Javicle - a JSON Video Composition Language
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
Jonathan Cobb d9b9ecd1d6 update docs 3 years ago
bin add more command line utilities 3 years ago
src add remove-track operation, refactor a bit 3 years ago
utils use filter_complex for overlay. fix tests. 3 years ago
.gitignore ignore test script, adjust log output 3 years ago
.gitmodules initial commit 3 years ago
LICENSE.txt add license 3 years ago
README.md update docs 3 years ago
pom.xml simplify json formatting. complicates parser, but it's a relatively clean solution 3 years ago

README.md

Javicle - a JSON Video Composition Language

JVCL (pronounced “Javicle”) is a JSON DSL for audio/video transformations.

Under the hood, it’s all shell commands: ffmpeg, mediainfo and so on.

JVCL provides higher-level semantics for working with these lower level tools.

A Quick Example

Say you want to split a portion of a video into ten-second chunks. With ffmpeg and bash, you might do something like this:

    INCR=10
    for ((i=10;i<130;i=(i+INCR))); do
      ffmpeg -i /tmp/my/source.mp4 -ss ${i} -t $((i+INCR)) /tmp/my/slice_${i}_$((i+INCR)).mp4
    done

With JVCL, you’d write this spec file and save it to a file (for example my-spec.jvcl):

{
  "assets": [ {"name": "src", "path": "/tmp/my/source.mp4"} ],
  "operations": [{
      "operation": "split",
      "creates": "src_split_files",
      "source": "src",
      "interval": "10s",
      "start": "10s",
      "end": "130s"
  }]
}

and then run it like this:

jvcl my-spec.jvcl

Yes, the JVCL is longer, but I think many would agree it is easier to read and maintain.

As the number of media assets and operations grows, hand-crafted shell scripts with magical ffmpeg incantations become ever more inscrutable.

JVCL is designed for readability and maintainability. JVCL will continue to evolve towards greater coverage of the full capabilities of ffmpeg.

Who is JVCL not for?

If you like GUIs, JVCL is probably not for you.

JVCL is not a replacement for Final Cut Pro or even iMovie.

Who is JVCL for?

JVCL is for people who like CLIs and automation.

JVCL is for people with relatively simple video composition needs (for now), since the range of operations supported is limited.

Concepts

In JVCL there are a few main concepts: spec files, assets and operations.

JVCL Spec Files

A JVCL spec file is just a regular JSON file that happens to contain a single JSON object, whose properties are assets and operations.

When you run jvcl on a spec file, it will load the assets, then perform the operations in order.

Unlike most JSON, comments are allowed in JVCL spec files:

  • A line comment starts with // and continue to the end of the line
  • A multi-line block syntax starts with /* and ends with */

Executing a JVCL Spec

To execute a spec stored in the file my-spec.json, you would run:

jvcl my-spec.jvcl

or use stdin:

cat my-spec.jvcl | jvcl

Scratch Directory

Output assets will be placed in the scratch directory, unless otherwise specified in the spec file. By default, JVCL will create a new temporary directory to use as the scratch directory. You can set the scratch directory explicitly using the -t or --temp-dir option:

jvcl -t /some/tempdir my-spec.json

Command Help

To view a list of all jvcl command-line options, run jvcl -h or jvcl --help

Assets

Assets are your media files: generally image, audio and video files.

All assets have a name and a path.

Input assets are defined using the assets array of a JVCL spec.

For input assets, the path can be a file or a URL. URL-based assets will be downloaded to the scratch directory. This can be overridden using the dest property on the asset.

Operations produce one or more output assets, as specified in the creates property of an operation JSON object.

For output assets, the path will be within the scratch directory. You can override this using the dest property on the creates object associated with the operation.

Asset Properties

Assets expose properties that can be referenced in operations. The properties currently exposed are:

  • duration: duration of the audio/video in seconds (audio and video assets only)
  • width: width of the video in pixels (video and image assets only)
  • height: width of the video in pixels (video and image assets only)

Operations

Operations are transformations to perform on the inputs.

An operation can produce one or more new assets, which can then be referenced in later operations.

