oss-upload')
|
||||
console.log(` 上传 ${missingUrl.length} 个视频到 OSS...`)
|
||||
for (const item of missingUrl) {
|
||||
const videoPath = path.resolve(inputDir, item.video)
|
||||
try {
|
||||
const { url } = await uploadFile(videoPath)
|
||||
const url = await uploadToOSS(videoPath)
|
||||
item.videoUrl = url
|
||||
// 回写 manifest
|
||||
if (manifestFile) {
|
||||
@@ -492,17 +490,12 @@ async function addImages(draftUrl, items, imgUrls, timeline, width, height, anim
|
||||
}
|
||||
|
||||
if (animation) {
|
||||
const parts = animation.split('+')
|
||||
for (const part of parts) {
|
||||
const name = part.trim()
|
||||
// 组合动画(持续整段):缩放、三分割 等
|
||||
if (name === '缩放' || name === '缩放 II') {
|
||||
info.loop_animation = name
|
||||
} else {
|
||||
// 默认作为入场动画
|
||||
info.in_animation = name
|
||||
}
|
||||
}
|
||||
const parts = animation.split('+').map(p => p.trim()).filter(Boolean)
|
||||
const groupNames = ['缩放', '缩放 II']
|
||||
const groupAnims = parts.filter(p => groupNames.includes(p))
|
||||
const inAnims = parts.filter(p => !groupNames.includes(p))
|
||||
if (groupAnims.length > 0) info.loop_animation = groupAnims.join('|')
|
||||
if (inAnims.length > 0) info.in_animation = inAnims.join('|')
|
||||
}
|
||||
|
||||
return info
|
||||
@@ -637,19 +630,9 @@ async function addVideos(draftUrl, inputDir, items, timeline, width, height, tra
|
||||
}
|
||||
|
||||
// ============================================================================
|
||||
// 音频上传(本地文件 → OSS 公网 URL)
|
||||
// 音频批量上传(本地文件 → OSS 公网 URL)
|
||||
// ============================================================================
|
||||
|
||||
async function uploadAudioToOSS(filePath) {
|
||||
try {
|
||||
const oss = require(path.join(__dirname, 'oss-upload'))
|
||||
const { url } = await oss.uploadFile(filePath)
|
||||
return url
|
||||
} catch (err) {
|
||||
throw new Error(`音频上传 OSS 失败: ${err.message}`)
|
||||
}
|
||||
}
|
||||
|
||||
async function batchUploadAudio(inputDir, items) {
|
||||
const urls = {}
|
||||
for (const item of items) {
|
||||
@@ -665,7 +648,7 @@ async function batchUploadAudio(inputDir, items) {
|
||||
continue
|
||||
}
|
||||
try {
|
||||
urls[seg.audio] = await uploadAudioToOSS(filePath)
|
||||
urls[seg.audio] = await uploadToOSS(filePath)
|
||||
console.log(` 上传: ${path.basename(filePath)} -> OK`)
|
||||
} catch (err) {
|
||||
console.error(` 上传失败: ${path.basename(filePath)} - ${err.message}`)
|
||||
@@ -686,7 +669,7 @@ async function batchUploadAudio(inputDir, items) {
|
||||
continue
|
||||
}
|
||||
try {
|
||||
urls[item.audio] = await uploadAudioToOSS(filePath)
|
||||
urls[item.audio] = await uploadToOSS(filePath)
|
||||
console.log(` 上传: ${path.basename(filePath)} -> OK`)
|
||||
} catch (err) {
|
||||
console.error(` 上传失败: ${path.basename(filePath)} - ${err.message}`)
|
||||
@@ -868,8 +851,6 @@ async function addSubtitles(draftUrl, items, timeline, style = {}, split = false
|
||||
start: currentTime,
|
||||
end: currentTime + duration,
|
||||
text: seg.text,
|
||||
keyword: '',
|
||||
keyword_color: '',
|
||||
}
|
||||
|
||||
if (inAnimation) cap.in_animation = inAnimation
|
||||
@@ -903,8 +884,6 @@ async function addSubtitles(draftUrl, items, timeline, style = {}, split = false
|
||||
start: currentTime,
|
||||
end: currentTime + duration,
|
||||
text: sentence,
|
||||
keyword: '',
|
||||
keyword_color: '',
|
||||
}
|
||||
|
||||
if (inAnimation) cap.in_animation = inAnimation
|
||||
@@ -918,16 +897,10 @@ async function addSubtitles(draftUrl, items, timeline, style = {}, split = false
|
||||
}
|
||||
} else {
|
||||
// 原始模式:一句字幕
|
||||
const keyword = ''
|
||||
const keywordColor = style.highlightColor || style.color || '#FFFFFF'
|
||||
|
||||
const cap = {
|
||||
start: tl.start,
|
||||
end: tl.end,
|
||||
text,
|
||||
keyword,
|
||||
keyword_color: keyword ? keywordColor : '',
|
||||
keyword_font_size: 18,
|
||||
}
|
||||
|
||||
if (inAnimation) cap.in_animation = inAnimation
|
||||
@@ -1040,7 +1013,7 @@ async function main() {
|
||||
console.log(' --duration 4 默认每段时长/秒(无TTS时的fallback,默认 4)')
|
||||
console.log(' --voiceover true|false 是否添加TTS配音轨道(默认 true)')
|
||||
console.log(' --subtitles true|false 是否添加字幕(默认 true)')
|
||||
console.log(' --split-captions true|false 分句字幕模式(默认 false,长句按标点切分)')
|
||||
console.log(' --split-captions true|false 分句字幕模式(默认 true,按标点切分)')
|
||||
console.log(' --bgm <url> 背景音乐 URL')
|
||||
console.log(' --effects "名称1,名称2" 特效名称(逗号分隔)')
|
||||
console.log(' --filter "名称:强度" 滤镜(强度 0-100)')
|
||||
|
||||
@@ -23,7 +23,7 @@ async function phaseAssemble(manifest, manifestPath, options) {
|
||||
subtitles: mode === 'images' ? 'true' : 'false',
|
||||
voiceover: manifest.items.some(it => it.audio) ? 'true' : 'false',
|
||||
duration: '4',
|
||||
animation: capcutConfig.animation || '缩放',
|
||||
animation: capcutConfig.animation || '渐显+放大',
|
||||
}
|
||||
|
||||
if (capcutConfig.defaultBGM) assembleArgs.bgm = capcutConfig.defaultBGM
|
||||
|
||||
Reference in New Issue
Block a user