NewsWorksSoftwareTextBioContact
background image

Preparing for Tracker in February

2006-01-20

I’m currently preparing for another tour with Gitte Bastiansen and Frode Thorsen in February. This time we’ll be giving concerts at schools north of Lillehammer in the Oppland region. The line up is:

Frode Thorsen – recorders
Gitte Bastiansen – dance with sensors
myself – live electronic processing

We have been using the DIEM Digital Dance system flex sensors attached to the body with wireless transmition of data to me. The sensors are starting to get worn out after approx. 50 concert + rehearsals and development. I’ve checked the possibilities of getting new sensors. They are more expensive than I expected: DKR 250 – a piece (that’s the equal of NOK 275 -). Add to that 25% for customs when importing. We probably need 20 or so new sensors so this adds up.

I did some searching around today to find information on various other sensors prices etc. in case we could find a cheaper solution but they generally seem to be costly.

I also found an interesting paper online by Lisa K Simone and Derek G Kamper on Design considerations for a wearable monitor to measure finger posture from Journal of NeuroEngineering and Rehabilitation suggesting that Flexpoint sensors might be both the cheapest and most reliable. They tested a number of different sensors and these were the only ones that provided more or less repeatable values over time. Their general comment on available sensor systems is worth taking into consideration:

Bend sensors are used in a number of university and home projects despite our findings that most are not repeatable for moderate to fine resolution measurements. Instead most are appropriate for binary ON/OFF applications or applications that do not require high resolution or highly repeatable results.

Earlier on we’ve attenpted to use the bend sensors as continuous controllers in various ways with varying results to say the least. The sensors provide MIDI CC messages with a supposed range of 0-127. In real world the range tend to be much more limited with a total range somewhre between 10 and 20 and also drifting over time. I’ve supposed that the drift was caused by variations over time of how the sensors were positioned at the body. But according to Simone and Kamper the electronic properties of the bend sensor itself might change over time and quite rapidly so:

The first sensor evaluated was the Abrams-Gentile Entertainment Inc. (New York NY) sensor patent #5 86 785.000 Attempts to measure repeatable bend resistance versus calibration tube diameter failed because the measured resistance decayed over time. The Abrams-Gentile sensor exhibits the most common behavior that we will refer to as Type A behavior and it appears in Figure 2 as line “AG”. The sensor reached a peak resistance value just as it was wrapped around the calibration tube with an immediate decay in resistance over time. We expected that the sensor values would be constant however the drift in measured resistance prevented an accurate and repeatable measurement of bend. To eliminate other potential sources of error the analysis was repeated on ten other sensors and the problem finally isolated to the sensors by testing each directly using an ohmmeter.

The average decay in resistance while on the tube was computed. After 30 seconds the average error for the Abrams-Gentile sensors was 9.5% of full scale and 24.4% of step function rise resistance (Table 1). The Abrams-Gentile sensor never settled on a final resistance value but over an extended two-day data collection session continued to slowly decay. While these sensors are appropriate for many applications such as position detectors and indicators of gross movement we determined that they are not appropriate for accurate and repeatable measurements of finger flexion.

Considering my own experience with the DIEM system this makes a lot of sense. We were hardly able to use those sensors in a meaningful way at all until I started calculating running means of input values and deviations from those means and used normalised versions of those values to control musical parameters. For this I used the running average abilities of lp.stacey from the Max library Litter Power Package by Peter Castine.

If time permits I’ll try to completely reconsider how to use and interprete the sensors for the upcoming tour.

Please note that this is not meant to be a harsh critiique of the DIEM sensor system in general rather an aknowledgement that expectations towards the quality of the data from the flex sensors have to be lowered quite a bit compared to what we hoped to be able to achieve initially. Apart from that the DIEM system have proved very stable and reliable. We’ve never had any comunication problems of any sort between the sender and the transmitting system. This part seems to work flawless.

I have another major addition to work on as well: I’ve got myself a Coninuum Fingerboard and I’m planning for it to get it’s debut during this tour.

Up

comments powered by Disqus

 
 
 

Creative Commons License Licensed under a Creative Commons Attribution 3.0 Norway License. Web site hosted by BEK.