Winter Week 6 - Testing and Filtering

This week we spent our lab time testing the microphone input to the LCDK. We spent a lot of time trying to make sure the microphone worked properly, and discovered that there was a lot of additional noise compared to the direct line input. We used an online tool to find the appropriate low-pass filter to use. We selected a Butterworth IIR filter, which was implemented using the bilinear transform method.

In order to test the filter, we began the process of scanning the cutoff frequency to determine the right amount of attenuation. We tested around 1000 - 2000 Hz and found that the filter helped reduce a large amount of noise. In the next lab session, we plan to fine tune the filter to maximize the chord recognition accuracy.

Lastly, my partner and I decided to output the chord sequence to a file in a simple format before trying to encode into either MIDI or XML. We found that these formats are somewhat difficult to implement in real-time due to the memory constraints of the LCDK. Instead, we are going to use a readable text-based format that can also be output to the console.

We have reached our week 7 milestones and only need to do some minor testing and fine-tuning to try to improve the chord recognition accuracy. We are also close to completing our final milestones.