Kling 2.6 Motion Control Tests 🎬
Os dejo algunas pruebas que hice con Kling... La verdad es que sigo pensando que le queda mucho trabajo para llegar a un resultado decente 👀
No hay consistencia en ningún momento, la cara se deforma con cada segundo que pasa por no decir que de cada 10 videos que intento generar, 8 de ellos me dan error 🥲
De momento sirve para hacer videos graciosos para internet o las redes sociales pero en ningún caso para un proyecto profesional 😅
Igualmente, si quieres que te mande los vídeos de referencia que usé para hacer estos vídeos comenta "ARIA" y te los mando por mensajes 💌
How soy_aria_cruz Made This Funny Face Motion Control AI Video - and How to Recreate It
This case study examines a high-engagement AI video test by creator @soy_aria_cruz, featuring a cinematic yet playful editorial portrait. The video showcases a young woman with dark hair and large round glasses performing a "crossed-eye" funny face. Utilizing Kling 2.6's motion control, the content bridges the gap between high-end AI generation and relatable, UGC-style (User Generated Content) humor. The warm, natural bedroom lighting and sharp focus on facial micro-expressions create an uncanny but captivating loop that encourages viewers to interact with the creator's automation funnel.
What You’re Seeing
The video is a tight close-up (CU) of a young woman. She has dark hair pulled back, olive skin, and is wearing distinctive large, round, silver-rimmed glasses and thin hoop earrings. The background is a softly blurred domestic interior, suggesting a bedroom or home office, with natural light coming from a window to the side, creating soft highlights on her face. Her primary action is a playful, exaggerated expression: she crosses her eyes while maintaining a tight, slightly amused smile. The motion is subtle, focusing on the eye movement and slight facial muscle shifts, which tests the limits of AI temporal consistency.
Shot-by-Shot Breakdown
Time Range
Visual Content
Shot Language
Lighting & Tone
Viewer Intent
00:00 – 00:04
Woman crosses her eyes and smiles at the camera.
Static Close-up (CU), 50mm feel.
Warm, natural side-lighting; soft bokeh.
Hook: Surprise/Humor through an "imperfect" AI expression.
Why It Went Viral
The success of this video lies in its subversion of the "Perfect AI" trope. Most AI-generated influencers or videos strive for flawless, "model-like" beauty. By intentionally generating a "funny face" (crossed eyes), the creator taps into biological curiosity and humor. It feels more human and less like a bot, which builds immediate trust and relatability with the audience. Furthermore, the video serves as a "tech demo," attracting other creators interested in the latest AI tools like Kling 2.6.
From a platform perspective, the video uses a high-friction engagement trigger. The text overlay "Comenta: ARIA" (Comment: ARIA) combined with a promise of a link creates a surge in comments. Instagram's algorithm interprets this high comment-to-view ratio as a signal of extreme relevance, pushing the video to a wider "Explore" audience. The short, looping nature of the video also ensures high "Watch Time" percentages, as viewers often watch it twice to see if the AI "glitches" during the eye movement.
5 Viral Hypotheses
The "Imperfection" Hook: AI that looks "too perfect" is often scrolled past. Hypothesis: Intentionally creating "ugly" or "silly" expressions increases stop-rate because it looks more like a real person's blooper.
The Tech-Curiosity Loop: Mentioning specific tools (Kling 2.6) in the caption targets a niche but highly active community of AI enthusiasts who save and share for reference.
The Direct CTA: Using a specific keyword for comments ("ARIA") simplifies the user's task, leading to higher conversion than a generic "comment below."
High Contrast Lighting: The use of "Rembrandt-style" side lighting makes the AI generation look more like professional cinematography, increasing the perceived value of the content.
The "Uncanny Valley" Bridge: By successfully animating eyes (the hardest part of AI), the video creates a "How did they do that?" moment that drives repeat views.
How to Recreate (Step-by-Step)
Character Definition: Use a consistent prompt or a reference image (IP-Adapter) to define your character. Focus on unique identifiers like "large round wire-frame glasses" and "thin hoop earrings."
Platform Selection: Use Kling AI (v2.6) or Luma Dream Machine for high-fidelity facial motion.
The "Funny Face" Prompt: Specifically prompt for the expression. Example: "A young woman looking at the camera, playfully crossing her eyes, tight smile, goofy expression."
Motion Control: In Kling, use the "Motion Brush" or "Camera Control" to ensure the head stays relatively static while the eyes move inward.
Lighting Setup: Specify "natural window light from the side, soft shadows, cinematic indoor photography."
Text Overlay: Use a clean, sans-serif font. Place your CTA (e.g., "Comment: [KEYWORD]") in the lower third of the frame to avoid covering the face.
Engagement Automation: Set up a tool like ManyChat to automatically DM the link to anyone who comments the keyword.
Publishing: Post as a Reel with a trending but low-volume "Aesthetic" or "Lo-fi" audio track to keep the focus on the visuals.
Growth Playbook
Opening Hook Lines
"Is AI finally getting the 'human' touch?"
"Testing Kling 2.6 motion control: The results are... weird."
"Stop making perfect AI. Make it real instead."
Caption Templates
Option 1: The Tech Review
Testing out the new Kling 2.6 motion brush! 🎬 I wanted to see if it could handle a non-standard expression like this. Consistency is still a challenge, but we're getting closer. Comment 'TECH' if you want my AI workflow! 👇
Option 2: The Engagement Funnel
AI or Real? You decide. 👁️ I’m sharing the exact prompts I used to get this level of detail. Comment 'ARIA' and I’ll send the link to your DMs! 🚀 #AIVideo #KlingAI