The "photo vs render" decision is product-specific
For a long time, product photography was the only viable option. 3D rendering, once specialist and expensive, has commoditised. Both have their place.Where photography wins
- Materials with subtle, hard-to-simulate properties — natural fabrics, leather, wood grain
- Lifestyle context — product in human use, real environment, real light
- Authenticity emphasis — handmade, artisan, "real" as a brand value
- Time-sensitive content — fast turnaround on real samples
- Fashion / textile — drape, flow, texture interaction with light
Where 3D rendering wins
- Configurable products — many SKU variants
- Pre-launch visuals — product doesn't exist yet
- Industrial / mechanical products — exploded views, cutaways
- Perfect lighting / control — absolute consistency across the catalogue
- Animation / interactive — 360° viewers, AR, configurators
- Hero / aspirational shots — impossible-to-photograph compositions
The hybrid — increasingly common
- Photography for hero / lifestyle
- 3D for catalogue variants and configurators
- 3D for AR experiences
- Photo backplates with 3D product comp'd in
The 3D pipeline
- CAD source — for industrial / engineered products
- Modelled-from-scratch — for consumer / non-CAD products
- Materials / shading — PBR workflow. Albedo, roughness, metalness, normal, displacement maps
- Lighting — HDRI environment + key fills
- Rendering — Cycles (Blender), V-Ray, Corona, Octane, Arnold, Redshift. GPU-accelerated dominant in 2026
- Post-processing — composite, colour grade, retouch
Cost comparison (rough, 2026)
- Studio photography — TL 2 000-10 000 per product
- 3D rendering — TL 1 500-8 000 per render initially; subsequent variants near-zero
- AR / interactive 3D — TL 5 000-20 000+ per asset
For products with many variants, 3D wins economically by the second or third SKU.
The "AI photography" question
Generative image tools have crossed the threshold for some product imagery in 2026:- Background generation for catalogue
- Lifestyle context (placing the real product in generated scenes)
- Mood / aesthetic generation for marketing
Where it's still rough:
- Exact product accuracy
- Brand consistency across many images
- Legal clarity around training data and trademark
The hybrid we see working: real product photo + AI-generated background, manually composited.
File formats and delivery
- Master files at high resolution
- Web-ready exports (AVIF, WebP, JPEG fallback)
- Multiple sizes / aspect ratios
- Metadata (alt text, captions)
- Asset library structure that supports findability
Brand consistency
- Lighting language — defined and documented
- Colour cast / treatment — calibrated, consistent
- Composition rules — common framing
- Background — solid, gradient, lifestyle — pick a system
One pattern we'd warn about
Letting "we have CAD, just render it" produce sterile catalogue imagery. Engineering CAD doesn't include the materials, lighting, or composition that make a render look right.One pattern that always pays off
A reference style guide for product imagery — three or four reference shots that define the language.What's your imagery split?
