Appearance
Use Reflectance Transformation Imaging (RTI) when the information you care about lives in a surface's shallow texture and how it catches light — incised inscriptions, tool marks, seal impressions, the ghost of erased writing — on objects that are essentially flat. RTI captures one viewpoint under dozens of lighting directions and merges them into a file you can relight interactively. Choose photogrammetry instead when you need true measurable 3D form, and reach for RTI when raking light reveals what flat-on photography hides.
When does RTI beat photogrammetry?
The decision is about where the meaning sits. Photogrammetry reconstructs geometry — the actual three-dimensional shape you can rotate and measure. RTI captures reflectance and micro-relief from a fixed viewpoint. So:
- A weathered slate gravestone where the inscription is almost flush with the surface: RTI. The letters are millimetres deep; relighting from a low angle makes them legible.
- A carved corbel you need as a rotatable, scaled model: photogrammetry.
- A cuneiform tablet where wedges are shallow impressions read by shadow: RTI.
- A whole monument facade for measurement: photogrammetry, or laser scanning.
If the question is "what does the surface say when raked by light?", RTI is usually the answer.
What kinds of objects suit RTI?
RTI shines on roughly planar, opaque, matte-to-semi-matte surfaces:
| Object | Why RTI fits |
|---|---|
| Worn coins and medals | Shallow relief read by directional light |
| Stone and slate inscriptions | Eroded letters become legible when relit |
| Clay tablets, seals, bullae | Impressed marks defined by shadow |
| Manuscripts with relief (drypoint, erased ink) | Surface disturbance, not colour, carries the trace |
| Paintings (raking detail) | Brushwork, craquelure, canvas weave |
It fits poorly where the signal is colour rather than texture, or where the object is deeply three-dimensional.
What does an RTI capture actually involve?
The accessible highlight-based method needs only a camera, a flash, and reflective spheres. The workflow:
text
1. Fix the camera on a tripod, object flat and parallel to the sensor.
2. Place two black reflective spheres in frame, beside the object.
3. Shoot 30-60 frames, moving the flash to a different position each
time (vary the angle across a virtual dome, ~15 deg increments).
4. Keep distance roughly constant; spheres record each light position.
5. Process in RTIBuilder; it reads the highlight on each sphere to
solve the light direction per frame, then fits a polynomial (PTM)
or hemispherical-harmonic (HSH) model per pixel.
6. View and relight in RTIViewer; export snapshots or the .rti file.The Cultural Heritage Imaging (CHI) tools — RTIBuilder and RTIViewer — are free, which keeps the entry cost to a tripod, a flash, and two black snooker balls.
How is RTI different from raking light photography?
A single raking-light photograph captures one lighting direction; you have to guess the best angle, and you are stuck with it. RTI is raking light from every direction at once, merged so a viewer can sweep the light interactively afterwards and apply enhancement modes (specular, normals visualisation) that make shallow features pop. The payoff is that a researcher who was not present at capture can relight the surface to test their own reading — a reproducibility win a single photo cannot offer.
What are the real trade-offs and costs?
- 2.5D, not 3D. RTI records one viewpoint's surface response. It is not a rotatable model and cannot measure depth like photogrammetry.
- Geometry limits. Deep undercuts, glossy, reflective, or transparent surfaces, and very large objects all defeat it.
- Time. Dozens of carefully aligned exposures per object; the camera must not move at all.
- Discipline. Consistent flash distance, in-frame spheres, and stable exposure are essential, or the light-direction solve degrades.
The cost is patience and technique, not money. For shallow surface detail it remains one of the highest-value-per-pound techniques in heritage imaging.
Should I combine RTI with other methods?
Often, yes. A common pairing is photogrammetry for the object's overall scaled geometry plus RTI for the legibility of a specific inscribed panel. Multispectral imaging complements RTI when the trace is partly chemical (faded ink) rather than purely topographic. Decide per question: geometry, surface relief, or spectral signature — each method answers a different one.
Key Takeaways
- RTI captures surface reflectance and shallow relief from one viewpoint, relightable afterwards.
- Choose RTI for flat-ish objects where light angle reveals the detail; choose photogrammetry for true 3D form.
- Ideal subjects: coins, stone inscriptions, tablets, seals, drypoint and erased-ink traces.
- The highlight method needs only a camera, flash, and two reflective spheres; CHI tools are free.
- RTI is 2.5D and fails on deep undercuts, glossy, or transparent surfaces.
- It is interactive, reproducible raking light — far more useful than a single fixed-angle photo.
- Combine with photogrammetry or multispectral imaging when questions span geometry and chemistry.
Frequently Asked Questions
What is Reflectance Transformation Imaging used for?
RTI captures how a surface reflects light from many directions, letting you relight it interactively afterwards. It excels at revealing shallow surface detail — incised inscriptions, tool marks, impressed seals, faded ink relief — on roughly flat objects.
When should I use RTI instead of photogrammetry?
Use RTI when the information is in surface texture and reflectance rather than 3D shape: a worn coin, a slate epitaph, a clay tablet. Use photogrammetry when you need true geometry, measurable form, or a full 3D model you can rotate.
What equipment do I need for RTI?
A fixed camera, a controllable light source, and reflective spheres for the highlight-based method. You can start with a tripod, a handheld flash, and two black snooker balls; a dome rig automates the light positions but costs more.
Is RTI better than raking light photography?
RTI is raking light from dozens of directions, combined into one interactive file. It removes the need to guess the single best light angle and lets a researcher relight the surface freely later, which a single raking-light photo cannot.
What are the main limitations of RTI?
RTI is essentially 2.5D: it captures a single viewpoint's surface response, not a full rotatable 3D model, and struggles with deep undercuts, glossy or transparent surfaces, and very large objects. It also needs a stable camera and dozens of exposures.
Can I do RTI for free?
Yes. The Cultural Heritage Imaging RTIBuilder and RTIViewer tools are free, and the highlight method needs only a camera, a flash, and reflective spheres. The cost is mainly time and careful capture technique.