AI art "張國榮《追》(4/6)"
Generate New Artwork

張國榮《追》(4/6)

More to See

Commenti

2persons, 1boy, 1girl, couple, romance, intimate, medium shot, side view, leaning on railing, head on shoulder, closed eyes, half-closed eyes, light smile, windblown hair, urban rooftop balcony, twilight, blue hour, city lights, bokeh, night sky, in-image text, lyric text, painterly, brush texture, color blocks, loose lines, teal and gold A medium side-angle view captures a young man and a young woman leaning side by side against a metal balcony railing high above a glowing metropolis at dusk. The girl leans her temple gently against the boy's shoulder, her eyes closed in serene contentment with a tender, subtle smile across her lips, while her right hand lightly grips the cool iron rail. The boy tilts his chin down toward her, his gaze resting warmly upon her profile, his left forearm resting along the bar beside hers, his fingers curling near her hand without quite touching. In-image typography is integrated cleanly into the composition: vertically aligned in slender, elegant white lettering across the cool upper-left sky reads verbatim "好光陰 縱沒太多 一分鐘那又如何 會與你 共同渡過 都不枉過", while a horizontal subtitle line rendered in delicate glowing script across the lower-right midground reads verbatim "瘋戀多 錯誤更多 如能從新做過 我會說 願能為你 提前做錯". A gentle evening draft lifts fine strands of their dark hair and rustles the hem of the boy's dark coat in the same eastward direction. The scene is immersed in a cool teal-blue twilight atmosphere, struck from below and behind by the warm golden highlights of blurred city bokeh and amber street lamps spilling upward into their silhouettes. Rendered in a soft, painterly anime style with visible brush textures, layered color blocks, and loose, sketch-like linework. Bittersweet and deeply tender.
Mostra parametri
Immissione richiesta originale
好光陰 縱沒太多 一分鐘那又如何 會與你 共同渡過 都不枉過 瘋戀多 錯誤更多 如能從新做過 我會說 願能為你 提前做錯
Modalità
pro
Tag qualità
﹝Prompt﹞
Dimensione
1104 x 1824
Negativo
lowres, worst quality, low quality, bad anatomy, explicit
Helper di Prompt

Modello e LoRA utilizzati