Noah-Grimaldi / auto-gaming-montage-maker

This is a tool for editors that automatically edits video game videos/clips pertaining to Fortnite, Apex Legends, Valorant, Call Of Duty, Overwatch, Minecraft, R6, PUBG, Destiny and Rocket League in order to make them into montages based off eliminations, knocks, wins, goals, or punches. It can also remove silent portions and add automatic captions
18 stars 4 forks source link
moviepy python roboflow yolov8

Automatic Video Editing (auto-game-montage, auto-silence-removal, and auto-captions)

This is a tool for editors that automatically edits video game videos/clips pertaining to Fortnite, Apex Legends, Valorant, Call Of Duty, Overwatch, Minecraft, Rainbow Six Siege, Destiny 2, PUBG, and Rocket League in order to make them into montages based off eliminations, knocks, wins, goals, or punches. It can also remove silent portions and add automatic captions to your video file with a simple GUI.

FOR REGULAR USAGE:

Simple Instruction

Windows:

MAC:

Linux:

Examples/Tutorial

What specific video edits does the program do?

What games does it edit?

Acceptable audio/video types

Acceptable video resolutions/fps

Demo for Auto Gaming Montages

Demo for Silent Portions and Automatic Captions

FOR DEVELOPER USAGE:

Clone the repository with --recurse-submodules so PySimpleGUI's older version will be included:

git clone --recurse-submodules https://github.com/Noah-Grimaldi/auto-gaming-montage-maker.git

Run editor_gui.py, which programmatically references the other files or download Windows, MAC, or Linux executables.

Table for different game's model speed and accuracy rating

CSGO Model Minecraft Model Overwatch Model Valorant Model Other games (template matching)
Average Speed (On 1 minute clip) 35 seconds 25 seconds 1 minute 35 seconds 55 seconds
Accuracy (1-5 rating) 5 3 3 5 5 (Besides R6, which is a 3)

How does the program edit the games?

Roboflow datasets I used or created

Already created datasets:

Datasets I created:

Facilitated Testing

If you need to have test video examples from Youtube, this is a great tool on Github for obtaining the videos to use in your program for testing.

After writing this I'm looking into the legality of this tool and I'm not 100% sure the youtube-dl is legal, although it is a larger open source community project that has been on Github for over 3 years now, so I'm assuming it would have been taken down at this point if it were not (I could be wrong, this is just a heads up).

Options explained for threaded methods

For edit_video.py:

For video-remove-silence:

For video_transcribe_audio.py:

Dependencies

After you've activated the venv on your specified platform, run the command: pip install -r requirements.txt

Platform Support

Windows/Mac/Linux

Credits

Credit to @excitoon for video-remove-silence