{
  "format": "ai-knowledge-map/tutorial-v1",
  "canonical": "https://ai-knowledge-map.com/software/video-creation/video-generation/",
  "language": "zh-Hans",
  "title": "第一章 专业视频生成平台、基本介绍与教程",
  "summary": "分清视频模型、平台与工作流，了解即梦、通义万相和海螺的教程入口，学习提示词、参考素材、镜头运动及一致性等视频生成控制点。",
  "questions": [
    {
      "question": "AI 视频模型、生成平台和工作流有什么区别？",
      "answer": "模型负责生成能力，平台提供使用入口与操作条件，工作流把素材准备、生成、检查和成片处理连接起来。",
      "sections": [
        {
          "id": "course-video-generation-section-1",
          "title": "1. 先分清模型、平台和工作流",
          "url": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-1"
        }
      ]
    },
    {
      "question": "怎样从即梦、通义万相和海螺的教程开始学习？",
      "answer": "本章按即梦、通义万相和海螺分别给出官方教程入口；按所用平台与模型版本选择资料，不把学习顺序当作统一性能排名。",
      "sections": [
        {
          "id": "course-video-generation-section-2",
          "title": "2. 三个生成平台",
          "url": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-2"
        },
        {
          "id": "course-video-generation-section-3",
          "title": "3. 三个平台怎样开始学习",
          "url": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-3"
        }
      ]
    },
    {
      "question": "生成视频时，应该怎样控制内容、镜头和一致性？",
      "answer": "先明确动作和镜头目标，再给参考素材分工，安排首尾帧与过渡，并逐段检查主体、镜头和一致性。",
      "sections": [
        {
          "id": "course-video-generation-section-4",
          "title": "4. 掌握生成视频的几个控制点",
          "url": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-4"
        }
      ]
    }
  ],
  "scope": "摘要与短答依据正式课程整理。以下完整章节保留原始来源名称、配置细节、适用版本与实测限制。此文件为公开阅读格式，不代表独立审核或实测结果。",
  "sources": [
    {
      "url": "https://jimeng.jianying.com/",
      "title": "即梦入口",
      "section": "course-video-generation-section-2",
      "context": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-2"
    },
    {
      "url": "https://seed.bytedance.com/en/seedance2_5",
      "title": "模型介绍",
      "section": "course-video-generation-section-2",
      "context": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-2"
    },
    {
      "url": "https://docs.volcengine.com/docs/ark/seedance-2-5-prompt-guide?lang=zh",
      "title": "官方提示词教程",
      "section": "course-video-generation-section-2",
      "context": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-2"
    },
    {
      "url": "https://www.alibabacloud.com/blog/wan3-0-30-second-ai-video-generation-from-any-input_603452",
      "title": "Alibaba Cloud 官方模型介绍与入口",
      "section": "course-video-generation-section-2",
      "context": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-2"
    },
    {
      "url": "https://hailuoai.video/",
      "title": "海螺入口",
      "section": "course-video-generation-section-2",
      "context": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-2"
    },
    {
      "url": "https://platform.minimax.io/docs/guides/video-generation",
      "title": "MiniMax 官方视频教程",
      "section": "course-video-generation-section-2",
      "context": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-2"
    },
    {
      "url": "https://xyq.jianying.com/tutorials/seedance-2-5",
      "title": "小云雀 Seedance 2.5 官方操作手册（小云雀界面）",
