Nowadays there are so many stories in the music world about one artist ripping off another artist's song. Hopefully this software will save singers and bands legal fees by checking if their song sounds too similar to another song before releasing it to the public.

What it does

Based on an algorithm that compares the cord progression and pitch, we take two MIDI files and print out a percentage of how close these two songs sound. We also implemented way to check how similar the lyrics are as well by inserting text files of the lyrics.

How we built it

We built it in python and QTPy. We mainly used the algorithm to program the logic behind the melody comparison and came up with our own way to compare the lyrics.

Challenges we ran into

Algorithms are hard :(. We spent a lot of time just trying to understand this algorithm. Most of the team has also never programmed in Python before so it was a lot of learning. Indentation errors. Ew. API's for reading midi files are also poopy.(We are still talking to the interwebs about solving our current issue)

Accomplishments that we're proud of

It kind of works.

What we learned

Python. We will never complain about brackets again (curse you indentation errors). How to go from an academic paper about an algorithm to real code. Also since it was quite musically based, we learned music thingz.

What's next for Project241Firenze

Getting more precise MIDI file readings. Maybe getting rid of MIDI files, going to mp3 instead which the converts to MIDI. Having vocal recognition to automatically get lyrics without having the user input them. 12+ hours of sleep.

Built With

Share this project: