提示词

nerate a photorealistic talking video from the reference image. The main subject is a middle-aged Armenian man wearing a light beige coat, patterned shirt, and gray hat. Preserve his exact facial identity from the source image. Scene takes place outdoors in a realistic Armenian courtyard/street setting in daylight. He is speaking to a man standing near him. Performance style is very natural, like a scene from a classic Armenian comedy: grounded, believable, warm, slightly ironic. At first he looks at the other man, then while saying “իսկ սա” he raises his hand and points with his index finger toward a refrigerator beside them. The refrigerator must be visible in the shot and clearly be the object he is referring to. His expression shows genuine curiosity and mild confusion, as if asking about electrical voltage compatibility. Use detailed lip sync for Armenian speech, with natural pauses and mouth shapes. Keep motion subtle and realistic, no morphing, no facial distortion, no extra fingers, no broken hands, no flickering. Duration 6–8 seconds. One continuous shot, medium framing, realistic lighting, cinematic quality, high facial detail, stable identity preservation. Spoken line in Armenian: «Արամ ջան, իսկ սա 120 վոլտո՞վ ա աշխատում, թե՞ 220» ⸻ Negative prompt No subtitles, no text, no dubbing mismatch, no extra voices, no background music, no comedy exaggeration, no cartoon style, no face distortion, no flicker, no warped hands, no incorrect lip sync, no sudden camera zoom, no scene change, no low quality blur, no overacting.

生成结果

nerate a photorealistic talking video from the reference image. The main subject is a middle-aged Armenian man wearin...

参数

参考图片