Vector Field Compositing

frame_02800-WEB SLIDER


Extending formation processes toward urban scales of intervention, the research investigates form generation tailored to urban context by developing techniques of abstracting architectural intention. Custom methods of information exchange, patterning, compositing and layering information rich expressions tailored to urban spaces.

Creating Vector Clouds

The basis for vector creation resides in the sequential aggregation of particulates. As particulates in suspension deposit against one another the location data is recorded with the nearest neighbouring particle as a vector. A cloud of vectors is ultimately created recording the process of formation of the entire deposition simulation process. This vector cloud provides the sequence of formation rather than the deposition geometry itself. (Figure XYZ – Growth and Translation of a Deposition)

Weighting Field Expressions

A range of vector fields can represent formation data from any number of simulations. For instance depositions were created to respond to areas of Thames foreshore geometry. Additional fields of vectors were created for programmatic preference, and urban connectivity. Together we created many abstracted views of how deposition occurred.

Our interests however were in how to hybridize the formation sequences. Rather than mix the geometries themselves our method involved mixing the creation sequence themselves, hence the necessity for vector field data. Vector fields were merged in space, and each field amplified in turn in steps, weighting the strength of field data per deposition type.


Compositing Competing Clouds


>> Download Github Repository for Vector Field Compositor
Composite Solution SpacesComposite Level of Detail



Credits & Special Thanks

AA DRL Studio Fall 2013

Course Tutor: Shajay Bhooshan – Zaha Hadid Architects

Consultation: Ashwin Shah  – Zaha Hadid Architects



Processing 2.0, Maya 2014

CPU: MacbookPro Dual Corei5@2.53Ghz

GPU: NVIDIA GeForce GT 330M, 256MB



vector, field, compositing, flow field, cloud, data, visualisation


>>References & Further Reading

Daniel Shiffman. Autonomous Agents. The Nature of Code. 2009.

Marshmellow Laser Feast. McLaren P1 Launch. 2013.

David Laidlaw. et all. Comparing 2D Vector Field Visualization Methods: A User Study. 2005.