Paper Spirit Zine

Paper Spirit Zine

Create a vertical editorial poster from a supplied photo. Preserve the factual photograph and its atmosphere first. Let the lower paper field become a scene afterimage: sometimes an animal-like vessel, sometimes only a ridge, shadow, waterline, branch, or breath mark that has a living contour.

Do not start from the animal. Start from what makes the original photo better than any transformed version: light, air, distance, weather, scale, emptiness, color temperature, or the small human cue. The animal must appear only if the source geometry already suggests it. It should feel discovered after looking, not designed first.

Return the generated image plus one brief Chinese rationale by default. Include the full prompt only when the user asks.

Core Structure

Use a two-world poster by default:

• top field: truthful cropped photograph, usually 35-48% of canvas height;
• bottom field: warm paper or mineral off-white field, usually 52-65%;
• transition: a visible but quiet photo-to-paper handoff, torn edge, soft paper abrasion, or faint photographic residue;
• main lower subject: one source-matched afterimage, which may be an animal-like partial vessel, a sleeping ridge, a shoreline breath, a shadow body, a cloud wing, or a tree/road trace;
• text: one quiet micro-text line, Chinese 8 Han characters or fewer or English 5 words or fewer.

The lower field must improve or deepen the original photo. If the transformation is less beautiful than the source, reduce the lower form, restore source atmosphere, and let the shape become more ambiguous.

Do not simply paste the original photo into a complete animal outline. Redesign the source so the lower form and scene share a visual logic. Prefer partial traces, negative space, broken contours, and open bodies over clean silhouettes.

Photo-First Hierarchy

Protect the source photograph's strongest feeling before choosing any animal:

• keep the original mood readable at thumbnail size;
• preserve the best light and air instead of replacing them with a concept;
• preserve scale relationships, such as tiny people under mountains, a small house inside forest, a road through city light, or one tree in empty snow;
• use the lower paper field to echo the source's native gesture, not to explain it;
• avoid any lower animal that makes the original feel smaller, flatter, more decorative, or more generic.

Use a quick test before generating: if the prompt can be summarized as "put the photo inside an animal," it is wrong. If it can be summarized as "let the photo's own gesture leave a living trace on paper," it is closer.

Authorial Proposition

Before drawing the lower field, decide what the photograph keeps as an idea:

• a mountain keeps its breath;
• a lake keeps its silence;
• a forest keeps its density;
• a road keeps its pull;
• a city keeps its noise or glow;
• a portrait keeps its guarded posture;
• a house keeps its warmth inside a larger landscape.

The lower afterimage must answer that proposition. If the animal only names a category, such as "forest = deer" or "city = moth", the result is too literal. Add one transformation rule: make the ridge become a back, the waterline become a belly, the road become a spine, the cloud become breath, or the tree line become fur. If no animal is needed, use a non-animal living trace.

Scene Reading

Before generating, build an internal scene card:

• core subject and place identity;
• whether the scene is people-led: visible face, gaze, gesture, body posture, or human relationship;
• dominant gesture: horizon, path, ridge, gaze, flight, road, shoreline, architecture, or light beam;
• native mood and color temperature;
• most important source texture: water, cloud, forest, wall, road, building, grass, snow, light, shadow, or crowd;
• quiet field available for paper;
• living-form candidates that naturally match the scene's subject, movement, or emotion.

Choose the lower form from the photograph rather than from personal preference. It may be an animal, but it may also remain an almost-animal landscape trace.

People-Led Photo Priority

When the supplied photo contains a clear person, face, gaze, pose, or human relationship, the person is the primary subject. The animal may not replace, outrank, or visually compete with the person, but the lower animal vessel must still contain the person.

For people-led photos:

• keep the person large, complete, recognizable, and emotionally central in the photographic section;
• preserve the face, gaze direction, body posture, clothing silhouette, and the key human-environment relationship;
• allocate more canvas authority to the photograph, usually 48-60% of the poster height;
• make the lower animal an authored container for the person, not an empty symbolic animal;
• place a recognizable source-derived human cue inside the animal vessel: face angle, gaze, seated or walking pose, clothing mass, hand gesture, or body silhouette;
• derive the animal from the person's gesture, clothing mass, shadow, hair, gaze, or surrounding light rather than drawing a standalone animal portrait;
• let the animal share one contour, axis, or rhythm with the person, such as a curled seated posture, a turned head, a walking direction, a hand gesture, or a cast shadow;
• preserve key surrounding fragments inside the same vessel only after the human cue is readable, such as road lights, window glow, crosswalk stripes, water, trees, or wall texture;
• avoid detailed animal faces, realistic animal eyes, cute animal behavior, and any animal rendering that pulls attention away from the human cue inside the vessel.

If the lower animal reads as a standalone animal portrait without the person inside it, the composition has failed. Rebuild the animal as a vessel around the human cue and source scene.

Form Selection

Use these defaults only when the source geometry supports them and the animal improves the photo:

• water, lake, sea, island, shoreline, mist over water: whale, fish, ray, heron;
• sky, clouds, mountain air, migration, open distance: bird, crane, eagle, swallow;
• forest, leaves, park, autumn, stillness: deer, fox, hare;
• road, grassland, journey, sunset movement: horse, dog, bird;
• interior, cafe, bedroom, self-portrait, quiet daily scene: cat, dog, moth, small bird;
• dense city, neon, night, glass, crowds: crow, cat, koi, moth;
• ancient architecture, temples, walls: crane, horse, swallow, lion-like guardian only if it remains restrained.

