Case Snapshot

This 9.5-second vertical clip is a practical WAN 2.2 Animate dance-copy test built around one important constraint: keep the performer close to camera. A glamorous woman in a dark velvet slit dress and oversized blue fur hat dances against a snowy night background while a left-side reference strip shows the portrait-plus-motion input setup. The choreography stays mostly in the upper body with sways, crossed arms, and torso twists. That is not an accident. It is exactly the kind of framing that gives the model the best chance of preserving the face.

What You're Seeing

The left-side strip makes the workflow transparent

The source portrait, motion reference, plus sign, arrow, and WAN 22 label turn the video into a readable test, not just an aesthetic dance clip.

The blue fur hat is a strong identity anchor

It helps the subject stay memorable and also makes continuity errors easier to notice.

The dance is intentionally medium-close

This is the key lesson from the creator's caption: if the subject stays near the camera, the face holds together better.

The background is decorative but simple

Snow-like particles and cool blue darkness add atmosphere without competing with the dancer.

The dress movement stays manageable

The slit dress adds visual interest, but the choreography avoids footwork-heavy patterns that would expose more body-consistency failures.

Shot-by-shot Breakdown

Why It Works

It respects the model's current limits

The creator does not ask WAN 2.2 to solve full-body viral choreography from a distant camera. She keeps the subject close and the motion readable.

The outfit and hat make the clip visually distinctive

That helps the sample stand out from generic dance tests while still serving as a technical benchmark.

The reference strip increases trust

Viewers can immediately understand that the result came from a source image plus dance reference rather than from manual editing or pure guessing.

The Wan 2.2 Rule

Close to camera means better face consistency

This clip directly supports the creator's stated conclusion: near-camera dance motion holds the face together better than distant full-body movement.

Upper-body dance is the sweet spot

Arm crossings, shoulder sways, and torso twists are demanding enough to feel like a dance copy, but they do not overload the model the way footwork-heavy choreographies do.

The model is useful for social reels before it is useful for perfect dance films

This is strong enough for creator content, mood reels, and experiments, even if it is not production-perfect.

Prompt Breakdown

The reference strip must stay visible

It is not decoration. It is the explanatory engine of the clip.

The character design should include one strong accessory

The oversized blue fur hat gives the audience an easy continuity marker.

The background should support, not dominate

Snowy particles and dark blue atmosphere provide style without distracting from the dance-copy test.

How to Recreate It

Step 1: Choose a near-camera dance reference

Do not start with a far-away TikTok dance if your goal is face stability.

Step 2: Build one strong character look

A memorable hat, dress, or accessory makes identity drift easier to spot and stronger when preserved.

Step 3: Keep the choreography mostly above the waist

Upper-body-driven dances are a better fit for current copy-motion quality.

Step 4: Show the inputs on screen

Reference transparency helps the content perform as both tutorial and proof.

Step 5: End before the movement complexity outruns the model

Short dance tests are stronger when they stop while the face still looks coherent.

Growth Playbook

3 opening hook lines

• If you want AI dance copy to work, keep the dancer close to the camera.
• This WAN 2.2 test proves face consistency survives much better in medium-close choreography.
• The mistake most people make is choosing dances that are too far away and too full-body.

4 caption templates

1. Hook: "Copy viral dances with AI." Value: "This WAN 2.2 Animate sample works better because the movement stays close to camera and mostly in the upper body." Question: "Do you want the link?" CTA: "Comment ARIA."
2. Hook: "The face breaks much less when the dancer is in the foreground." Value: "That is the biggest lesson from these tests, more important than the outfit or background." Question: "What dance style should I test next?" CTA: "Write ARIA below."
3. Hook: "This is one of the cleaner WAN 2.2 dance-copy examples I've seen." Value: "The trick is respecting the current limits instead of forcing impossible choreography." Question: "Would you use this for social content?" CTA: "Type ARIA."
4. Hook: "AI dance is still imperfect, but this setup gets much closer." Value: "Near-camera arm and torso moves are far more reliable than distant footwork-heavy references." Question: "Should I compare it to other models?" CTA: "Comment ARIA."

Hashtag strategy

Broad: #AIVideo #DanceAI #WAN22 #MotionCopy. These support broad discovery.

Mid-tier: #WAN22Animate #AIDanceTest #DancePrompt #ReferenceToVideo. These fit the actual workflow more closely.

Niche long-tail: #ViralDanceCopyAI #WAN22DanceCopy #CloseUpDancePrompt #AIDanceConsistency. These target viewers searching for this exact dance-copy use case.

FAQ