Student Services Centre

Eye-kon produced three verified views at AVR Level 3, the level at which a representation confirms the use and appearance of materials as well as position and scale. All three formed part of the planning submission and supported the accompanying heritage and townscape work.

Three positions, chosen to ask three different questions. Peas Hill at street level, ten metres above datum, tests how the building meets the historic street. The Grand Arcade roof at 33 metres tests it against the rooftops. Great St Mary's Tower at 38 metres tests it in the skyline of the historic core, from the one place in Cambridge where that skyline can be read whole.

Production Process

1 | Reconnaissance, photography and survey

The camera positions were established on site with the University's Planning Officer, walked using a lens equivalent to the one the photographer would later use, with a lighting study run in 3ds Max beforehand to establish the best time of day for each position.

Photography was taken on a Phase One P45+ digital back mounted on an Arca Swiss RL3D with a 25 mm lens, giving a 67.9 degree horizontal and 53.5 degree vertical field of view. The two ground level positions were shot from a levelled tripod at 1.6 metres above the surveyed ground point. Great St Mary's Tower has no floor to stand a tripod on, so that camera sat 1.38 metres above the top of the stone bench. Location, height, angle, time, direction, vertical shift, aperture and shutter speed were recorded for every frame.

The survey worked to a briefing document with the required points marked up on photographs and described in writing. Forty one numbered control points were surveyed across the three views, chosen near and far, high and low, so that parallax has enough range to fix a virtual camera in only one possible position. The survey returned a 3D AutoCAD file with a locus at every point, which the camera match points snap to directly.

2 | 3D modelling and camera matching

The building was modelled in 3ds Max from Bennetts Associates' AutoCAD drawings and set into a scene built on the survey plan and 3D survey data, so it holds its real position relative to every surveyed point.

Virtual cameras were then derived using the camera match routine, which takes the surveyed points as they appear in the photograph and deduces mathematically where a camera must be to see all of them in those positions. This is the verifiable part of the process: if widely divergent points all register, no other camera position is possible, because parallax would place them differently relative to one another.

3 | Material and lighting calibration

Lighting was reconstructed in V-Ray with a sun model set to the site's latitude and longitude and to the time of day of each photograph. Virtual camera exposure, including shutter speed and aperture, was matched to the photographic metadata. Materials, including the brick, glazing and natural stone, were developed in discussion with Bennetts Associates so that what the images show is what the architects specified.

4 | Verified view production

Images were rendered at the full pixel dimensions of the host photograph and composited in Adobe Photoshop, following Appendix C of the London View Management Framework. Rendered at matching dimensions the model drops into place in exactly the correct position; the operator's job from there is masking foreground buildings back over it.

Each view carries an annotated graticule: thick lines at ten degree increments, intermediate lines every degree, measured away from the optical axis, with four red arrows marking the horizontal and vertical position of that axis. From those a reader can take the horizontal and vertical field of view directly off the image, tell whether vertical rise or shift was applied, and tell whether the frame has been cropped.