Manish Kumar

6 posts

Manish Kumar

Manish Kumar

@laiho249

Talks about AI & all things future

Katılım Kasım 2023
38 Takip Edilen0 Takipçiler
Manish Kumar
Manish Kumar@laiho249·
@techhalla Further, I think you can also animate the 3D asset. Would be super cool! But overall loved the idea.
English
1
0
1
40
TechHalla
TechHalla@techhalla·
this is the future of education teachers vibe coding interactive apps with custom 3D models instead of boring slides. Make your own assets with Tripo this way 👇
English
22
61
586
43.3K
Lia
Lia@liajunx·
Step 1: paste this prompt in chat gpt [SUBJECT PRESERVATION] Use the uploaded image as the identity reference for the main subject. Preserve overall identity including facial structure, proportions, skin tone, and hairstyle. Identity must remain consistent, but allow minor natural variation caused by lighting, camera angle, and broadcast capture conditions. Do not transform the subject into a different person. [EXPRESSION] Maintain the same emotional state as the reference image. However, change gaze direction: the subject is NOT looking at the camera. The subject is looking slightly to the side, focused on the baseball game happening on the field. Expression should feel natural and candid, consistent with watching live play in a stadium. Allow subtle micro-variation consistent with broadcast sports footage. [OUTFIT] The subject is wearing the “[팀 이름]” KBO jersey. All jersey design elements (colors, logo placement, typography, and patterns) must match the official uniform exactly. Do not redesign, reinterpret, or simplify the jersey. Allow natural fabric folds and distortion due to sitting posture. [SCENE] A realistic KBO professional baseball stadium broadcast capture. The subject is seated in spectator stands during a live game. Posture: upright seated posture, back naturally straight. Arms resting comfortably close to the body. Natural, candid audience moment. [POSTURE CONTROL – IMPORTANT] The overall posture should feel natural, but maintain a clean and composed appearance as seen in broadcast footage. [ENVIRONMENT] Stadium seating is fully occupied with spectators. Crowd behavior includes cheering, filming with smartphones, holding clappers and fan merchandise. Background spectators must be out of focus or blurred for anonymity. Only the main subject remains clearly visible. The crowd should feel dense, dynamic, and slightly imperfect like real broadcast footage. [LIGHTING] Night-time stadium floodlights from above. Uneven overhead illumination with realistic exposure variation. Natural shadows under eyes, nose, and chin. No studio lighting or beauty lighting effects. [CAMERA / STYLE] Medium close-up (waist-up framing). Shot on an 85mm lens. Slight 3/4 side angle (not front-facing). Subject is oriented toward the field, not toward the camera. Head turned slightly toward the field. Candid broadcast perspective, as if captured unintentionally during live TV coverage. [VISUAL STYLE] Broadcast screenshot aesthetic: mild motion blur light compression artifacts pixel noise imperfect sharpness The image should feel like a paused TV broadcast frame, not a cinematic or studio-quality photo. [ASPECT RATIO] 16:9 horizontal widescreen only. Do not output vertical or square framing. [OVERLAY] Include a minimal KBO broadcast scoreboard at the top. Clean and realistic TV graphics style. Do not obstruct the subject’s face. [FINAL CONSTRAINT] The subject must remain clearly recognizable as the same person from the input image. No identity shift or face redesign. Natural variation from broadcast conditions is allowed for realism.
English
2
0
17
6.4K
Krea
Krea@krea_ai·
Krea Video is here! this is how it works 👇 - (sound on)
English
3.5K
392
2.7K
406.4K