I’ll write it up asap This week I was finishing up preparing my two main deliverables for interim demo. The first is the filtration circuit I have been spending most of my time on. Unfortunately, due to delays on parts I ordered, I am still yet to have a fully functional hardware circuit, but I have finished building it and incorporated some tweaks as advised by Professor Sullivan.
The biggest was adding in an RC high pass filter to filter out unwanted DC offset. Here is the calculation I did to get the best values:
DC Offset calculation- In order to do this, we know that DC has a frequency of 0 Hertz and the lower end of the frequency range of most music is 20 Hertz. This means that the high pass filter needs to be as low as possible (in order to prevent distortions of the music) but still clear the 0 Hertz. We know the cutoff frequency of a high pass filter is given by 1/2piRC, and I wanted it to be less than 5 Hertz. Using the standard values of resistors and capacitors that I had available, and attempting to ensure that the capacitance is as small as possible to prevent charging time from adding delays, I calculated using my largest resistor value (100KOhms) and solved C = 0.637 microfarads in order to achieve this. However, because I don’t have a 0.637 microfarad capacitor, I calculated using a 1 microfarad capacitor and found a cutoff frequency = 1.59 Hertz.
Additionally, I began working on the second deliverable, scoring using STT. At this moment, I have made it so that we have a streaming STT model that can get a transcription of the user’s audio. The transcription is aligned with timestamps for each word, or in this case lyrics, and I am comparing that with the timed lyrics gathered from lyricsify. This process is not yet finished but will be done before demo.
Next week I will primarily be focused on doing demos, but I also hope to get the last parts delivered and built into the circuit during the week, even if it is not ready for the first day of demos.
Right now, my biggest challenge will be interfacing the circuit with laptop output and ensuring that the circuit is operating as expected. Additionally, if my parts don’t arrive soon I am a little worried that it will lead to a big time crunch towards the end. However, because of this, I am focusing most of my effort into perfecting and implementing scoring.