Musical Note Identifier

User interface and guitar notes played
Recorded audio: guitar playing chords C, D, E. 1) Chromagram 2) Binary chromagram, filtered by intensity 3) Spectrogram of audio signal

Inspiration

We are three UBC Physics students who have interests in music, audio, the physics of sound, and signal analysis (and are always singing off key)!

What it does

Our program will allow the user to press the Start Recording button to record an audio signal and then click Display Notes to show the notes that were recorded. Lastly, the user can click Generate Chromagram to see a visual of the audio you recorded in a chromagram!

How we built it

We started with using the Librosa Python package to analyze audio signals and generate meaningful pitch data. We started with analyzing a single, constant note and then worked up to identifying multiple notes within a chord and a varying audio signal. We implemented the ability to record an audio signal to process instead of using preprocessed (clear) audio data. Finally, we designed a GUI using PyQt5 that allows the user to take in a recording in a desktop environment.

Challenges we ran into

We ran into some issues with the sensitivity of our audio signals, signal noise and designing the GUI.

Accomplishments that we're proud of

Making a GUI for the first time!

What we learned

We learned more about Fast Fourier Transforms and time-frequency analyses!

What's next for Musical Note Identifier

The next steps for our project is to make the GUI cooler and to eventually be able to analyze complete songs, which is a big jump in complexity to be able to separate frequencies played by different instruments and isolate their notes.