      "section": "course-video-generation-section-3",
      "context": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-3"
    },
    {
      "url": "https://help.aliyun.com/zh/model-studio/wan3-video-generation-prompt-guide",
      "title": "万相 3.0 官方视频生成 Prompt 指南（中文）",
      "section": "course-video-generation-section-3",
      "context": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-3"
    },
    {
      "url": "https://help.aliyun.com/zh/model-studio/wan3-video-generation-guide",
      "title": "万相 3.0 官方视频生成使用指南（API）",
      "section": "course-video-generation-section-3",
      "context": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-3"
    },
    {
      "url": "https://design.minimax.io/h3",
      "title": "MiniMax H3 官方教程与工作流中心",
      "section": "course-video-generation-section-3",
      "context": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-3"
    },
    {
      "url": "https://helpx.adobe.com/firefly/web/work-with-audio-and-video/work-with-video/writing-effective-text-prompts-for-video-generation.html",
      "title": "Adobe 提示词教程",
      "section": "course-video-generation-section-4",
      "context": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-4"
    },
    {
      "url": "https://support.google.com/flow/answer/16353334?hl=en",
      "title": "Flow 首尾帧说明",
      "section": "course-video-generation-section-4",
      "context": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-4"
    }
  ],
  "sections": [
    {
      "id": "course-video-generation-introduction",
      "title": "阅读说明",
      "url": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-introduction",
      "html": "<p>本板块面向已经会通过自然语言表达需求、希望进一步控制视频结果的读者。第一章认识专业生成平台与教程；第二章学习生成大致内容后的精细处理，以及让 AI 直接驱动工具的方法；第三章提供新人可以跟做的现成工作流。</p>\n<p>资料核对日期：2026 年 10 月 31 日。本文已完成资料核对与三章正文编写；接入和生成步骤尚未完成本机软件操作、付费调用或效果验收。第三章收录原作者公开的案例与教程，作者演示与本站实际复现分别说明。调研取舍与工作流文件记录见 [资料核对记录](VIDEO_CREATION_RESEARCH.md)。</p>"
    },
    {
      "id": "course-video-generation-section-1",
      "title": "1. 先分清模型、平台和工作流",
      "url": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-1",
      "html": "<table class=\"d-table\"><thead><tr><th>输入与任务</th><th>主要控制什么</th><th>首次练习可以怎样做</th></tr></thead><tbody><tr><td>文字生成视频</td><td>主体、动作、场景与摄影方式</td><td>一个主体、一个动作、一个短镜头</td></tr><tr><td>图片生成视频</td><td>初始外观与构图，再补充运动</td><td>用产品或场景图生成缓慢运动</td></tr><tr><td>首尾帧生成</td><td>起点、终点及两者之间的变化</td><td>同一对象的两种相近状态</td></tr><tr><td>多素材参考</td><td>人物、道具、动作、镜头或声音</td><td>给每份素材指定一种用途</td></tr><tr><td>多片段成片</td><td>顺序、节奏、转场及音画衔接</td><td>把三个通过检查的镜头组成短片</td></tr></tbody></table>"
    },
    {
      "id": "course-video-generation-section-2",
      "title": "2. 三个生成平台",
      "url": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-2",
