云端工作流效果与耗时对比

subject_definitions:
<Subject 1> is the adult man from <Picture 1>. Preserve his facial identity, hairstyle and clothing. Use the reference as an identity sheet and render one person in a single scene.
<Subject 2> is the adult woman from <Picture 2>. Preserve her facial identity, hairstyle and clothing. Use the reference as an identity sheet and render one person in a single scene.

summary:
[reference generation] A five-second continuous cinematic medium two-shot of the same two people in a quiet sunlit courtyard, with separate male and female off-screen narration followed by one short on-camera line.

retention_analysis:
<Subject 1>: fully_preserved - retain the same face, hair and clothing from <Picture 1>.
<Subject 2>: fully_preserved - retain the same face, hair and clothing from <Picture 2>.

detailed_description:
Live-action cinematic photography with soft daylight and natural skin texture. [Shot 1] A single continuous eye-level medium two-shot. <Subject 1> stands on the left and <Subject 2> stands on the right in a peaceful courtyard. The camera slowly pushes forward. At 00:00.200, a male narrator (S4) says in an off-screen voiceover, <d>[Chinese] 故事开始了。</d> At 00:02.000, a female narrator (S5) says in an off-screen voiceover, <d>[Chinese] 我们会再见。</d> Both visible people remain silent with their lips closed during these two voiceovers. At 00:03.700, <Subject 1> (S1) looks toward <Subject 2> and says gently, <d>[Chinese] 明天见。</d> Only his lips move for his own line. The woman gives a small natural smile. The image is a clean cinematic frame with no on-screen text.

overall_soundscape:
Gentle outdoor ambience. S4 is a male off-screen voice; S5 is a female off-screen voice. They are separate from the visible man's S1 on-camera speech. Each voice is clear, sequential and never overlaps another line.

non_diegetic_music:
N/A