[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"product:omniflash":3},{"id":4,"name":5,"slug":6,"slogan":7,"description":8,"organization":9,"images":10,"avatar_url":11,"image_url":12,"website_url":13,"access_links":14,"status":19,"build_by":20,"categories":21,"stats":25,"upload_relationship":29,"create_at":30,"update_at":31,"owner_id":26,"upload_user_id":32,"reviewer_id":33,"video_info":34,"prerelease_cover_url":35,"scheduled_publish_at":34,"user_scheduled":36,"emoji_reacts":37,"tag":34,"pass_review":34,"user_star":36},2144,"Omni Flash","omniflash","由 Google 推出、登顶 LMarena 的全模态 4K AI 视频生成与对话式连贯剪辑创意工坊。","{\"type\": \"doc\", \"content\": [{\"type\": \"paragraph\", \"content\": [{\"text\": \"【产品介绍】 \", \"type\": \"text\"}]}, {\"type\": \"paragraph\", \"content\": [{\"text\": \"Omni Flash（全称 Gemini Omni Flash，由 Google DeepMind 研发）是一款颠覆传统分步渲染逻辑的革命性全模态 AI 视频生成与对话式剪辑平台。 \", \"type\": \"text\"}]}, {\"type\": \"paragraph\", \"content\": [{\"text\": \"作为 Google 推出的全新一代多模态旗舰媒体模型，Omni Flash 旨在取代创作者手动拼凑的零散工具链，在 Lmarena (AILab) 的文生视频竞技场中登顶榜首。 \", \"type\": \"text\"}]}, {\"type\": \"paragraph\", \"content\": [{\"text\": \"该平台核心采用“单一多模态引擎”设计，能一次性读取文本、图像和音频，同步输出带有原生同步音频、完美口型同步（唇形同步）以及锁定角色一致性的影院级 4K 视频。 系统打破了“单次生成、不满意就重来”的行业痛点，首创“状态化\u002F对话式剪辑”机制，让创作者能够通过日常聊天直接微调焦距、镜头或局部细节。 产品无缝集成于 Google Flow 创意工作室、Gemini 官方应用以及 YouTube Shorts 矩阵中，为独立导演、营销团队及品牌电商家提供了一站式、电影级连续性创作的极致生产力工具。\", \"type\": \"text\"}]}, {\"type\": \"paragraph\"}, {\"type\": \"paragraph\", \"content\": [{\"text\": \"【产品功能】\", \"type\": \"text\"}]}, {\"type\": \"bulletList\", \"content\": [{\"type\": \"listItem\", \"content\": [{\"type\": \"paragraph\", \"content\": [{\"text\": \"统一多模态一次性渲染 单一引擎同时读取文字、画面与声音，将构图运动、人物台词及环境配乐在一秒内同步规划并融为一体，生成无中间解耦的高连贯场景。\", \"type\": \"text\"}]}]}, {\"type\": \"listItem\", \"content\": [{\"type\": \"paragraph\", \"content\": [{\"text\": \"电影级视听语言镜头控制 深度理解电影语法，支持通过“跟拍、推拉变焦、动机性布光及三分法构图”等专业摄影术语直接指挥大模型，精准调控景深与焦距。\", \"type\": \"text\"}]}]}, {\"type\": \"listItem\", \"content\": [{\"type\": \"paragraph\", \"content\": [{\"text\": \"跨镜头锁定角色连续性 支持多达 9 张图像\u002F视频参考锚定，用户只需一次锁定人物面部、服装或特定品牌资产，系统即可在后续的所有批次和多镜头切中保持视觉连续性。\", \"type\": \"text\"}]}]}, {\"type\": \"listItem\", \"content\": [{\"type\": \"paragraph\", \"content\": [{\"text\": \"状态化对话式流式剪辑 支持像与真人剪辑师沟通一样，通过自然语言对话（如“特写镜头多停留两拍”）直接局部改写或重写视频片段，无需重新渲染整个场景。\", \"type\": \"text\"}]}]}, {\"type\": \"listItem\", \"content\": [{\"type\": \"paragraph\", \"content\": [{\"text\": \"一站式全球发布生态物料 一键导出商业授权、无水印且带来源元数据的 4K 全清母版，原生集成于 YouTube Shorts 等分发渠道，提供极速预览反馈与海外多语种本地化配音替换。\", \"type\": \"text\"}]}]}]}, {\"type\": \"paragraph\"}]}","Google","https:\u002F\u002Fwatcha.tos-cn-beijing.volces.com\u002Fprod\u002Fuser\u002Fuploads\u002F10009357_1780558719_b4e21851-9bbf-498f-9ae1-7eeab1340554.png;https:\u002F\u002Fwatcha.tos-cn-beijing.volces.com\u002Fprod\u002Fuser\u002Fuploads\u002F10009357_1780558719_521183f9-3999-4274-993a-3c7789f1ed0e.png;https:\u002F\u002Fwatcha.tos-cn-beijing.volces.com\u002Fprod\u002Fuser\u002Fuploads\u002F10009357_1780558719_2c3467ce-c12b-483b-a547-fff02f43a744.png","https:\u002F\u002Fwatcha.tos-cn-beijing.volces.com\u002Fprod\u002Fuser\u002Fuploads\u002F10009357_1780558683_4fa0aa04-6a82-463b-abfd-5ab9e181f8e6.png","https:\u002F\u002Fwatcha.tos-cn-beijing.volces.com\u002Fprod\u002Fuser\u002Fuploads\u002F10009357_1780558692_2592f2b8-0a7b-49d6-8735-f5b174ce244c.png","https:\u002F\u002Fomniflash.ai\u002Fzh",{"items":15},[16],{"platform":17,"url":13,"is_primary":18},"WEB",true,"PUBLISHED",[],[22],{"id":23,"name":24},4,"视频创作",{"upvotes":26,"stars":27,"review_count":27,"reply_count":26,"update_at":28},0,1,"2026-09-08T07:03:24.982Z","THIRD_PARTY","2026-06-04T07:38:51.014Z","2026-06-05T06:03:36.348Z",10009357,10004627,null,"",false,{"reacts":38},[]]