Ping: Midi Controller- false object detection

n0izmkr

This is my first stamp project and my first forum posting. I have looked through the other forum posts relating to the Ping, but have been unable to find posts relating to my particular problem. Please let me know if I need to provide any further details of my project.

I have a built a theremin-style midi controller, utilizing two Ping Ultrasonic sensors and the BS2. The controller works great in a "quiet" enviornment. However, when put into tight rehersal and performances spaces, it receives false hits.

The controller is played in a live electronica band- computers, drums, bass, trumpet, and trombone. The controller seems to be sporadically detecting hits from the trumpet and cymbals. The severety of the situation fluctuates, depending on the size/shape of the performance space.

When we tested one instrument at a time, the false hits occured in a patteren that could be related to the height of the performer and the direction in which he was playing. So, it does seem to be a result of soundwave collisions, or the receiver detecting 40khz waves from other sources.

Currently, the workaround for this problem is turning the sensors off with a switch (SPST and main power supply) when the controller isn't being played. I would like to find a more refined way to deal with this. I'm hoping there are small changes I can make to the physical design to prevent the false detection. I'm also hoping this won't need a drastic change in the physical housing, as the performer who I built this for has already spent a couple of months playing/learning the instrumet. I am hesitant to add more code, for fear of slowing down the response time. Yes, a dilemma. However, if someone has encountered this issue in their own projects, and knows a possible code solution, I will post a link to the code.

Possible solutions I have considered include adding another sensor, such as an IR switch, or adding acoustical treatment. I have researched acoustic treatment, but as of yet, have not found anything specifically rated for ultrasonic waves.

Measurements of the Ping boxes are: 3.5"x2.5"x1.5". The sensors are recessed 2cm inside the box.

Any advice, insight, etc would be greatly appreciated!

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Nicole Carroll
aka n0izmkr
www.nicolecarrollmusic.com

Phil Pilgrim (PhiPi)

You're definitely getting 40KHz interference from other sources. Cymbals are an almost ideal source of such interference, and a trumpet could easily have harmonics in the 40KHz range. You could try to shield the Pings from these sources, but I doubt you'd have much luck. One software strategy that might help is to take several readings, then use the median value as your distance reading. There's also a way to do a running median calculation, which I presented here. What's not clear, however, is whether the response time would be fast enough for your app.

-Phil

Franklin

You might be able to convert to IR or use IR to switch the pings in and out of the circuit.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
- Stephen

n0izmkr

I had considered implementing an IR switch, but unfortunately this would not solve the problem because in small spaces, the waves from the cymbals/trumpet are overriding the detection of the hand over the ping sensors, so there would still be false detection.

From what I understand, the PIR sensors only output high/low signal. I don't think I could replace the ping sensors with the IR sensors, since I am mapping the sensor's field of view to precise midi values.

Please correct me if I am wrong regarding the PIR output.


Regarding the running median- I'll have to think on this one. I'm not sure it would be the best solution for this application, as I need real-time, precise response, with near zero latency. However, thus far I have had no luck with material barrier solutions, so I'm going to give this a try.


Thank you for your input!


n

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Nicole Carroll
aka n0izmkr
www.nicolecarrollmusic.com

Phil Pilgrim (PhiPi)

You could also try measuring the hand distance as outlined here ... or wait for Parallax's laser distance sensor to debut. By using a white glove against a black background and diffuse front lighting, you could also key the instrument with the TSL1401-DB imager by measuring horizontal, rather than vertical, displacement. Your performer would have to get used to the new paradigm, though.

-Phil

Franklin

http://www.youritronics.com/infrared-theremin-musical-instrument/

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
- Stephen