@aswless_854771076/ai_short_studio_cli 0.1.161 → 0.1.162

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@aswless_854771076/ai_short_studio_cli",
3
- "version": "0.1.161",
3
+ "version": "0.1.162",
4
4
  "description": "VVICAT 无限画布项目命令行工具",
5
5
  "license": "MIT",
6
6
  "private": false,
@@ -13,9 +13,9 @@
13
13
 
14
14
  ### 声音一致性绑定
15
15
 
16
- 初始化时单独询问每个需要说话的人物的声音来源,并把选择写入 `audio-plan.voice_consistency`:
16
+ 初始化时汇总确认需要说话的人物的声音来源;用户已确认的选择不重复询问,并把选择写入 `audio-plan.voice_consistency`:
17
17
 
18
- - `first_appearance_sample`(默认):只把人物第一次出场且有清晰对白的已有视频或原声截取为声纹样本,样本本身不是固定音色 ID。先展示截取时长、来源镜头、波形/试听链接和权利状态,用户确认后再由当前 Provider 创建或选择真实 `voice_id`;当前 Provider 默认支持该绑定,只有实时能力明确返回不支持时才分流。没有已有来源音频时不得用尚未生成的首镜反推音色,必须停下让用户上传样本或选择 Provider voice。
18
+ - `first_generated_video`(默认):首次成功生成且含清晰对白的视频建立人物声音基准,后续镜头继承;已有来源样本时可切换 `first_appearance_sample`:只把人物第一次出场且有清晰对白的已有视频或原声截取为声纹样本,样本本身不是固定音色 ID。先展示截取时长、来源镜头、波形/试听链接和权利状态,用户确认后,若目标视频模型支持参考音频,则使用真实 AUDIO 节点绑定到 `reference-audios`,并回读声音计划逐人物引用的 nodeId、assetId、versionId、mediaId;只有实时 schema 明确接受 Provider voice ID 时才创建或选择该 ID,不把独立 TTS 音色 ID 当作视频参考音频。首镜尚未生成时不得伪造 voice_id 或媒体证据;先保留 reference=null、限制和复听计划,首镜完成后回写真实视频版本与音频证据,再供后续镜头继承。
19
19
  - `user_upload`:用户上传并确认有权使用的声音样本。
20
20
  - `global_asset`:从全局声音资产库选择已审核的 voice ID 和版本。
21
21
  - `provider_voice_id`:选择实时 Provider 目录中已确认的声音 ID。
@@ -53,6 +53,8 @@
53
53
  - 两类编辑都基于当前 `selectedOutputVersionId` 追加版本并使下游 stale。同节点、同基线、同载荷重复提交或当前选版已满足目标时返回已有版本;画布冲突刷新当前选版后重试一次,资产版本竞争保留原基线重试一次,身份错误和其他 409 直接停止核对。
54
54
  - `canvas apply` 只处理节点、边、删除和视口,不代替 settings 或结构化 artifact 编辑。
55
55
 
56
+ 节点创建后必须在同一批确定最终排版:按真实依赖边将上游放左、下游放右,间距按实际节点宽高计算,默认节点为 360×380px,列间至少留 120px、行间至少留 84px,不能把间距当作左上角坐标步长;保留已有节点坐标,只为新节点选择不遮挡现有节点的最近空位。不要把新节点先堆在画布中心再交给用户整理,排版只写节点 x/y,不写入 config,也不为排版创建媒体任务。
57
+
56
58
  ## 安全边界
57
59
 
