For sprint 2, we set our goal to be:
In this sprint we looked into developing the “stick on a motor” drummer into a more humanoid mechanism that involves something similar to a wrist flick that can play along with a metronome. In parallel, we wanted to decompose the SparkFun audio sound detection chip in order improve its performance and add a filtering stage.
We realized last sprint that we spent a lot of time working, but we didn’t always have direction in our work. To combat that this sprint we decided to set goals at the beginning of each meeting to be accomplished by the end. This deliberate communication served both to help us stay on track and to keep us up to date on what everyone else was working on. It was really helpful to know what the other members of the team were working on and if they were having any challenges that they needed extra help with.
We accomplished:
During this sprint we successfully have a two axis of rotation snare drummer. It currently has a passive wrist flick however brake line has been ordered to make it a more active motion. This arm prototype was integrated with small DC motors and the SparkFun sound detection chip. This allowed software to calibrate to the new system and have the drumstick reflect the bpm of the metronome. A delay was calculated so the drummer would begin playing on beat and we are working on calibrating the motion so it stays on beat. Concurrently, the SparkFun chip was decomposed on a breadboard earlier in the sprint. We integrated the new breadboard circuit with the wrist flick snare drummer mechanism. There was a lot of experimentation with different filtering this week. We attempted several different order filters with different cut-off frequencies but we did not get any of the results we were expecting from the audio signal. In order to better understand the filters we were using we ran bode plots and made Matlab simulations of the filters in order to design them more deliberately. In addition we narrowed the scope of our project from being able to listen and detect the beat of any song, to being able to detect the beat of just a few songs. These songs were selected by looking at spectrograms of their frequencies to see which songs had consistent and repetitive kick drum tracks.
Overall, we had a productive sprint and made a lot of progress on both beat detection and the mechanical system. Something that we learned this sprint that we want to take forward is to try hard to never work alone! This will hopefully help us keep up team morale during the late nights and make sure we keep communicating.
In this sprint we looked into developing the “stick on a motor” drummer into a more humanoid mechanism that involves something similar to a wrist flick that can play along with a metronome. In parallel, we wanted to decompose the SparkFun audio sound detection chip in order improve its performance and add a filtering stage.
We realized last sprint that we spent a lot of time working, but we didn’t always have direction in our work. To combat that this sprint we decided to set goals at the beginning of each meeting to be accomplished by the end. This deliberate communication served both to help us stay on track and to keep us up to date on what everyone else was working on. It was really helpful to know what the other members of the team were working on and if they were having any challenges that they needed extra help with.
We accomplished:
During this sprint we successfully have a two axis of rotation snare drummer. It currently has a passive wrist flick however brake line has been ordered to make it a more active motion. This arm prototype was integrated with small DC motors and the SparkFun sound detection chip. This allowed software to calibrate to the new system and have the drumstick reflect the bpm of the metronome. A delay was calculated so the drummer would begin playing on beat and we are working on calibrating the motion so it stays on beat. Concurrently, the SparkFun chip was decomposed on a breadboard earlier in the sprint. We integrated the new breadboard circuit with the wrist flick snare drummer mechanism. There was a lot of experimentation with different filtering this week. We attempted several different order filters with different cut-off frequencies but we did not get any of the results we were expecting from the audio signal. In order to better understand the filters we were using we ran bode plots and made Matlab simulations of the filters in order to design them more deliberately. In addition we narrowed the scope of our project from being able to listen and detect the beat of any song, to being able to detect the beat of just a few songs. These songs were selected by looking at spectrograms of their frequencies to see which songs had consistent and repetitive kick drum tracks.
Overall, we had a productive sprint and made a lot of progress on both beat detection and the mechanical system. Something that we learned this sprint that we want to take forward is to try hard to never work alone! This will hopefully help us keep up team morale during the late nights and make sure we keep communicating.
|
|
|