Skip to content

seandkim/Montgomery-AI

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

62 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Montgomery AI: Automatic Guitar Transcription

Montgomery AI is a tool that converts guitar performance videos into tablature (tab) scores. Using AI technologies like Facebook's Segment Anything and Google's MediaPipe, it analyzes video and audio to identify which notes are being played. This project combines computer vision and audio processing to simplify guitar transcription.

Overview of Methods

Setup

  1. Download Segment Anything 2.1
  2. Download MediaPipe
  3. Navigate to root folder.
  4. conda activate montgomery
  5. python -m montgomery.main

To Run

  • pip install -e .: one time only
  • python -m montgomery.main

Version

  • python=3.12
  • torch=2.5.1
  • numpy=1.26.4

These can be downloaded via conda create -n montgomery python=3.12 numpy=1.26.4 matplotlib pytorch=2.5.1 torchvision torchaudio -c pytorch

Baseline

The result of the program was compared to ChatGPT. The following query was used: "How do you play the intro riff for "______" by _____? Give me the tab score"

Test videos

Youtube videos were downloaded via "https://cnvmp3.com/", "https://yt1d.com/en12/". The link can be found in input.json

About

Automatic Guitar Transcription using Meta Segment Anything and Google MediaPipe

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages