Week 6 - Midterm Project
Me and Daniel decided to work on a spooky Theremin-like musical instrument, we used two time-of-flight sensors to get the user input and mapped those readings to various parameters of the sound. We started by implementing simple audio synthesis using the Mozzi library since it had added functionality and classes compared to the built-in tone function of the Arduino Nano. Then we ordered an I2C multiplexer to get a reading from both of our sensors. We noticed that the multiplexer for some reason detected two I2C breakout boards on all the ports. I searched around the net and found that people had similar issues, seems to be a bug in certain chips but it apparently does not affect the output (link to the forum this was discussed and screenshot of Adafruit support's message below). It took us a while to figure out a proper code that would get us a reading from both of the sensors.
Once we could get a reading from the sensors and had the ability to switch between them, we had to figure out what we wanted each sensor to do. We initially had lots of ideas in mind but then we faced the limitations of the Arduino Nano and the Mozzi library. But within their abilities, we impletmented delay and filter effects. Also a square wave was used to get a saturated and rich tone. One thing I noticed with our readings, was that when we mapped the reading to our target range, most often we wouldn't want our input range to have a minimum of zero or less than 40, but the normal map function would output unreliable results once it got an input value lower than the input minimum, so I wrote my own map function with a workaround:
1 2 3 4 5 |
float _map(float source, float min1, float max1, float min2, float max2) { source < min1 ? source = min1 : source = source; source > max1 ? source = max1 : source = source; return (source - min1) / (max1 - min1) * (max2 - min2) + min2; } |
Then we added a switch to control between several states, everytime the switch is pressed the state changes and once it reaches the 4th state it loops back to the first one. The first state uses an EventDelay which is equivalent to the normal delay in Arduino's library. Within the delayed period, we can hear a pitch drop. One of the sensors changes the interval between event delays and the other one changes the initial beginning pitch. The second one was our attempt at creating a scale and mapping the reading of our sensors to the notes within that scale. I personally expected a more defined pitch generation but I guess the Arduino Nano 33 IoT's hardware is not ideal for sound synthesis. Still we could get some interesting chiptune-like effects. The third state implements a lowpass filter, one sensor changes the notes on a scale and the other controls the cutoff frequency of our filter. The fourth state is a feedback delay, but as mentioned in the reference of the mozzi library the range of the delaytime is between 0 and 7 milliseconds, which is hardly noticeable to the human ear. Once the delaytime is below 20ms even if technically it is a delay, most often it is referred to as a chorus or a modulation effect.
This project in my opinion was a great introduction to sound synthesis and sound manipulation using the bare minimum. It was a challenge to be stripped off of all the musical advancements that we take for granted and work with a voltage and a speaker and a microcontroller. The limitations of the Arduino Nano definitely changed the course of our project development. The delay is not at all what one would expect from a delay effect but given the limited memory, it's not possible to have a long delay buffer and hence we had to work with what we had. Despite all the frustrations with things not working as we expected, I had lots of fun and eureka moments throughout the development of this project.
— Nima Niazi