Courses / AI Images & Video / Lesson 3 of 7
Style, Light, and Composition
Learn the vocabulary photographers, illustrators, and art directors use, so you can control how an AI image looks and feels instead of hoping for the best.
Start with an example
Two marketing students are asked to make an image of the same thing: a cup of coffee on a café table.
Student A writes: “A cup of coffee on a table, nice photo.” The result is fine but forgettable, like a thousand other stock photos.
Student B writes: “Editorial photograph of a ceramic cup of black coffee on a marble café table, early morning, soft window light from the left casting a long gentle shadow, steam rising, shallow depth of field with a blurred street outside the window, cup placed on the right third of the frame, muted palette of cream, gray, and warm brown, 4:5 vertical.”
Student B’s image looks like it belongs in a magazine. Same subject. Same tool. The difference is vocabulary: Student B knows the words that photographers and art directors use, and the model learned those words from millions of captioned professional images.
This lesson gives you that vocabulary.
1. Style and medium: the overall look
Style words tell the model what kind of image to make. They have a bigger effect than any other single word.
| Style word | What you get | Good for |
|---|---|---|
| Photograph / editorial photo | Realistic, like a magazine | Products, food, lifestyle |
| Documentary photo | Natural, candid, unposed | Community and nonprofit stories |
| Flat vector illustration | Clean shapes, solid colors | Infographics, icons, websites |
| Watercolor | Soft edges, paint texture | Invitations, gentle or artistic brands |
| 3D render | Smooth, dimensional, toy-like | Tech products, characters, mascots |
| Ink sketch / line drawing | Hand-drawn, minimal | Diagrams, menus, journals |
| Collage / paper cut-out | Layered, textured, playful | Posters, social graphics |
You can sharpen a style with descriptive modifiers: “minimalist flat vector,” “vintage 1970s film photograph,” “bold, high-contrast poster illustration.”
A note on artists’ names: Some people prompt “in the style of [famous living artist].” Avoid this. Describe the qualities you like instead, such as bold outlines, muted earth tones, or exaggerated proportions. It respects the artist’s work, avoids legal gray areas, and builds your own skill. Lesson 6 covers this in depth.
2. Light: the emotion of the image
Light shapes mood more than almost anything else. You can control three things:
Direction: where light comes from
| Direction | Effect |
|---|---|
| Front light | Even, clear, flat. Good for product clarity |
| Side light | Shape and texture; dramatic and dimensional |
| Backlight | Glowing edges and silhouettes; dreamy or heroic |
| Top light (noon sun) | Harsh shadows under eyes and objects; usually unflattering |
Quality: hard or soft
- Hard light (direct sun, a bare bulb): sharp shadows, high contrast, intense.
- Soft light (cloudy day, window light, a large lamp): gentle shadows, flattering, calm.
Color: warm or cool
- Warm (golden hour, candlelight, tungsten bulbs): cozy, inviting, nostalgic.
- Cool (blue hour, overcast, fluorescent): calm, clinical, modern, or lonely.
- Colored light (neon, stage lights): energetic, nightlife, futuristic.
Put them together: “soft, warm window light from the left” names the quality, color, and direction in six words.
3. Camera and framing: how close and from where
Image models learned from captions that often included camera details, so photography terms work remarkably well.
Shot size
| Term | Shows |
|---|---|
| Extreme close-up | A detail: an eye, a stitch, a drop of sauce |
| Close-up | A face or a single product |
| Medium shot | A person from the waist up |
| Full shot | A whole person, head to toe |
| Wide shot | The person in their environment |
Camera angle
- Eye level: natural and relatable.
- Low angle (looking up): powerful, impressive, heroic.
- High angle (looking down): smaller, vulnerable, or overview.
- Top-down / flat lay: directly overhead; popular for food and products.
Lens and depth of field
- Wide-angle lens (24mm): shows more space, can exaggerate size; good for interiors.
- Standard lens (35–50mm): natural, like human vision.
- Telephoto lens (85mm+): compresses the background; flattering for portraits.
- Shallow depth of field: sharp subject, blurry background. Focuses attention.
- Deep depth of field: everything sharp. Good for landscapes and busy scenes.
4. Composition: where things go
Composition is how elements are arranged inside the frame. It guides where the viewer looks.
- Rule of thirds: Imagine the frame divided into a 3×3 grid. Placing the subject on a line or intersection, rather than dead center, usually feels more balanced and dynamic.
- Negative space: Empty areas around the subject. Essential when you’ll add a headline.
- Leading lines: Roads, railings, or shelves that draw the eye toward the subject.
- Symmetry: Centered and balanced. Feels calm, formal, or grand.
- Foreground framing: Something soft in the foreground, like leaves or a doorway, adds depth.