58
60
  refresh token 只保存在系统钥匙串;密码和 API Key 只从交互式隐藏输入、stdin 或秘密管理器提供。不得读取、复制或删除本地认证缓存规避认证。项目供应商读取只返回遮罩状态,写入 Key 必须使用安全 stdin。删除既有项目、覆盖完整画布或改模型前先确认影响。
@@ -30,7 +30,7 @@ Rules:
30
30
  7. New plans must set `audio_strategy.mode=native-first`, `provider_selection=prefer-native`, and `fallback_allowed=true`. `fallback_reasons` must contain exactly: `provider-no-native-audio`, `voice-identity-drift`, `speech-intelligibility-failed`, `narration-performance-failed`, `audio-sync-failed`, and `native-ambience-failed`.
31
31
  8. Keep source and presentation orthogonal on every line: `delivery_mode=native|post_dub|external_audio`, `presentation=visible-dialogue|offscreen-dialogue|narration`, and `fallback_mode=none|post-dub|cinematic-tts|sound-design`. Never put narration or offscreen in delivery_mode.
32
32
  9. Use null for `source_audio` and `voice_binding` when no selected asset or real voice evidence exists. A native-audio fallback requires `native_audio_exception` with an allowed reason, actual evidence, an exact millisecond range, and final `mix_sources`. Offscreen dialogue or narration that was never generated is not a native-audio failure and must keep `native_audio_exception` null.
33
- 9a. `voice_consistency.source_mode` must be one of `first_appearance_sample|user_upload|global_asset|provider_voice_id|none`, taken from the user-confirmed project initialization. `first_appearance_sample` only extracts a real interval from an existing selected video or source audio where the character first appears with clear dialogue, and records `source_asset_key`, `source_version_id`, `start_ms`, `end_ms`, `sha256`, preview evidence, and confirmation; the sample is not a `voice_id`: after confirmation the Provider must create a real `voice_id`, pass capability and listening review, and only then bind it. Without existing source audio, preview, or confirmation put it in unresolved; never infer the voice from an ungenerated first shot. `user_upload` needs an upload receipt and rights confirmation, `global_asset` needs a reviewed global voice ID/version, and `provider_voice_id` needs live catalog evidence; `none` keeps voice_binding null.
33
+ 9a. `voice_consistency.source_mode` must be one of `first_generated_video|first_appearance_sample|user_upload|global_asset|provider_voice_id|none`, taken from the user-confirmed project initialization. `first_appearance_sample` only extracts a real interval from an existing selected video or source audio where the character first appears with clear dialogue, and records `source_asset_key`, `source_version_id`, `start_ms`, `end_ms`, `sha256`, preview evidence, and confirmation; the sample is not a `voice_id`: when the target video model supports reference audio, connect the real selected AUDIO node to later shots through `reference-audios` and record matching nodeId, assetId, versionId and mediaId. Create or select a Provider voice ID only when the target model live schema explicitly accepts that ID; a standalone TTS clone ID is not a video audio reference. For `first_generated_video`, keep the binding pending until a first shot has been generated and selected, then extract a measured speaker interval, listen and bind it to later shots. Without existing source audio, preview, or confirmation put it in unresolved; never invent sources or claim an unfinished binding is complete. `user_upload` needs an upload receipt and rights confirmation, `global_asset` needs a reviewed global voice ID/version, and `provider_voice_id` needs live catalog evidence; `none` keeps voice_binding null.
34
34
  9b. Missing offscreen dialogue uses `delivery_mode=post_dub`, `presentation=offscreen-dialogue`, and a clone of the character's original voice made with the model required by the current audio specification. Its voice binding must include `voice_role=character`, `source=clone`, the real `target_model`, and a real `upload_receipt_id`. Narration uses `delivery_mode=post_dub`, `presentation=narration`, and an independently designed professional narrator voice. Its binding must include `voice_role=narrator`, `source=design`, `professional_narration=true`, a real `request_id`, and a `cinematic_profile` identical to performance. Both routes must bind source `speech-timing`; never estimate timing from text.
35
35
  10. Every narration line requires an executable cinematic `performance` with `tone_arc`, ordered `emotion_beats[]`, `pace`, `breath_and_pause`, and `distance_and_space`; “emotional” alone is insufficient. Non-narration performance is null. Native background sound means ambience and action sound only; BGM remains an independent licensed-music workflow.
36
36
 
