Sourcing and Adapting Web Content
The web is an enormous library of reference material for rendering. Using it well is a skill — evaluating quality, understanding use rights, and transforming raw sources into compositing-ready content.
Fair use in design education
In the United States, the fair use doctrine allows limited use of copyrighted material without permission for purposes including education, commentary, and transformative creative work. Design renderings created for educational purposes — where sourced images are transformed through cropping, color adjustment, distortion, and compositing into new visual works — generally fall within fair use boundaries.
The key qualifier is transformative use: the sourced material must be significantly altered or recontextualized in the new work, not reproduced as-is. A rendering where a sourced texture is distorted, color-graded, and composited with other elements to produce a new image is transformative. A rendering where a sourced photograph is placed unchanged in a layout is not. The practical standard: if a viewer looking at your final work could identify the specific source photograph you used, it is not sufficiently transformed.
For professional work beyond this course, the standards change significantly. Know the license of every image you use in a paid deliverable. When in doubt, source from explicitly licensed or public domain repositories.
What makes a good source image
| Quality criterion | How to evaluate it | Why it matters |
|---|---|---|
| Resolution | Is the image large enough to place at the required size at 150 PPI without upsampling? At 150 PPI, a 6" × 6" area in the composition needs at least 900 × 900 pixels. | Upsampled imagery shows pixelation and loses material detail. Low-resolution sources produce renders that look degraded at output size. |
| Lighting angle | Does the lighting direction in the source match or approximate your scene's sun angle? | A texture photographed with light coming from the left placed in a scene with light from the right creates an immediate visual inconsistency. Color correction can shift color temperature; it cannot reverse shadow direction. |
| Extraction feasibility | Is the subject clearly separated from its background by color or value contrast? Or are the edges complex and low-contrast? | Difficult extractions take disproportionate time relative to their visual contribution. Evaluate extraction difficulty before committing to a source. |
| Perspective angle | For plan textures, is the source photographed from roughly overhead? For elevation textures, from perpendicular? Or is significant distortion needed? | The more perspective correction required, the more distortion quality degradation. A source photographed at 45 degrees can be distorted to overhead, but detail will compress at one end. |
| Color temperature match | Is the source's overall warm/cool balance compatible with your scene, or is heavy correction required? | Heavy color correction (large shifts in white balance) degrades color accuracy and can introduce unwanted color casts in specific tonal ranges. |
Free and openly licensed sources
| Source | Type | Best for |
|---|---|---|
| Unsplash (unsplash.com) | Free, no attribution required for most uses | People photography, sky images, atmospheric landscape photos |
| Pexels (pexels.com) | Free, Pexels license — no attribution required | People, architecture, textures, natural environments |
| Pixabay (pixabay.com) | Free, Pixabay license | General photography, some textures and vegetation |
| Poly Haven (polyhaven.com) | CC0 — fully public domain, no restrictions ever | High-quality physically based material textures: concrete, gravel, grass, bark, stone, fabric — all photographed overhead at consistent scale |
| ambientCG (ambientcg.com) | CC0 — fully public domain | PBR material textures — similar to Poly Haven, extensive library |
| Skalgubbar (skalgubbar.se) | Free for non-commercial and educational use | Extracted people at various scales and activities — specifically designed for architectural rendering |
| Dezgo, Adobe Stock Free Tier | Mixed licensing — check each image | People, landscapes, architectural contexts |
The chopping methodology
Rarely does a single source image provide exactly what a rendering needs. Professional compositing work draws on parts of multiple source images — a canopy from one tree photograph, a trunk from another, a shadow texture from a third. The art is in combining components from different sources so convincingly that the result reads as a unified element.
The process:
- Identify what you need specifically — not "a tree" but "a tree canopy that reads as deciduous, seen from slightly below, with light from the upper right, against a sky background"
- Source multiple candidates — collect 4–6 images that might provide what you need, even if none is perfect
- Extract the useful portion from each — not the whole image, but the part that contributes what you need (Card 09–10 for extraction technique)
- Transform to fit — scale, distort, and warp each component into position (Card 23)
- Unify color and lighting — adjust each component's color temperature and tonal range to match the scene (Card 16)
- Blend seams — use Wacom soft brushes on mask edges (Card 14) to eliminate visible hard boundaries between composited components
Avoiding stock photography clichés
Stock photography has visual signatures that trained eyes immediately recognize — the perfectly lit business handshake, the family running in a park, the lone architect staring at a blueprint. These images are so frequently used in presentations that their presence signals generic effort rather than specific intent. Avoid them.
For rendering work, the most useful entourage images are anonymous: candid people in motion, figures seen from behind or at middle distance where their faces aren't identifiable, people doing mundane activities (walking, sitting, talking to someone) rather than performing for the camera. Images where the figure is naturally integrated into a spatial context are more convincing than studio photographs placed into a composited scene.
Try this
Source a paving texture and three human figures for a plan view rendering using only Poly Haven (for the texture) and Skalgubbar (for the figures). Before placing anything, evaluate each source against the five quality criteria in the table above. Note which criteria each source passes and which it doesn't. Then adapt accordingly: distort the texture to overhead view (Card 23), extract and scale the figures (Cards 09–10), and color-match all elements (Card 16). The constraints of quality sourcing will shape your choices more than any stylistic preference.
What breaks
Using a low-resolution image because it has exactly the right content — a blurry, pixelated person placed in a rendering degrades the entire image's quality. A higher-resolution person with a slightly different pose that requires Puppet Warp to adjust is always preferable. Resolution problems are permanent; pose problems are fixable.
Sourcing everything from the same aesthetic category — if all your figure photographs come from the same stock library, they carry the same lighting, the same color temperature, and often the same visual style, making the compositing uniformity feel artificial. Source from multiple different contexts and adjust each to match the scene.