Seedance 2.0
參考圖中的表演者手持無線麥克風,在移動光束與煙霧中現場演唱原創日文感謝歌曲
提示詞
15 秒,16:9,搭配原創日文歌曲的現場演唱會音樂影片。 使用 <Picture 1> 作為表演者身份的嚴格參考。保留 <Picture 1> 中完全一致的臉部、髮型、服裝、比例與風格。 創作一首關於以下主題的原創日文流行歌曲: 「私を見つけてくれてありがとう」 讓歌曲真摯、感性且振奮人心。創作原創日文歌詞、旋律、歌聲與器樂音樂。演唱內容必須清楚使用日文。 保留 <Picture 1> 中表演者原有的視覺質感。如果參考圖是動漫、插畫、Q版、賽璐珞上色或非寫實風格,請讓表演者完整維持相同的繪畫風格。不要將表演者變成寫實或半寫實風格。如果參考圖是寫實風格,請維持表演者的寫實呈現。 建立充滿動感的室內演唱會表演。表演者手持無線麥克風唱歌,自然移動,並與觀眾和鏡頭產生交流。開頭略帶情感,並在結尾變得更加明亮愉快。 只保留一位主要表演者。不要加入其他特邀表演者或複製人。 不要使用靜態鏡頭。每個鏡頭都應有運動:推近、拉遠、側向跟拍、弧線運鏡、低角度英雄鏡頭、舞台廣角鏡頭,以及最後的英雄鏡頭。 使用演唱會燈光、移動光束、煙霧、觀眾燈光與舞台縱深。表演者身後的大型螢幕應顯示表演者正在演唱。 讓影片充滿活力、製作品質精緻,並適合社群媒體發布。 避免複製人重複、身份特徵偏移、服裝變更、臉部偏移、解剖結構錯誤、字幕、歌詞文字、標誌、浮水印、干擾性的使用者介面圖像,以及渲染風格偏移。
15 seconds, 16:9, live concert music video with an original Japanese song. Use <Picture 1> as the strict identity reference for the performer. Preserve the exact face, hairstyle, outfit, proportions, and style from <Picture 1>. Create an original Japanese pop song about: 「私を見つけてくれてありがとう」 Make it sincere, emotional, and uplifting. Compose original Japanese lyrics, melody, vocals, and instrumental music. The singing must be clearly in Japanese. Preserve the performer’s original visual texture from <Picture 1>. If the reference is anime, illustrated, chibi, cel-shaded, or non-photoreal, keep the performer fully in that same drawn style. Do not make the performer photoreal or semi-photoreal. If the reference is photoreal, keep the performer photoreal. Create a dynamic indoor concert performance. The performer sings with a handheld wireless microphone, moves naturally, and connects with the audience and camera. Start slightly emotional and become brighter and more joyful by the end. Keep one main performer only. No extra featured performers or clones. Do not use static shots. Every shot should have motion: push-ins, pull-backs, side tracking, curved moves, low-angle hero shots, wide stage shots, and a final hero shot. Use concert lighting, moving beams, haze, audience lights, and stage depth. The large monitor behind the performer should show the performer singing. Keep the video energetic, polished, and social-media-ready. Avoid clone duplication, identity drift, costume changes, face drift, broken anatomy, subtitles, lyric text, logos, watermarks, intrusive UI graphics, and rendering-style drift.
- 提示詞為日文:直接貼上即可,模型能理解。
此提示詞的權利屬於作者 @otosa_vt。任何商業使用(推廣 AI 工具、聚合平台、Telegram 機器人)由你自行負責,mozgi.io 不承擔任何責任。 詳情 →











