← Back to galleryIllustration
Naive oil pastel travel line sketch — After
ChatGPT
IllustrationChatGPT

Naive oil pastel travel line sketch

Creator @Hamburgerai

Prompt

Please convert each travel photo I upload into an independent 「NAIVE OIL-PASTEL TRAVEL LINE|稚拙油画棒旅行线描」.

Output each photo separately, without multi-image stitching. The overall layout uses a 3:4 vertical top-and-bottom comparison composition.

The upper part occupies about 48%–50% of the frame, retaining the clear original photo with substantial area; the lower part occupies about 50%–52%, refining the same scene into a minimalist, naive yet aesthetically controlled oil-pastel travel line drawing.

Use clear, straight horizontal boundaries between the upper and lower areas; do not use gradients, feathering, or mutual penetration.

The upper part must faithfully retain the original photo.

Accurately maintain the subject identity, number of subjects, orientation, posture, action, key architectural structures, vehicle proportions, boat direction, ridge morphology, spatial relationship between figures and environment, natural textures, real light and shadow, and original color atmosphere.

Apply only slight color grading for travel magazines, independent publications, or photography exhibitions, and add an extremely thin film grain.

Do not redraw the photo, do not stretch, mirror, move, replace, or add subjects; do not change vehicle direction, figure posture, building outlines, mountain peak morphology, or number of boats. The upper part must still clearly be a real photographic work.

The lower part should not redraw the complete composition of the photo.

First understand the photo, then select the most memorable subject from it, for example:

- A single boat
- A lighthouse or tower
- A pavilion, hut, or independent building
- A roadside vehicle
- An isolated mountain peak
- One to two figures
- An animal or travel object with a clear outline

Keep only the minimum visual facts that make the subject instantly recognizable:

- Number of subjects
- Overall outer silhouette
- Orientation or direction of travel
- Posture
- Most important proportional relationships
- One to three key structures
- One clue explaining the environment

For boats, keep the bow, stern, hull curve, cabin, or motor; for vehicles, keep the wheel distance, body direction, seats, and front structure; for buildings, keep the roofline, main entrance, and support relationships; for mountain peaks, keep the unique ridge and one necessary horizon line; for figures, keep the number, posture, orientation, hairstyle, or main clothing color, without depicting facial features, clothing folds, or fine human structures.

Delete dense vegetation, repeated window panes, tourist shadows, advertising signs, complex water ripples, cloud details, rock textures, and photographic-level light and shadow.

The background should retain at most one to three broken path lines, shorelines, water surface lines, horizons, ridges, or ground hints. Do not form a complete background, and do not shrink the entire original scene to fit it all onto the paper.

The lower part uses warm off-white or light ivory matte drawing paper; the paper surface has natural fibers, small particles, and obvious paper texture resistance.

The subject usually occupies only 22%–42% of the width and 20%–38% of the height of the conversion area, placed in the lower-middle, left-leaning, or right-leaning, following the direction of the subject in the original photo.

Retain at least 65%–80% of completely unpainted continuous paper white. The paper must become the largest color block in the lower half.

Use childlike, direct, but aesthetically maturely controlled oil pastels, thick crayons, or waxy chalk lines to redraw the subject.

Lines must show:

- Obvious thickness
- Uneven pressure
- Waxy particles
- Paper texture resistance
- Natural breaks
- Partial white gaps
- Occasional buildup
- Hesitation and back-folding
- Slight double lines
- Partial misalignment
- Unclosed outlines
- Active stopping at secondary structures

Do not use a uniform black outer outline.

Extract 3–6 colors from the original photo, using mutually separate colored lines to distinguish structures.

For example:

- The subject's signature color for the main outer outline
- Gray-blue or charcoal gray for tires, shadows, or ridges
- Ocher, sienna, or taupe for buildings and ground
- Gray-green or olive green for a small number of plant hints
- Light blue or gray-blue for one or two water surface or sky direction lines

Almost no full coloring.

Only add a small amount of loose, transparent short cross-hatching, thick scratches, or waxy color blocks at the most critical positions of the subject. Allow colors to be slightly darker where crayons overlap, but do not create realistic volume.

Do not draw complete shadows, do not use smooth gradients, and do not pursue correct and smooth perspective. It should be like an experienced illustrator deliberately retaining a somewhat clumsy hand-feel.

The lower half adds a small amount of small text by default.

Small text includes:

1. A small title of 1–3 English words;
2. A very short observation of 5–10 English words.

Before generating, please clearly replace:

TITLE: “[English Title]”
OBSERVATION: “[Short sentence based on the real content of the photo]”

Observation sentences should not introduce scenic knowledge or write vague travel quotes. It should describe actions, colors, positions, or feelings truly visible in the photo, such as:

- “yellow waits where the alley cools”
- “the boat rests below the pale cliff”
- “one figure stands beneath the peaks”
- “a small door beneath stacked stone”

Text should be placed in the continuous paper white away from the subject, and the total area must not exceed 3%–5% of the lower area.

Text should use:

- Slightly slanted blunt pencil handwriting
- Small oil pastel signatures
- Simple handwritten sans-serif with slight breaks
- Typefaces with paper teeth, white gaps, and waxy particles

Text color is extracted from the subject's outline color, gray-blue, taupe, or ocher. Do not use smooth, neat, floating digital fonts.

Do not add author names, accounts, signatures, brands, logos, QR codes, URLs, fictional locations, fictional dates, fake Chinese, or garbled text.

The overall style should be: simple, quiet, relaxed, slightly clumsy, like a small sketch casually drawn in a paper notebook during a trip. It can be naive but not childish; simple but not sloppy; inaccurate, but the subject's identity, direction, and posture must be accurate.

Avoid: children’s classroom drawing, kindergarten cartoon, cute sticker aesthetics, flower-vase template, colored-pencil detailed rendering, adult coloring book, uniform black outline, automatic edge tracing, clean vector icon, smooth pen line, full oil-pastel filling, heavy oil painting, realistic crayon portrait, complete background, individual leaves, repeated water waves, photographic lighting, gradient shadows, full-page illustration, decorative typography, fake text, author signature, logo, watermark, blurry split boundary.

The most important rule: First strip the photo down to just a few of the most recognizable lines of the subject, then let the crayon particles, breaks, hesitant outlines, and small text signatures work together with the large areas of untouched paper to complete the memory.

Copy the text, add your own details, and try it in your image model.