Harmonic Split - Web Application Consisting of Guitar Tuner Made in JavaScript and AI Song Splitter - Spleeter Made in Python

About This Web Application:

Made by:

Frimu Aurel-Viorel

• GitHub

Nisipeanu Ionut

• GitHub

Made for the Scientific Communication Session 2023 Organized by:

Politehnica University of Bucharest - Faculty of Engineering in Foreign Languages

• English link - here
• Romanian link - here
• French link - here

Got 2nd place in category - Computers and Information Technology in English

Main Page

Guitar Tuner Page

Audio Splitter Page

Usage & Requirements:

Python Version:

Built and tested on Python 3.10.6

Operating System:

Built and tested on Windows 10 Pro version 21H1, should (in theory) work on most Windows 10 versions and Windows 11

Necessary Libraries / Modules:

• werkzeug.utils
• flask
• spleeter

How to Get Them?

• pip install werkzeug.utils
• pip install flask
• pip install spleeter

Necessary Changes to the Code:

Change UPLOAD_FOLDER = 'YOUR UPLOAD FOLDER GOES HERE' from line 10 of flask_main.py to the folder you want it to save the archives of the split songs - use os.path.join('app', 'subdir', 'dir', 'filename.foo') for system agnostic paths! (as of right now the app doesn't delete the archives or the songs after sending, will probably add in a future update)

The AI Model:

As of right now the file size limit GitHub has doesn't let me upload the model but it can be downloaded from here then the content should be unarchived twice using 7-zip and placed in the pretrained_models/5stems folder

Running it:

After running the flask_main.py file, the web application will be available at http://127.0.0.1:5000 / localhost:5000 so just navigate to that address in your browser of choice and start using the app, for the song splitter only you can use my audio splitter only found here

Recommended Specs:

It's better to use a machine with a beefier CPU and at least 8GB of RAM as the AI model runs on the CPU and uses around 4GB of RAM per song that's being split

Tested on a laptop with:

• Ryzen 7 5800H
• 16GB RAM / 32GB RAM