MusicToVisualizedVideo

Batch converts folder of mp3 music files and creates output video with album cover , lyrics
( only if original file have it ) and audio visualization. Uses python 3.12 and FFMpeg.

How to use:

Download project (as zip or clone)
Install python 3.12 (or use portable if you like)
Download ffmpeg https://www.ffmpeg.org/download.html
add ffmpeg folder to os path or just drop ffprobe and ffmpeg files inside script folder
then

  pip install -r  requirements.txt

(or use venv if you like it, i am to lazy to describe)

python.exe mtvv.py  ./ ./out

or if you using nvidia and downloaded some font and placed inside folder (mine most used variant)

python mtvv.py ./ ./out --batch-size 30  --vrate 1550 --arate 128 --font NotoSerifJP-VariableFont_wght.ttf --frate 60 --codec h264_nvenc

Where - ./ is input folder and ./out is output folder (for a sample we assume you placed script directly to music folder).

usage: mtvv.py [-h] [--batch-size BATCH_SIZE] [--vrate VRATE] [--arate ARATE] [--font FONT] [--shuffle SHUFFLE]
               [--frate FRATE] [--codec CODEC] [--vis-type VIS_TYPE] [--test] [--wavecolor WAVECOLOR]
               [--wavecolor2 WAVECOLOR2]
               input_folder output_folder

positional arguments:
  input_folder          Folder containing MP3 files
  output_folder         Folder to save MP4 videos

options:
  -h, --help            show this help message and exit
  --batch-size BATCH_SIZE
                        Number of tracks per video (adequate max value is 30, default: 25)
  --vrate VRATE         Out video bitrate in kbits (default: 550)
  --arate ARATE         Audio bitrate in kbits out video (default: 192)
  --font FONT           Font file: default = arial.ttf
  --shuffle SHUFFLE     Set to 1 to shuffle input list.
  --frate FRATE         Video framerate (default 30).
  --codec CODEC         Codec, default - software encoding by libx264.For nvidia best - h264_nvenc.
  --vis-type VIS_TYPE   Visualization type: 0 for sphere showwaves (with geq), 1 for just showwaves, 2 for full-width
                        showwaves bottom visualization, 3 for top/bottom simultaneous visualization, 4 - avectorscope.
                        (default: 0)
  --test                Run in test mode - process only 60 seconds of each track
  --wavecolor WAVECOLOR
                        Wave color in hex or from ffmpeg color table (default: album art dominant color)
  --wavecolor2 WAVECOLOR2
                        Secondary wave color in hex or from ffmpeg color table (default: 0x9400D3)

How it works:

Its grabs files from input folder, create output folder with temp inside.
Next its analyze first input batch of files and extract metadata with album cover and lyrics.
Lyrics converted to long transparent image thats will be added to chunk segment output temp video file (you can check it then its fully process first file in first batch).
Then ffmpeg combines all of it and add audio visualisation to segment and proceed next mp3 file.
After all files in batch is processed, scripts calls ffmpeg to concat segments to final batch output, write it to processed_files.json and iterate to next batch of files.
If you stop it after batch, on a next launch with same output folder, it will check for processed_files.json and ignore already batched files.
To stop it doing that, just remove processed_files.json.