Keeping it real

Natural products are highly valued by consumers yet it has been difficult to reproduce their properties fully in synthetic materials. The key, say the NPL, is knowing how we perceive naturalness  

Natural products are highly valued by consumers yet it has been difficult to reproduce their properties fully in synthetic materials. the key, say the NPL, is knowing how we perceive naturalness

Scientists at the National Physical Laboratory (NPL) are leading a multidisciplinary team working towards producing the world’s first model that will predict how we perceive naturalness. The results could help make synthetic products so good that they are interpreted by our senses as being fully equivalent to the ‘real thing’, but with the benefits of reduced environmental impact and increased durability.

It is often easier and cheaper to manufacture imitation or synthetic materials than to mine or harvest their natural equivalents such as wood, stone or fur. But consumers may reject products made from materials which don’t look or feel natural, and are often willing to pay a higher price for something which they regard as natural. However the highest price is that paid by the environment when materials are drawn from a diminishing pool of natural resources.

The Measurement of Naturalness project, or MONAT, involves measurements by physicists, measurement scientists, neuroscientists and psychologists to probe and unravel some of the secrets of the human perceptual process. This is unique and important work that nobody has undertaken before. It will lead to a new understanding of the ways in which the physical properties of materials and objects interact with, and are interpreted by, our sensory and cognitive systems (Figure 1) and therefore has implications not only for naturalness, but many other perceptual parameters, such as ripeness, cleanliness and quality.

The MONAT project is exploring each stage in this process, in order to develop a fuller understanding of how the human perceptual system functions and provide a complete link from the physical (objective) properties of a material right through to the perceptual (subjective) impression of whether it is natural. When we look at and touch a surface, for example, interactions between the material and the sensory transducers in our skin and eyes generate sensory impulses, which then pass along nerve fibres to the brain. The strength of the signals depends not only on the physical characteristics of the material, such as its roughness, colour and texture, but also on factors such as the sensitivity of the sensory transducers and the environmental conditions. Surface structure at the nano-scale, for example, will not be sensed by our fingertips, but may affect the visual appearance. Once they reach the brain, the nerve impulses are combined and interpreted to generate the final perceptual response.

The multidisciplinary team is developing and applying innovative, leading-edge techniques and knowledge from the fields of metrology, psychophysics, neuropsychology and mathematics to evaluate relevant physical characteristics of a surface, such as its reflectance, texture and roughness, and to link these to what’s happening in the brain when a person sees or touches that surface. The information gathered will underpin the development of software models that bridge the gap between measurement of the physical attributes of a material and the human perception of whether that material is natural. These models can then be used to predict human reactions to new materials and objects. Importantly, new measurement and analysis capabilities are being developed for properties that it is currently not possible to evaluate (e.g. visual texture and multi-angular appearance of materials) and adapting leading-edge methods for use in novel ways (such as the use of functional neuroimaging techniques to establish dynamic neural pathways and interactions between these). The project is also using cutting-edge MEMS technology to develop a new measurement tool for touch, based on a bio-mimetic approach.

Several different types of material have been chosen for study, selected in each case to cover a range from ‘natural’ to ‘synthetic’. Investigations in the first phase of the project have focussed on wood and wood effect materials: vinyl, veneer, laminate, photocopies and real wood. The real wood was subjected to varying degrees of processing i.e. sawn, sanded, weathered, oiled, waxed or varnished. In order to ensure consistency of presentation for the perceptual tests, and avoid any clues being given by the edges of the samples, they were mounted in grey plastic boxes, each with the same viewing aperture of 80 x 80mm.

Various physical properties that are considered by experts to be relevant to the

Figure 1: The perceptual process

perceived naturalness of wood materials have been measured: surface friction in several directions; surface topography; hardness; visual texture at several angles; average colour; and gloss. Many of these measurements were performed using special measurement systems that are being developed and improved within the project to ensure both that all relevant information can be determined and that the measurement can be made at a resolution and dynamic sensitivity appropriate for comparison with human sensory systems.

Friction coefficients, for example, were determined using an instrument developed at NPL to measure the frictional interaction between the operator’s finger and the sample surface using the same movement patterns as used in the psychophysics and neuropsychology experiments within the project. NPL is also developing a novel multi-spectral goniometric system for measurements related to visual appearance. This is known as IRIS/GASP (image replication imaging spectrometer and gonio-apparent spectrophotometer) and is able to capture spatial, spectral and texture information across the full sample area and to examine changes in these characteristics as a function of the angle at which the surface is illuminated or viewed.

In the psychophysical and neuropsychological experiments observers are required to make visual, tactile and visuo-tactile assessments of the samples under strictly controlled environmental conditions. The psychophysical experiments have used four different protocols (labelled scaling, magnitude estimation, ranked ordering and a binary decision task) all of which have been found to yield highly consistent results in the first phase of the project (on wood). A binary decision task, involving assessment of the sample as either natural or unnatural, was also used for the neuropsychology investigations, which were carried out using a 3T functional magnetic resonance imaging (fMRI) scanner.

