During the lab work sessions this week we mostly focused on the design review presentation for next week. We worked as a group on the slide deck and we sent the slides to our TA for feedback on Friday. All 3 of us worked independently on our respective aspects of the game – Stephen worked on the hardware, Shreya on signal processing, and George on graphics programming.
One concern we currently have is regarding the piezoelectric sensors, as an issue with their detection accuracy could pose a minor setback. As mitigations to this concern, we ordered several types of piezoelectric sensors in case one fails to meet our expectations and we also plan to develop a small prototype as soon as possible in case another approach needs to be taken. Stephen ordered these parts earlier this week and they arrived on 2/17, so we plan on doing small tech demos with these parts after the design presentation to make sure that they work as intended.
Another risk that we discussed this week was related detecting beats in the input music. We are going to use a Python module called Librosa for detecting the beats per minute (BPM) of the audio, but we are a little unsure to what extent this BPM measurement will help us identify the beats of the song. In this case, a “beat” is more than just the pulse from the BPM – it also refers to the groove and rhythm of the song. We want to sample the song at regular intervals relative to the BPM of the song, and use these samples to identify important beats. These beats can be used to generate the beatmap, which is the series of notes that rain down towards the player while their song is playing. This week our team discussed strategies for beat tracking and whether we should reduce our target metric of 80% for beat accuracy to make this goal more achievable; but we haven’t reached a conclusion on this yet.
Overall, we are on schedule as a team to have an MVP by the end of March, and there are no updates to out Gantt chart in the past week.