@@ -30,7 +30,7 @@
30
30
  7. 新计划固定 `audio_strategy.mode=native-first`、`provider_selection=prefer-native`、`fallback_allowed=true`;`fallback_reasons` 必须完整且仅含 `provider-no-native-audio`、`voice-identity-drift`、`speech-intelligibility-failed`、`narration-performance-failed`、`audio-sync-failed`、`native-ambience-failed`。
31
31
  8. 每行分别填写声音来源 `delivery_mode=native|post_dub|external_audio`、画面呈现 `presentation=visible-dialogue|offscreen-dialogue|narration`、预案 `fallback_mode=none|post-dub|cinematic-tts|sound-design`,不得把 narration/offscreen 填进 delivery_mode。
32
32
  9. `source_audio` 和 `voice_binding` 没有已选资产或真实音色证据时为 null,不得伪造。原生失败兜底时 `native_audio_exception` 必须写受控 reason、实际 evidence、毫秒 range 和最终 `mix_sources`;原本未生成的场外音或旁白不是原生失败,`native_audio_exception` 必须为 null。
33
- 9a. `voice_consistency.source_mode` 只能是 `first_appearance_sample|user_upload|global_asset|provider_voice_id|none`,且必须来自项目初始化时用户确认的配置。`first_appearance_sample` 只从人物第一次出场且有清晰对白的已有视频或原声截取真实区间,写入 `source_asset_key`、`source_version_id`、`start_ms`、`end_ms`、`sha256`、试听证据和确认记录;样本不是 `voice_id`,必须在用户确认后经 Provider 创建真实 `voice_id` 并通过能力检查和试听审核。没有已有来源音频、未展示试听或未确认时写入 unresolved,不得自动绑定或用尚未生成的首镜反推。`user_upload` 必须有上传回执和权利确认,`global_asset` 必须有全局 voice ID/版本和审核证据,`provider_voice_id` 必须有实时目录证据;`none` 时 voice_binding 为 null。
33
+ 9a. `voice_consistency.source_mode` 只能是 `first_generated_video|first_appearance_sample|user_upload|global_asset|provider_voice_id|none`,且必须来自项目初始化时用户确认的配置。`first_appearance_sample` 只从人物第一次出场且有清晰对白的已有视频或原声截取真实区间,写入 `source_asset_key`、`source_version_id`、`start_ms`、`end_ms`、`sha256`、试听证据和确认记录;样本不是 `voice_id`;目标视频模型支持参考音频时,将真实已选 AUDIO 节点接入后续镜头的 `reference-audios`,并保存一致的 nodeId、assetId、versionId、mediaId。只有目标模型实时 schema 明确接受 Provider voice ID 时,才创建或选择对应 ID,不把独立 TTS 克隆 ID 当作视频参考音频。`first_generated_video` 在首镜生成前保留待绑定状态,首镜选版后按实测说话区间取样、试听并绑定后续镜头;没有已有来源音频、未展示试听或未确认时写入 unresolved,不得伪造来源或声称已绑定。`user_upload` 必须有上传回执和权利确认,`global_asset` 必须有全局 voice ID/版本和审核证据,`provider_voice_id` 必须有实时目录证据;`none` 时 voice_binding 为 null。
34
34
  9b. 原本未生成的场外音设 `delivery_mode=post_dub`、`presentation=offscreen-dialogue`,绑定当前声音规范指定模型克隆的角色原声音色;voice_bindings 必须含 `voice_role=character`、`source=clone`、真实 `target_model` 和 `upload_receipt_id`。旁白设 `delivery_mode=post_dub`、`presentation=narration`,使用独立专业设计音色;voice_bindings 必须含 `voice_role=narrator`、`source=design`、`professional_narration=true`、真实 `request_id`,且 `cinematic_profile` 与 performance 逐字段一致。两者都必须先绑定原声 speech-timing,不得按文字估算时间。
35
35
  10. `presentation=narration` 必须写电影感 `performance`:`tone_arc`、按顺序的 `emotion_beats[]`、`pace`、`breath_and_pause`、`distance_and_space`,不能只写“富有情感”;非旁白为 null。背景音原生直出只包括环境声和动作声,BGM 仍由独立授权配乐流程负责。
36
36
 