Figure 2: Schematic of the multi-spectral goniometric system developed at NPL, known as IRIS/GASP

In the first stage of the modelling, data for all the physical measurements (i.e. the outputs from the various visual texture algorithms and for the touch-related parameters) have been categorised using a variety of linear and non-linear regression techniques, to determine correlations with the results of the psychophysical tests. Although considerable further research is needed, early indications are that it seems possible to link the physical measurements with the degree of perceived naturalness determined by human subjects, which is the primary objective of the project (see Figure 3).

The outcomes from this project could have a great impact on the development of synthetic replacements for materials such as wood, animal skin and furs, marble, stone and plants. Applications across this range are manifold and range from wood used for flooring, furniture, vehicle dashboards, paneling and staircases, to skin for prosthetic limbs, or plants and flowers used in design and decoration. They may also be useful in the development of improved virtual reality systems for recreation and training purposes (e.g. flight simulators and surgical training).

Of the reasons why manufacturers would be interested in synthetic materials

Figure 3: Partial least squares regression (R2=0.87) on wood samples. The x-axis shows the visually-perceived naturalness and the y-axis is the predicted perceived naturalness as calculated from an optimised model based on measurements of the physical visual texture of the surface. Each point on the graph represents a different wood sample. Red indicates those samples generally assessed as being ‘definitely not natural’, purple are ‘probably not natural, turquoise are ‘probably natural’ and green are ‘definitely natural’.

one of the most obvious is the environmental impact. Natural products come from finite resources and increasingly have a negative impact on the environment and communities. Synthetic materials come from more manageable sources where supply can be predicted and controlled, providing a reliable source. While it could be argued that an increased number of synthetic products will only lead to more landfill as consumers find it easier to replace things, it is worth pointing out that synthetic materials can be engineered to be more robust and designed to keep. For instance, you could have a mahogany table that looks and feels real but is impervious to natural damage such rot or wood worm. Additionally with a synthetic material it is easier to obtain an exact replica to replace part of a larger object should it become damaged – for example, the damaged door of a wardrobe could be replaced with an exact copy of the original, thereby extending its life.

In conjunction with the MONAT project, NPL undertook a real-time experiment at the Royal Society’s Summer Science Exhibition this June.  The public were invited to touch and feel 20 wood and wood effect samples and assess whether they were real or not; more than 1000 experiments were completed during the 4 days of the exhibition. The data are still being analysed but initial results show that some samples are fairly consistently identifiable as natural or synthetic wood while others cause much more confusion. NPL is currently investigating how these findings relate to the physical properties of the different samples and will correlate the outcomes with those from the more rigorously controlled psychophysical investigations in the MONAT project.

As well as the real-time experiment the exhibition included a range of interactive exhibits that explore the perceptual process. The first of these showed how we can use body parts to measure an object, as the ancient Egyptians did with the cubit, a standard measure related to the Pharaoh’s arm length. There were visual, tactile and auditory experiments designed to demonstrate the limitations of the senses as measurement devices, by exposing how perceptions can be fooled by illusions. Videos highlighted how the use of Magnetic Resonance Imaging (MRI) brain scans is helping us understand the perceptual process, by allowing researchers to discover which areas of the brain are stimulated when people carry out specific tasks, such as using their vision and touch senses to explore natural and non natural wood samples.  Following the successful launch at the Royal Society’s Summer Science Exhibition it is now planned to use it at other venues. All the data collected will help to ensure that the models being developed within MONAT are as robust as possible.

The MONAT project is part-funded by the EU under the ‘Measuring the Impossible’ Programme. Other projects are investigating subjects as diverse as eyewitness memory, emotional response to computer games, measuring body language and understanding how music-induced emotions are processed in the brain.

The full MONAT research group is made up of:

• National Physical Laboratory – making the physical measurements of materials being studied, developing new measurement instrumentation, and modelling the link between physical properties and perception
• Trinity College, Dublin – neuroscience investigations e.g. using functional MRI scanners to measure areas of the brain that respond to particular sensations such as surface roughness or colour
• Unilever Research (Port Sunlight Laboratory) – the consumer products company is using its research expertise to bring together the different disciplines involved in the project and ensure that the physicists, neuroscientists, and psychologists all talk the same language
• Parc Científic de Barcelona, Universitat de Barcelona – psychology investigations into how different people make judgements of naturalness
• Laboratoire de Physique Statistique, Centre National de la Recherche Scientifique and Laboratoire d'Electronique et Technologie de l'Information – developing a biomimetic finger to reproduce the touch perception of a human fingertip
• Biometris, Wageningen University and Research Centre – data analysis and modelling