I looove her songs!!! thx @oshuclips for stopping me 🎧⚡️❤️ #oshuclips

imma

@imma.gram · Digital creator

INSTAGRAM · 2024-07-12Source

-1likes

25comments

Remix This

Prompt

HIGH-GRANULARITY VISUAL INVENTORY
- Subject(s): split-screen meme/reaction layout. Left panel shows virtual influencer close-up with pink bob and surprised open-mouth expression. Right panel shows two people outdoors in park; one interviewing another wearing headphones. Text overlay on right reads “LISTENING TO?” in yellow with black outline.
- Clothing/materials: left subject in dark top with “VIRTUAL HUMAN” print. Right panel has casual summer clothing (light tees/shorts, red cap).
- Environment: left indoor neutral background; right sunny grassy park with trees.
- Composition: 9:16 vertical with black margins and two equal side-by-side panels inside; high contrast between controlled portrait and candid street-interview clip.
- Lighting: left soft indoor light; right bright daylight.
- Palette: pink hair accent, green park tones, yellow text overlay.
- Style: social-media remix frame / reaction content.

MASTER PROMPT
[Subject] A vertical split-screen social clip frame: left side is a pink-bob virtual influencer reacting with surprised open mouth; right side is a street-interview moment in a park with two people, one holding a mic and one wearing headphones.
[Environment] Left indoor neutral backdrop, right outdoor sunny grass and trees.
[Composition/Camera] 9:16 canvas with black surrounding margins and two side-by-side panels; add yellow caption text on right panel: “LISTENING TO?”.
[Lighting] Soft indoor portrait lighting on left; bright natural daylight on right.
[Style/Rendering] Viral short-form remix aesthetic, candid and energetic.
[Detail constraints] keep split layout, keep pink-hair reaction face, keep park interview scene and yellow caption text.

NEGATIVE PROMPT
single full-frame portrait, no split, no text overlay, dark night park, studio interview set, anime style, painterly render, low-res blur, missing headphones or mic cues.

SUGGESTED PARAMETERS
- AR 9:16
- Lens mix: left 50mm portrait, right 28-35mm candid
- Steps 24-34
- CFG 5.0-6.3
- Sampler DPM++ 2M Karras
- Seed 873440291

DELTA PROMPT STRATEGY
1) No split -> "strict two-panel side-by-side layout"
2) Left expression wrong -> "surprised open-mouth reaction"
3) Hair wrong -> "short pink blunt bob"
4) Right scene wrong -> "sunny park interview with mic and headphones"
5) Text missing -> "yellow outlined text: LISTENING TO?"
6) No black frame margins -> "social-clip style black padding around panels"
7) Lighting mismatch -> "indoor soft left, outdoor bright right"
8) Panels uneven -> "balanced vertical split with similar panel widths"
9) No meme vibe -> "casual reactive social-media remix tone"
10) Over-polished -> "raw screenshot-like capture quality"

The Music Street Interview: How imma.gram Built This AI Art

This frame blends two familiar social formats: reaction face and street interview. The combination is fast to understand and naturally comment-friendly.

Why it went viral

The first hook is contrast: controlled avatar reaction on one side, unpredictable real-world clip on the other. This creates instant narrative tension.

The second hook is clear text prompt. “LISTENING TO?” gives viewers a direct participation cue and invites guesses or song recommendations.

Signal	Evidence (from this image)	Mechanism	Replication Action
Dual-format familiarity	reaction face + street interview	low comprehension friction	Combine two known content grammars
Prompt text CTA	“LISTENING TO?” overlay	drives comments	Use short direct question overlays
Expression anchor	surprised face on left panel	emotional entry point	Capture exaggerated but readable reaction

Use Cases & Transfers

Music discovery clips: ask audience to guess songs.
Trend response posts: reaction panel + source panel.
Collab cross-audience content: mix creator identity with street culture.

Transfer recipes

Recipe	Keep	Change	Slot template (EN)
Food review variant	reaction left + real clip right	swap interview with taste test	{reaction_panel} + {street_clip_panel} asking {prompt_text}
Fashion guess variant	split and text CTA	replace “listening” with outfit question	{avatar_reaction} with {public_style_clip} and {question_overlay}
Game challenge variant	meme pacing and side-by-side contrast	use gameplay on right panel	{left_reaction} vs {right_source_clip} in {vertical_split}

Aesthetic Read

What works here is pacing design, not beauty lighting. The frame is built for instant comprehension and rapid emotional response.

Observed cue	Why it matters	Recreate move
Hard panel split	clear narrative structure	keep strict visual division
Text-on-action panel	boosts engagement prompt	place one question overlay on source panel
Strong facial reaction	emotional hook	lock one readable expression on left

Prompt Technique Breakdown

Prompt chunk	What it controls	Swap ideas (EN, 2-3 options)
left reaction close-up	emotion anchor	surprised face \| laughing reaction \| shocked pause
right street interview clip	real-world context	park interview \| sidewalk challenge \| festival question clip
yellow prompt text	comment CTA	“WHAT SONG?” \| “GUESS THIS?” \| “YOUR PICK?”

Remix Steps

Lock split layout, reaction face, and one question overlay.
Change one variable per run: topic prompt, right-panel context, or reaction intensity.
Sequence: teaser split -> answer reveal -> audience reply compilation.