Lib.rs

ffzap ⚡ is a simple, fast command-line tool for processing media files with ffmpeg. As it's multithreaded and can run as many tasks in parallel as your system can handle, it's perfect for converting, compressing, or editing audio and video files quickly and efficiently.

Because it uses ffmpeg under the hood, it supports any media file processing that ffmpeg can handle.

1. Installation
2. Usage
   • Practical examples
3. Speed comparison to alternatives
4. Requirements
5. All available options
6. Migrating to 1.0.0
7. License

Installation

Homebrew (macOS / Linux)

brew tap CodeF0x/formulae
brew install ffzap

For macOS, both x86_64 and aarch64 builds are provided. For Linux, only x86_64.

Winget (Windows 10 / 11)

winget install CodeF0x.ffzap

Only x86_64.

Cargo (Universal -- 32bit and 64bit)

(Install Rust first if you haven't already)

cargo install ffzap

Every architecture Rust has compile-targets for.

Building from source

(Install Rust first if you haven't already)

git clone https://github.com/CodeF0x/ffzap
cd ffzap
cargo build --release

Every architecture Rust has compile-targets for.

The ffzap executable will be under target/release/

Or Download a prebuilt binary

and add it to your path.

Usage

ffzap's usage is almost identical to ffmpeg, consider this simple example:

ffzap -i vids/test-1.webm -f "-c:v libx264 -b:v 1000k" -o transcoded.mp4

Mind that the ffmpeg processing options go into the -f argument (short for --ffmpeg-options), need to be passed as a string and without the file name.

With a single file it doesn't really make sense to use ffzap, so consider this more advanced example:

ffzap -i vids/**/*.{mp4,mkv} -f "-c:v libx264 -b:v 1000k" -o transcoded/{{name}}.mp4 -t 4

This command takes all videos in vids and its subfolders ending in .mp4 and .mkv, processes them using the options provided by -f and saves them to a (new) directory called transcoded, keeping the original filename and changing the file extension to .mp4 while processing 4 files in parallel.

For more info on the -o syntax, run ffzap --help. For more ffmpeg options, visit ffmpeg's documentation.

More practical examples:

ffzap --file-list files.txt -f "-c:v libx265 -preset medium -crf 23 -c:a libopus -b:a 128k" -o "Output/{{name}}.mp4" -t 2

ffzap --file-list files.txt -f "-c:v mjpeg -q:v 2" -o "Output/{{name}}.jpg" -t 6

ffzap --file-list files.txt -f "-i watermark.png -filter_complex [1]format=rgba,lut=a=val*0.3[watermark];[0][watermark]overlay=(main_w-overlay_w)/2:(main_h-overlay_h)/2 -c:a copy" -o "{{name}}_watermark.mp4" -t 2

ffzap --file-list files.txt -f "-vf scale=1280:720 -c:a copy" -o "{{name}}_resized.mp4" -t 2

ffzap --file-list files.txt -o "{{name}}.mkv" -t 2

Speed comparison to commonly used alternatives (based on GitHub stars)

Preface: I want to point out that with this comparison, I am not saying the mentioned tools are bad or inferior to ffzap. They're achieving a great job and are used by many. My personal issue with these tools is that I think they don't use system resources efficiently / don't allow the user to take control and this test aims to show that.

ffzap is not a magical tool that speeds up your batch encoding processes by simply increasing the amount of ffmpeg instances. You need to be aware of what your system can and can't do and then leverage ffzap to take advantage of that.

Test system:

• Intel Core i7 13700K
• RTX 3090
• 32 GB RAM
• NVMe SSD

Test Videos (29):

• 1440p
• 60 FPS
• H264
• AAC / 192k

Re-Encoding settings:

• Video:
  • Codec: H265
  • Quality: CRF 24
  • Preset: medium
  • Resolution: 1080p
• Audio:
  • Copy

Used ffzap command:

ffzap -i Videos/**/*.mp4 -f "-c:v libx265 -crf 24 -preset medium -vf scale=1920:1080 -c:a copy" -o "Out/{{name}}.mp4" -t 2

In this example, we're comparing ffzap to other tools on how they utilize the system's CPU when re-encoding a video. An i7 13700K can run two parallel 1080p encodings comfortably when limited to 12 threads per ffmpeg instance, and we achieve that by using -t 2 (--threads 2) in ffzap. This of course slows down each individual process, but because we're running two jobs at once, it makes up for it.

All other mentioned tools (except FFBatch AV Converter) run only one encoding job at any given time, resulting in the following total encoding times:

ffzap and FFBatch AV Converter are the clear winners here. Given the small difference in total encoding time between the two, ffzap could just be as fast if not even faster if I would invest more time in finding the sweetspot between ffmpeg instances and threads per instance. For the sake of simplicity for this example I didn't to that.

Requirements

• a working installation of ffmpeg
• (just for installing / building via Cargo) a working installation of the Rust programming language

Available options

$ ffzap --help
⚡ A multithreaded CLI for digital media processing using ffmpeg. If ffmpeg can do it, ffzap can do it - as many files in parallel as your system can handle.

Usage: ffzap [OPTIONS] --output <OUTPUT>

Options:
-t, --thread-count <THREAD_COUNT>
        The amount of threads you want to utilize. most systems can handle 2. Go higher if you have a powerful computer. Default is 2. Can't be lower than 1
        
        [default: 2]

-f, --ffmpeg-options <FFMPEG_OPTIONS>
        Options you want to pass to ffmpeg. For the output file name, use --output

-i, --input <INPUT>...
        The files you want to process

    --file-list <FILE_LIST>
        Path to a file containing paths to process. One path per line

    --overwrite
        If ffmpeg should overwrite files if they already exist. Default is false

    --verbose
        If verbose logs should be shown while ffzap is running

    --delete
        Delete the source file after it was successfully processed. If the process fails, the file is kept

-o, --output <OUTPUT>
        Specify the output file pattern. Use placeholders to customize file paths:
        
        {{dir}}  - Entire specified file path, e.g. ./path/to/file.txt -> ?./path/to/
        
        {{name}} - Original file's name (without extension)
        
        {{ext}}  - Original file's extension
        
        Example: /destination/{{dir}}/{{name}}_transcoded.{{ext}}
        
        Outputs the file in /destination, mirroring the original structure and keeping both the file extension and name, while adding _transcoded to the name.

-h, --help
        Print help (see a summary with '-h')

-V, --version
        Print version

Migrating to 1.0.0

In version 1.0.0, the following changes were made:

• --input-directory has been deprecated and replaced by --input.
• --input-file has been deprecated and replaced by --file-list.

1. Replacing --input-directory with --input:

Instead of using --input-directory, you now have to use --input to specify the files you want to process:

ffzap --input <files here> -f "<options here>" -o "<output pattern here>"

Note: The short form -i remains unaffected by this change.

2. Replacing --input-file with --file-list:

Instead of --input-file, use --file-list to specify a file containing a list of files to process:

ffzap --file-list <path to list here> -f "<options here>" -o "<output pattern here>"

For further details and motivation behind these changes, refer to issue 16.

License

You are free to:

• Use: You can use this software for any purpose.
• Modify: You can modify the software as you like.
• Distribute: You can distribute the original or lightly modified software only if you credit the original author ( Tobias "CodeF0x" Oettl). Selling of the original or lightly modified versions is not allowed.
• Sell or redistribute substantially modified versions: If you make significant changes to this software, you may redistribute or sell the modified version without crediting the original author.

Conditions:

• No selling of this software in its original form or with minor modifications.
• Credit must be given to the original author (Tobias "CodeF0x" Oettl) if redistributing unmodified or minimally modified versions.