If the animal would feel forced, choose a non-animal form: sleeping ridge, torn shoreline, wind line, tree-root body, cloud-breath arc, road-spine, shadow field, or architectural echo. If two animals fit, choose the one whose silhouette best supports the source geometry. Prefer one readable but quiet form over multiple decorative animals.

Composition Rules

Make the lower trace large enough to be the primary lower-field form, usually 55-80% of poster width. Let it breathe; do not fill the paper field edge to edge.

Prefer these forms over a clean full-body silhouette:

• partial body: only back, head, wing, antlers, tail, flank, or jaw when these are truly needed;
• negative animal: the animal is readable through missing paper and surrounding scene fragments, but remains secondary to the source mood;
• landscape-animal hybrid: hills, roads, clouds, buildings, or shadows become anatomical lines;
• pure landscape afterimage: no animal face, only a living curve made from ridge, water, wind, or shadow;
• broken vessel: one side is photographic, the other side dissolves into dry ink, paper fibers, or blank field;
• off-center trace: the animal does not sit like a centered emblem.

Inside the animal:

• preserve one source-specific spatial relationship, such as shoreline over water, ridge over valley, building over plaza, road through city, or figure against light;
• preserve at least one source gesture outside literal texture, such as the slope becoming a back, cloud becoming breath, road becoming spine, window grid becoming ribs, or shadow becoming a hidden limb;
• let source texture follow the animal's anatomy where it feels natural, such as water bands following a whale belly, cloud drift following wings, tree lines following antlers, or road lights following a spine;
• use only a few photographic fragments, ink planes, torn-paper blocks, and line accents;
• keep at least 35-60% of the lower form quiet, pale, or incomplete;
• allow the contour to break, fade, or open in multiple places;
• remove facial details unless a face is essential to the source emotion.

The lower trace can be filled with photographic material, printed ink, translucent cut-paper, or a mix of these. It must remain readable at thumbnail size as a source-derived memory, not as a centered logo or specimen.

Photo Handling

Keep the top photograph truthful:

• preserve natural color, perspective, main subject identity, and recognizable place cues;
• crop deliberately, not randomly;
• do not over-smooth faces, architecture, animal bodies, or small objects in the photo;
• do not replace the photograph with an illustration.

The lower afterimage may reinterpret the photo, but it should still carry the same scene, not a generic landscape. When in doubt, make the source photograph more important and the lower shape quieter.

For people-led photos, the lower animal must carry the person inside it as the emotional anchor. Do not crop the person into a tiny decorative strip, erase them into texture, or use the lower animal as a standalone animal portrait.

Visual Language

Use a refined art-book zine aesthetic:

• flat scanned paper, off-white or mineral warm paper;
• torn fiber, dry ink, xerox, riso, letterpress, halftone, or cut-paper texture;
• one optional high-chroma source-derived accent, not a decorative splash;
• restrained linework that can describe animal anatomy only where useful;
• wide negative space and asymmetry;
• quiet typewriter, pencil, dry Song-style Chinese, or small letterpress text.

The result should feel like a memory specimen, field note, or paper afterimage, not a logo, tattoo design, poster template, children illustration, wildlife infographic, or animal-shaped demo.

Prompt Compiler

Write the image-generation prompt in four compact paragraphs:

1. Canvas, ratio, top photo placement, lower paper field, photo-to-paper transition, and overall hierarchy.
2. Factual scene details that must remain recognizable in the top photo.
3. Chosen afterimage form, the artistic proposition, why it matches the source, whether it is animal-like or non-animal, what parts are missing or broken, source fragments inside it, texture behavior, linework, color accent, and micro-text.
4. Material finish, mood, and hard avoids.

Use specific visible instructions. Do not include analysis labels, file paths, or metadata in the prompt.

Quality Gate

Before returning, check:

• Does the top photo still read as the supplied photo?
• Is the original photo's light, air, distance, and best feeling still stronger than the concept?
• If a person is present, does the person remain the first-read subject?
• Are the person's face, gaze, pose, and emotional role preserved?
• Does the lower form feel source-matched and proposition-driven instead of arbitrary?
• If animal-like, is it readable without becoming a sticker, logo, clean specimen, or wildlife illustration?
• In people-led scenes, does the lower animal vessel visibly contain the person or a clear human cue from the source?
• Does the scene inside the lower form preserve one real spatial relationship?
• Is the lower field spacious, tactile, and editorial?
• Is there a real material handoff between photo and paper?
• Is the micro-text quiet and short?
• Does the result avoid generic double exposure, fantasy animal posters, cute animal graphics, and logo-like silhouettes?

If the animal is too decorative, too complete, or makes the source photo less beautiful, regenerate with a quieter non-animal or almost-animal trace, fewer marks, more source atmosphere, a broken contour, and a more source-derived interior. If the scene disappears, restore the strongest source relationship inside the lower form.

If the human subject is weakened, regenerate with the person larger and more central in the top photo and with a recognizable human cue placed inside the lower animal vessel. Reduce standalone animal facial features and rebuild the animal around the person's posture, clothing mass, gaze, or shadow.

Hard Avoids

Avoid forcing an animal onto every photo, animal-shaped demos, cute cartoon animals, mascot energy, children's book style, fantasy spirit animals, glowing neon outlines, tattoo logos, sports logos, sticker borders, crisp digital masks, generic double exposure, full-scene photo pasted unedited into an animal, animal portraits that steal attention from people, people-led animal vessels with no person inside, detailed animal eyes in people-led scenes, shrinking a human subject into a decorative photo strip, busy collage, many animals, decorative splashes, heavy vintage stains, coffee paper, thick drop shadows, glossy mockups, ad layout, CTA text, large titles, watermarks, fake signatures, and unrelated symbols.