      "html": "<p>结合当前可查的独立评测、创作控制与课程用途，第一章只重点介绍下面三条路线。推荐顺序是课程的学习安排；文生、图生与编辑有不同榜单，不把它们合成未经定义的统一总排名。</p>\n<table class=\"d-table\"><thead><tr><th>平台与重点模型</th><th>推荐理由与适用任务</th><th>官方入口与教程</th></tr></thead><tbody><tr><td>即梦／Dreamina：Seedance 2.5</td><td>综合首选，适合从参考图出发制作多镜头内容，并继续做指令编辑。所查 Artificial Analysis 图生视频榜第 1、视频编辑榜第 2</td><td><a href=\"https://jimeng.jianying.com/\" target=\"_blank\" rel=\"noopener noreferrer\">即梦入口</a>、<a href=\"https://seed.bytedance.com/en/seedance2_5\" target=\"_blank\" rel=\"noopener noreferrer\">模型介绍</a>、<a href=\"https://docs.volcengine.com/docs/ark/seedance-2-5-prompt-guide?lang=zh\" target=\"_blank\" rel=\"noopener noreferrer\">官方提示词教程</a></td></tr><tr><td>通义万相：Wan 3.0</td><td>重点学习从文字生成镜头及后续编辑。所查 Artificial Analysis 文生视频与视频编辑榜均第 1</td><td><a href=\"https://www.alibabacloud.com/blog/wan3-0-30-second-ai-video-generation-from-any-input_603452\" target=\"_blank\" rel=\"noopener noreferrer\">Alibaba Cloud 官方模型介绍与入口</a></td></tr><tr><td>海螺：MiniMax H3</td><td>重点学习图片生成视频与多模态参考。所查 Arena 图生视频榜第 1；Artificial Analysis 文生视频榜第 3</td><td><a href=\"https://hailuoai.video/\" target=\"_blank\" rel=\"noopener noreferrer\">海螺入口</a>、<a href=\"https://platform.minimax.io/docs/guides/video-generation\" target=\"_blank\" rel=\"noopener noreferrer\">MiniMax 官方视频教程</a></td></tr></tbody></table>"
    },
    {
      "id": "course-video-generation-section-3",
      "title": "3. 三个平台怎样开始学习",
      "url": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-3",
      "html": "<table class=\"d-table\"><thead><tr><th>平台</th><th>首选教程</th><th>补充教程</th></tr></thead><tbody><tr><td>即梦／Dreamina：Seedance 2.5</td><td><a href=\"https://docs.volcengine.com/docs/ark/seedance-2-5-prompt-guide?lang=zh\" target=\"_blank\" rel=\"noopener noreferrer\">Seedance 2.5 官方提示词指南（中文）</a></td><td><a href=\"https://xyq.jianying.com/tutorials/seedance-2-5\" target=\"_blank\" rel=\"noopener noreferrer\">小云雀 Seedance 2.5 官方操作手册（小云雀界面）</a></td></tr><tr><td>通义万相：Wan 3.0</td><td><a href=\"https://help.aliyun.com/zh/model-studio/wan3-video-generation-prompt-guide\" target=\"_blank\" rel=\"noopener noreferrer\">万相 3.0 官方视频生成 Prompt 指南（中文）</a></td><td><a href=\"https://help.aliyun.com/zh/model-studio/wan3-video-generation-guide\" target=\"_blank\" rel=\"noopener noreferrer\">万相 3.0 官方视频生成使用指南（API）</a></td></tr><tr><td>海螺：MiniMax H3</td><td><a href=\"https://design.minimax.io/h3\" target=\"_blank\" rel=\"noopener noreferrer\">MiniMax H3 官方教程与工作流中心</a></td><td><a href=\"https://platform.minimax.io/docs/guides/video-generation\" target=\"_blank\" rel=\"noopener noreferrer\">MiniMax H3 官方视频生成指南（API）</a></td></tr></tbody></table>"
    },
    {
      "id": "course-video-generation-section-4",
      "title": "4. 掌握生成视频的几个控制点",
      "url": "https://ai-knowledge-map.com/software/video-creation/video-generation/#course-video-generation-section-4",
      "html": "<p><strong>先控制动作，再增加复杂度。</strong> 对一个短镜头，写清谁在做什么、动作怎样开始和结束、摄影机是否移动。画面已有主体时，重复堆叠外观形容词可能使真正重要的动作要求不够突出。需要多个镜头时，把每个镜头的作用分别写清；可参考 <a href=\"https://helpx.adobe.com/firefly/web/work-with-audio-and-video/work-with-video/writing-effective-text-prompts-for-video-generation.html\" target=\"_blank\" rel=\"noopener noreferrer\">Adobe 提示词教程</a>。</p>\n<p><strong>给参考素材分工。</strong> 人物图负责身份，产品图负责外观，视频负责动作或运镜，声音负责台词或节奏。多份参考互相冲突时，先减少素材，再检查结果。参考是一种控制条件，输出仍需逐帧或逐段检查。</p>\n<p><strong>首尾帧限定两端，过程仍需设计。</strong> 起点是关着的盒子、终点是打开的盒子，提示词应描述打开过程；如果两张图的场景、视角和对象差异很大，模型需要自行补出更多变化。先用画幅和构图接近的两张图练习。<a href=\"https://support.google.com/flow/answer/16353334?hl=en\" target=\"_blank\" rel=\"noopener noreferrer\">Flow 首尾帧说明</a></p>"
    }
  ]
}