conda create -n [myEnv]
conda activate [myEnv]
conda install --file Requirements.txt
pip install -r Requirements2.txt
python MusicVideoGenerator.py -songName [song.wav] -bpm [int] --output [outputPrefix] --dynamic [True/False]
- This program cleverly assembles your previously downloaded scaffold videos into a tempo synced music video based on a provided input song and bpm
- Analyses a .wav file (located in music/)
- Generates 3 seperate tempo and dynamically synced music videos out of random .mp4 videos located in videos/
- Combines them through the use of overlay FX into one video located in output/
- Applies chroma shifting to the final product to make the videos bleed into eachother for contrast.
- Louder sections of your input song have a higher chance of faster / more intense visuals and vice versa.
- Generate music videos from thin air.
Royalty Free:
There are also a small selection of videos in this repo to allow you to test the program. These are from pexels.
- Video is 720p, audio is 256kps AAC
Richard, this thing is so slow it's making my head spin. I could do it way faster in Generic X movie editing software!
- Look, fair. My counter to that is you can leave and eat a sandwhich, call a loved one etc rather than have to sit through the the video editing process.
- You may have luck speeding it up by using smaller pool of scaffold videos, or using the --dynamic False flag if you only need tempo synced visuals.
- My runtimes are generally ~60 - 70 minutes per super loud song (16gb RAM). More RAM allows more videos to be loaded into memory and speeds up the process significantly. I acknowlege that this is nowhere near optimal performance, this is my first proper project. I am more than happy to collaborate with this project going forward.
- A public link to this repo in the description would be appreciated :)
- Please get in contact: richardpienaar1@gmail.com