Prompt example: “Subject on the right third, clean negative space on the left, a leading line from the sidewalk drawing the eye to the storefront.”
5. Color palette
Name colors directly: “palette of navy, cream, and red.” Or describe a mood: “muted earth tones,” “bright primary colors,” “pastel,” “high-contrast black and white.”
For brand work, use your actual brand colors. Many tools respond to color names better than hex codes, so write “deep navy blue” rather than only “#001F4A.”
Putting it together: the professional prompt
“Documentary photograph of a young woman teaching a free coding class to teenagers in a community center, medium shot at eye level, warm late-afternoon light from tall windows on the right, shallow depth of field, teacher on the left third with students softly blurred behind her, palette of warm wood, navy, and cream, 16:9, no text or logos.”
Read it again and find all the parts: style (documentary photograph), subject, setting, shot size and angle, light (warm, late afternoon, from the right), depth of field, composition (left third), palette, format, and avoid list.
A worked example: the style study
Imani designs graphics for her family’s Caribbean restaurant. She wants a signature look for the menu board. She runs a style study, changing one variable at a time:
| Version | Changed | Result |
|---|---|---|
| A | Baseline: “Editorial photo of jerk chicken with rice and peas on a wooden table” | Good, but generic |
| B | Style → “bold, colorful flat vector illustration” | Fun and graphic, but loses appetite appeal |
| C | Light → “warm golden-hour side light, deep shadows” (back to photo) | Rich, warm, mouth-watering |
| D | Palette → “vibrant greens, golds, and reds inspired by tropical fruit” | Bright and festive; matches the restaurant’s walls |
She combines C and D: warm side light with a vibrant tropical palette. Because she changed one thing at a time, she knows exactly which words created the look, and she saves them in her visual prompt kit (Lesson 5) for every future menu item.
Think deeper: why camera words work with no camera
There’s no actual lens inside an AI. So why does typing “85mm lens” or “shallow depth of field” change the picture?
Because of who wrote the captions the model learned from. Photographers, stock photo sites, and art communities describe images with precise technical words. Over millions of examples, the model learned that “85mm” goes with flattering portraits and blurry backgrounds, and “wide angle” goes with big, open scenes.
That leads to a useful rule: the best prompt words are the ones professionals use. Words from photography, film, illustration, and design work well because they were used consistently. Words like “beautiful,” “amazing,” or “high quality” barely help, because people attached them to every kind of image imaginable.
Learning design basics also helps you judge your results, not just request them:
- Contrast pulls the eye to the brightest or sharpest spot.
- Hierarchy decides what people notice first, second, and third.
- Warm colors feel inviting; cool colors feel calm or distant.
- Symmetry feels stable and formal; asymmetry feels lively.
These ideas are centuries older than AI, and they’ll outlast every tool.
Questions to sit with:
- If the AI learned “professional quality” mostly from expensive photography, what kinds of people, places, and styles might it treat as less “professional”?
- Is knowing the right words to ask for a creative skill? Why or why not?
Project
Create your own style study with the template below. Generate the same subject four times, changing only style, then light, then camera angle, then palette. Label each version with the exact words you changed, pick your favorite, and write two sentences explaining why it works for your brand.
Prompt template
SUBJECT: [who/what, doing what] SETTING: [where, what's around] STYLE / MEDIUM: [photograph, editorial illustration, watercolor, 3D render, flat vector, ink sketch] LIGHT: [direction + quality + color, e.g. soft window light from the left, warm] CAMERA / FRAMING: [shot size, angle, lens, e.g. medium shot, low angle, 35mm, shallow depth of field] COMPOSITION: [subject placement, negative space, leading lines] COLOR PALETTE: [3 colors or a mood, e.g. navy, cream, and red] FORMAT: [1:1 / 4:5 / 9:16 / 16:9] AVOID: [text, logos, extra people, clutter]
Try it
Pick one simple subject, like a cup of coffee on a table. Generate it four times, changing only the lighting: soft window light, harsh noon sun, neon light at night, and candlelight. Write one word describing the mood of each. You've just learned that light controls emotion.
Your project
Create a 'style study': generate the same subject in four versions, changing only one variable each time (style, light, camera angle, then color palette). Label each image with the exact words you changed. Choose the version that best fits your brand and write two sentences explaining why it works.
Key takeaways
- Style and medium words set the overall look: photograph, watercolor, 3D render, flat vector, and more.
- Light has three parts you can control: direction, quality (hard or soft), and color (warm or cool).
- Camera words control framing: shot size, angle, lens, and depth of field.
- Composition rules like the rule of thirds, negative space, and leading lines guide the viewer's eye.
- Change one variable at a time so you learn exactly what each word does.
Finish your project and pass the knowledge check to complete this lesson.
