Honestly there is nothing much to update at all this week, since we are basically just rushing out our projects and trying to compile all the code together. I guess I’ll just go through briefly some of the problems that we are facing currently.
We need a method of classifying music into different kinds of emotions, so we have decided to extract certain features from the song such as the timbre, the intensity, the tempo and the pitch and then train a simple SVM model using these features. However, the main issue is that we have a lack of labeled songs to use, since a majority of background instrumental tracks online are copyrighted, making it extremely difficult to find songs that fit the genre we want, and are free to use.
We are also having some trouble integrating everything together as we want to be able to upload files unto a server and then change those files and get the server to send it back. However, renting a cloud server would cost money (which we do not have a budget for) and also requires a lot of backend work (which we do not have the time for).
I guess we have no choice but to work with the limitations to the best of our abilities and hope that our project turns out the way we want.