RTI Imaging – baby steps

First attempt using a RTI dome populated with 6 x 3 watt daylight LEDs controlled by a custom control box

without processing the image above is a typical record of the subject


My first attempt at RTI, that’s Reflectance Transformation Imaging to you and me 🙂 . I drifted into this type of photography after shooting graffiti at Gloucester cathedral and quickly realised that normal photographic techniques can struggle to bring out all the detail.

Using directional lighting placed at a low angle to throw a strong shadow into the dips and bumps of graffiti craved into stone gives some impression of what is there.

The RTI technique uses a single light source to take multiple photographs, typically 50 or more, with the light source being moved all around the subject during the capture session.

For my 1st attempt at RTI I used a carving in Welsh slate of a Draig, the Welsh Dragon. The sculpture was commissioned for my home several years ago but seemed ideal as a test subject.


1st Baby steps – for this test I was using:

  1. A Canon 5D4 and 24-70mm lens set at 47mm focal length with the target approximately 500mm from the camera and the diagonal coverage of the target was 400mm.
  2. A Sirui tripod with the camera suspended off an inverted centre column.
  3. For lighting I used a Godox AD200 flash head and a laser range finder mounted on a common bracket which I can mount onto a second tripod.
  4. A Phottix wireless camera trigger.
  5. The X-rite ColorChecker Passport Photo calibration card.
  6. A 10mm Silicon Nitride ball bearing and a 25mm Obsidian ball.

Some initial thoughts on the image capture process:

  1. Never underestimate the shadow being thrown by the light sphere when illuminated at 15 degrees. Check it with a test image to ensure the shadow doesn’t encroach on the subject. 
  2. Don’t frame the subject too tightly, leave room for the light sphere and check the shadow again. For info, the length of the shadow is 4 times the height of the light sphere at 15 degrees, it is long!
  3. The 10mm ball bearing only covered 160 pixels on the image, the recommended diameter should be around 250 pixels. The sequence worked but I suspect it was marginal. This is the blended light sphere image the RTIbuilder software produced and shows the highlights created by each flash position. It’s nowhere near perfect but adequate for this first test.
The blended Light sphere

4. I used the X-rite ColourChecker Passport Photo calibration chart to ensure my colour and exposure are as good as I can make them. The brightest image will always be when the flashgun is nearest to the surface normal so check the exposure at 65 degrees and avoid blowing out the highlights.

The off white square next to the pure white square should be used to set the white balance and to set the optimum exposure in Lightroom. The RGB value for this square should 200,200,200, in Lightroom this is approximately 90%, 90%, 90%

5. The black obsidian spheres are relatively cheap but the surface is fairly soft and easily marked, I would recommend buying the more expensive Silicon Nitride ball bearings which are very hard and the surface finish is excellent.

6. The distance from the flashgun to the subject should be 3-4 times the diagonal of the image. I used a laser rangefinder mounted to the side of the flashgun to set the distance and the angle of the flashgun.

7. It’s important to ensure that the camera does not record any of the subject due to the ambient light, so use a low iso (100) and small aperture (f11) and set the shutter speed to your highest sync speed, I used 1/160 second. It is not recommended  to use a smaller aperture because of diffraction, this optical effect causes the image to be softer even though there is an increase in depth of field.


2nd Baby steps – for this test I was using:

a dull 2 pence coin captured with a Canon 5D4 and Canon 100mm macro lens

 

For further information, software and guidance on the capture and image processing the Cultural Heritage Imaging website is well worth a visit.