@@ -4,7 +4,7 @@
4
4
 
5
5
  优先在视频模型支持时保留原生对白、环境声和动作声。逐镜审核语言、说话者、文本、可懂度、时间同步、环境动作声和未声明音乐。失败时记录实际区间与原因;合格部分不重复生成。
6
6
 
7
- 独立语音先用 `voice_analysis` 分离说话人与台词,再按项目初始化时确认的 `audio-plan.voice_consistency` 绑定声音。默认 `first_appearance_sample` 只从人物第一次出场且有清晰对白的已有视频或原声截取实际说话区间,保存来源镜头、起止时间、SHA-256 并先展示试听;样本不是 `voice_id`,用户确认后才调用当前 Provider 的 `voice-clone` 创建真实音色 ID,再试听、审核并绑定。没有已有来源音频时不得用尚未生成的首镜反推音色,必须让用户上传样本或选择 Provider voice。`user_upload` 使用用户上传且已确认有权使用的样本,`global_asset` 使用全局资产库已审核 voice ID,`provider_voice_id` 使用实时 Provider 目录声音,`none` 不建立跨镜绑定。当前 Provider 默认支持已确认 voice ID,只有实时能力明确返回不支持时,才切换支持的 Provider voice、后配音或明确放弃固定音色;不能把未收到不支持结果的 H3 原生音频判为未固定。生成前核对逐字文本和使用权,输出作为未选候选逐条试听、审核后再选版。
7
+ 独立语音先用 `voice_analysis` 分离说话人与台词,再按项目初始化时确认的 `audio-plan.voice_consistency` 绑定声音。默认 `first_generated_video` 先让首次成功生成且含清晰对白的视频建立声音基准;已有来源音频时才使用 `first_appearance_sample` 截取实际说话区间,保存来源镜头、起止时间、SHA-256 并先展示试听;样本不是 `voice_id`:目标视频模型支持参考音频时,确认实际说话区间后按实时目录创建 `video-audio-extract`,填写说话人、起点和时长,连接当前已选视频并执行;回读 AUDIO 结果,将其接入后续镜头的 `reference-audios`,并核对声音计划逐人物引用的节点、资产、版本和媒体。首镜尚未生成时保留待绑定状态,不能伪造音色或自动切为 none。只有实时 schema 明确接受 Provider voice ID 时,才调用其支持的声音创建或选择流程;独立 TTS 的克隆 ID 不能代替视频参考音频。`user_upload` 使用用户上传且已确认有权使用的样本,`global_asset` 使用全局资产库已审核 voice ID,`provider_voice_id` 使用实时 Provider 目录声音,`none` 不建立跨镜绑定。参考音频和 Provider voice ID 的可用性分别按目标模型实时能力核对,不根据 Provider 名称默认支持;不支持时说明限制和复听方案,不能把原生音频直出本身当作已固定音色。生成前核对逐字文本和使用权,输出作为未选候选逐条试听、审核后再选版。
8
8
 
9
9
  CosyVoice 自定义音色设计/克隆只在目标为百炼 CosyVoice、实时模型目录返回该能力且用户确认时,才可在独立 TTS 流程中使用真实音色 ID,并经过试听和审核;它不能静默覆盖 H3 原生对白。StarRouter 没有等价能力;MuseTalk 与旧 RunningHub 口型驱动仍暂时禁用。旧 audio-plan 中的失效配置不得继续提交付费任务;向用户说明缺失能力,并先重新确认当前可用模型和制作方案。不得把临时音色或未通过审核的配音默认为正式结果。
10
10