Into the thick of it!
Did a spike this week on the Librosa API documentation this week. Very dense stuff; Librosa seems to be a powerful and extensive library, but I think there are only a handful of specific features we will need to achieve the goal of the project. I’ve been running some test programs, seeing what data I can pull from some of my own collection of MP3s. Those little tests have given me a decent understanding of how librosa works on its own, so now it comes down to seeing how we can leverage it to meet our needs. Another member of my group is also doing the same with the Flask web framework, as we’re interested in making this application a web application.
Fleshing out the details
A member in my group who is savvy in front end design drew up some ideas for how the user interface could look. From my point of view, I think the concepts are great. It seems the direction we’re going right now is to use a visual aid, like a pie chart or bar graph, to denote the confidence that the program has in its genre classification of an uploaded audio clip. Nothing overly flashy, but bold and easy to read. That, to me, is actually more important in terms of accessibility. Front end is not really my forte, however, but I’m eager to learn and contribute to that portion of the project in any way that I can. As I mentioned before, though, I’m much more interested in the machine learning and audio preprocessing aspect of this project.
Moving forward
Due to some unforeseen circumstances, our group will be formally submitting our project plan tonight after our team meeting, so that will be the first order of business. After that, I have an inclination to start figuring the logistics of the audio processing portion of this assignment.
Also -> Countdown to baby: 23 Weeks
