a gesture controlled instrument mounted on a pole dance pole
Ghent University College & Logos Foundation
By now, we have designed and build a good 35 musical robots, together making up the <M&M> machine orchestra. We consider it a major achievement and are indeed pretty proud about its realisation and its musical versatility in very different musical contexts. However, the entire orchestra occupies quite a large floorspace and thus started to compromise more and more our activity and research into experimental dance and gesture controlled music making. Mainly for this reason we limited out dance productions always to a maximum of two dancers occupying minimal floor space. With these considerations in mind, it should'nt come as a surprise that I became highly interested in pole-dance. In this type of virtuoso and highly developed dance, movement is strictly confined to the vertical space delimited by the central pole and the length of the stretched body around it. The fact that there is a slight erotic touch to it and that it often is combined with striptease, came as an extra benefit, particularly with regard to applications for our gesture interfaces using either sonar or radar, that do work much better and reliable on naked bodies than with clothes. So we went on along this line and decided to develop a special pole equiped with radar gesture sensors for movement capture in a vertical plane.
The pole as we made it conforms to the pole-dance standards: diameter 51mm, stainless steel brushed with a fine grain. The length can reach a maximum of 4m50. The transducers and electronic circuitry is mounted on the top of the pole.
1. three microwave radar devices.
These make use of microwave sensors produced by Waldmann under typenumber HFMD10. Their operating frequency is specified at 9.35GHz. The emitted electromagnetic waves are reflected by reflective surfaces and if in movement, these will cause a Doppler shift between emitted and reflected signal. As to human bodies, the most reflective surface is the naked skin. We performed measurements showing that a pullover gives a damping in the order of at least 15dB, thus reducing the resolution of the interface effectively with a factor 6 or worse. Even a shiny T-shirt, reduces the amplitude of the reflections with 6dB. Hence our advise to always perform naked with this (and in fact, any doppler based) invisible instrument. For further details: see Picradar.
In our pole setup we use three circuits as described above, each set to a different MIDI-channel. The sensors are mounted on top of the pole under 120 degree angles and facing downwards.
2. Adapter board to connect the outputs from these 3 devices to 4 port midi interface connected to the (laptop) computer. This board also houses a linear stabilized power supply for the transducers. These require 15V dc and draw about 50mA each. This brings the total power requirements, including some safety margin to 15V/ 250mA, or no more than 4Watt. Standard 5pole 180° DIN connector cabling is used for the interconnections.
The software consists of two levels: a first level, encoded in the PIC microcontroller. This code samples the signals from the transducers at the required sampling rate and processes this data such that we get access to following parameters:
The second level is part of our <GMT> programming environment. Here the data from the three interfaces are combined in an attempt to make sense and gain relevant information with regard to gesture input from the pole dancer.
The relevant procedures, including many more features than we can describe here, are integrated in our DLL library g_lib.dll. The source code is available. The required exported functions are:
GetPicRadarPointer (devicenumber) AS DWORD
this returns a pointer to the structure defined for each transducer. This structure behaves like an object and sets and returns all relevant data, operational mode and parameters. Devices are numbered 0 to 2.
The derivation of meaningful information with regard to the gesture input is based on the same considerations and gesture typology as described in "Gesture controlled virtual musical instrument" (1999).
The data acquisition card we use has a 12 bit resolution. Since the normal noise level of the system is around 4, we should consider the last 2 bits as irrelevant. The practical resolution, and therefore the precision of the system, cannot be better than 10 bits. There is absolutely no advantage in using a higher resolution data acquisition card for this application. At least, unless lower noise or higher power microwave devices become available...
The <SensoPole> suite is a collection of studies for a naked performer, in which we tried to explore different ways of mapping gesture and dance information on musical activity produced by our robot orchestra, composed of following robots:
Everytime we finish a new robot and
add it to the M&M orchestra, we add a new chapter in this suite of pieces.
The pole and associated equipment is available for any competent composer wanting to develop a piece or performance using it. Since the use of the instruments requires software to be written, it is highly advisible to study our <GMT> software and its functionality with regard to this instrument. As an alternative, the public domain language PD can be used as well. Usefull PD patches have been developed by our collaborators Kristof Lauwers, Yvan Vander Sanden and Johannes Taelman. They are available upon demand.
(1) This project is part of the ongoing research of the author in gesture controlled devices over the last 30 years. Earlier systems, based on Sonar, Radar, infrared pyrodetection and other technologies are fully described in "Gesture controlled virtual musical instrument" (1999), in "Quadrada" (2003), "picradar" (2004) as well as in his doctoral dissertation 'An Invisible Instrument' (1993). Artistic productions and compositions using these interfaces and devices have been: <Standing Waves>, <Holosound>, <A Book of Moves>, <Virtual Jews Harp>, <Songbook>, <Slow Sham Rising>, <Gestrobo>, <Technofaustus> , "PicRadar Studies" etc.
(2) People interested in buying the sensors as described here (fully functional and inclusing the programmed PIC) can take contact with the author. Cost, depending on the version required start at 350€ for a single transducer. A complete system, consisting of 4 sensors, a powering box and power supply costs ca.1500€
(3) Microcode for the PIC in this project was written by Johannes Taelman.
First published on the web: 10.05.2007 by dr.Godfried-Willem Raes