Cinematic 14-second anime sequence. An autumn night… | Deepfake AI Video

Text-to-video prompt used to generate this video

Cinematic 14-second anime sequence. An autumn night festival on dry packed-earth streets. Paper lanterns, stall banners, yukata and casual coats. Multiple camera setups, ending on a locked beauty frame. Clear short dialogue. 24fps. 0–2s: Extreme wide establishing shot. Dry shrine approach, lantern rows, stall roofs, a taiko stand. Crowd moving. Slow push-in. 2–5s: Medium tracking. A girl in a patterned yukata and a boy in a light haori walk the lane. Camera orbits front to side. Dialogue: Her: “Don’t lose me in the stalls.” Him: “Then take my sleeve. I’m not going far.” 5–9s: Dynamic orbit. Short beat: they stop at a candy-apple stall (solid candy), then at a game stall. Lantern light on faces. Dry ground only. 9–12s: Close two-shot. Sparklers off-frame as dry light, not rain. They grin. 12–14s: Perfect anime shot (hold). Medium-wide. Pair under a lantern arch, festival bokeh behind them. Locked frame. Soft fade. Style: festival anime, autumn matsuri, multiple shots ending on a key visual, dry night streets only, masterpiece, best quality, ultra high res, hyper-detailed, no text on screen, no logos.