🎬 docker-ffmpeg-converter

Dynamic Docker service designed to simplify and streamline media conversion tasks

🌟 Overview

docker-ffmpeg-converter automates the process of file conversion by:

1. Monitoring an input directory for new files.
2. Utilizing ffmpeg to perform specified conversion operations.
3. Optionally removing the source files post conversion.

Whether you are building a small pipeline or conducting background conversions of multiple files, docker-ffmpeg-converter efficiently handles it all. Run multiple instances and make your conversion process more versatile and robust.

🚀 Deployment

Deploying docker-ffmpeg-converter is a breeze with Docker Compose. Below is an example YAML configuration file:

version: "3.9"
services: 
  video-converter:
    image: ghcr.io/kennethwussmann/docker-ffmpeg-converter:latest
    volumes:
      # This is where we will add input files and get output files
      - ./data:/data
    environment:
      # See below for all options and their meaning. This is just the required set.
      - SOURCE_DIRECTORY_PATH=/data/input
      - DESTINATION_DIRECTORY_PATH=/data/output
      - GLOB_PATTERNS=*.webm
      - FFMPEG_ARGS=-y -fflags +genpts -i %s -r 24 %s.mp4

🏷️ Tags

This image is built for arm64 and amd64.

• latest - Latest stable release
• x.x.x - Specific version under Semver (See all versions)
• develop - Unstable pre-release development version

🔧 Configuration

Configure the container through environment variables. Here's a breakdown of what you can customize:

Variable	Required	Description
SOURCE_DIRECTORY_PATH	Yes	Directory watched for new files
DESTINATION_DIRECTORY_PATH	Yes	Directory for converted files
GLOB_PATTERNS	Yes	Comma-separated list of glob patterns the service will use to filter files in the SOURCE_DIRECTORY_PATH
FFMPEG_ARGS	Yes	Arguments for the ffmpeg binary that specify what to do. See ffmpeg docs. You can use placeholders %s, see below.
REMOVE_SOURCE_AFTER_CONVERT	No (default: false)	true or false. Whether or not to delete source files after successful conversion
SCAN_INTERVAL	No (default: 10)	Interval in seconds when the service will search for new files
FILE_UNCHANGED_INTERVALS	No (default: 3)	How many cycles the service will wait for new files to stay unchanged until conversion starts. See below for detailed info.

💼 How the Service Works

📂 File Monitoring and Conversion

1. File Detection: The service continually polls the SOURCE_DIRECTORY_PATH, filtering files using the specified GLOB_PATTERNS.
2. File Verification: New files are cached and checked for size stability, based on FILE_UNCHANGED_INTERVALS.
3. Conversion: Stable files are handed to the converter, building an ffmpeg command with FFMPEG_ARGS.

Example:

FFMPEG_ARGS=-y -fflags +genpts -i %s -r 24 %s.mp4
SOURCE_DIRECTORY_PATH=/data/input
DESTINATION_DIRECTORY_PATH=/data/output

Resulting command:

/usr/bin/ffmpeg -y -fflags +genpts -i /data/input/myfile.webm -r 24 /data/input/myfile.mp4

Note: The order of %s is vital. The first represents the source file and the second the destination.

4. Post-Conversion: Depending on REMOVE_SOURCE_AFTER_CONVERT, the source file may be deleted after successful conversion.

⏱️ Custom Intervals and Patterns

• Scan Interval: Adjust the polling frequency with SCAN_INTERVAL, defining the seconds between each search.
• File Unchanged Intervals: Customize the number of cycles to wait for file size stability using FILE_UNCHANGED_INTERVALS.

🔗 Pipeline example

Here is a more complex example on how to use mulitple instances:

version: "3.9"
services: 
  # Same as above
  webm-to-mp4:
    image: ghcr.io/kennethwussmann/docker-ffmpeg-converter:latest
    volumes:
      # This is where we will add input files and get output files
      - ./data:/data
    environment:
      - SOURCE_DIRECTORY_PATH=/data/input
      - DESTINATION_DIRECTORY_PATH=/data/output
      - GLOB_PATTERNS=*.webm
      # Convert *.webm files to .mp4
      - FFMPEG_ARGS=-y -fflags +genpts -i %s -r 24 %s.mp4

  # Another instance to take thumbnails from the *.webm files
  webm-thumbnails:
    image: ghcr.io/kennethwussmann/docker-ffmpeg-converter:latest
    volumes:
      - ./data:/data
    environment:
      - SOURCE_DIRECTORY_PATH=/data/input
      - DESTINATION_DIRECTORY_PATH=/data/output
      - GLOB_PATTERNS=*.webm
      # Take multiple thumbnails from *.webm files
      - FFMPEG_ARGS=-y -i %s -vf fps=1/4 %s_%04d.png

The output directory will then contain multiple thumbnails and a converted MP4 file of our source material. Because the containers run in parallel they finish quickly as both tasks can run simulatanously.

🎉 Conclusion

docker-ffmpeg-converter is a great solution for seamless media conversion tasks, providing robust customization and a simplified deployment process. Experience the ease of automation with this powerful Docker service.

Feel free to explore, contribute, or seek support. Happy converting! 🎬