Most of the operation settings can be JavaScript expressions, for example:

"start": "someAsset.duration - 10"

The above would set the start value to ten seconds before the end of someAsset.

Supported Operations

Today, JVCL supports several basic operations.

For each operation listed below, the header links to an example from the JVCL test suite.

scale

Scale a video asset from one size to another. Scaling can be proportional or anamorphic.

split

Split an audio/video asset into multiple assets of equal time lengths.

concat

Concatenate audio/video assets together into one asset.

trim

Trim audio/video; crop a section of an asset, becomes a new asset.

overlay

Overlay one asset onto another.

ken-burns

For transforming still images into video via a fade-pan (aka Ken Burns) effect.

letterbox

Transform a video in one size to another size using black letterboxes on the sides or top/bottom. Handy for embedding mobile videos into other screen formats.

remove-track

Remove a track from a video asset.

Complex Example

Here is a complex example using multiple assets and operations.

Note that comments, which are not usually legal in JSON, are allowed in JVCL files.

If you have other JSON-aware tools that need to read JVLC files, you may not want to use this comment syntax. The asset and operation JSON objects also support a comment field, which can be used as well.

{
  "assets": [
    // file -- will be referenced directory
    {
      "comment": "first video, already local",
      "name": "vid1",
      "path": "/tmp/path/to/video1.mp4"
    },

    // URL -- will be downloaded to scratch directory and referenced from there
    {
      "comment": "second video, will be downloaded",
      "name": "vid2",
      "path": "https://archive.org/download/gov.archives.arc.1257628/gov.archives.arc.1257628_512kb.mp4"
    },

    // URL -- will be downloaded to `dest` directory and referenced from there
    {
      "comment": "third video, will be downloaded",
      "name": "vid3",
      "path": "https://archive.org/download/gov.archives.arc.49442/gov.archives.arc.49442_512kb.mp4",
      "dest": "src/test/resources/sources/"
    },

    // Image URL
    {
      "comment": "JPEG image, will be downloaded",
      "name": "img1",
      "path": "https://live.staticflickr.com/65535/48159911972_01efa0e5ea_b.jpg",
      "dest": "src/test/resources/sources/"
    }
  ],
  "operations": [
    // scale examples
    {
      "comment": "scale using explicity height x width",
      "operation": "scale",            // name of the operation
      "creates": "vid2_scaled",        // asset it creates
      "source": "vid2",                // source asset
      "width": "1024",                 // width of scaled asset. if omitted and height is present, width will be proportional
      "height": "768"                  // height of scaled asset. if omitted and width is present, height will be proportional
    },
    {
      "comment": "scale proportionally by a scale factor",
      "operation": "scale",            // name of the operation
      "creates": "vid2_big",           // asset it creates
      "source": "vid2",                // source asset
      "factor": "2.2"                 // scale factor. if factor is set, width and height are ignored.
    },

    // split example
    {
      "comment": "split one asset into many",
      "operation": "split",            // name of the operation
      "creates": "vid1_split_%",       // assets it creates, the '%' will be replaced with a counter
      "source": "vid1",                // split this source asset
      "interval": "10"                 // split every ten seconds
    },

    // concat examples
    {
      "comment": "re-combine previously split assets back together",
      "operation": "concat",           // name of the operation
      "creates": "recombined_vid1",    // assets it creates, the '%' will be replaced with a counter
      "source": ["vid1_split"]         // recombine all split assets
    },
    {
      "comment": "append vid2 to the end of vid1 and create a new asset",
      "operation": "concat",           // name of the operation
      "creates": "combined_vid2",      // asset it creates, can be referenced later
      "source": ["vid1", "vid2"]       // operation-specific: this says, concatenate these named assets
    },
    {
      "comment": "re-combine only some of the previously split assets",
      "operation": "concat",           // name of the operation
      "sources": ["vid1_splits[1..2]"],// concatentate these sources -- the 2nd and 3rd files only
      "creates": "combined_vid3"       // name of the output asset, will be written to scratch directory
    },

    // trim example
    {
      "comment": "trim all of the assets that were split above",
      "operation": "trim",          // name of the operation
      "creates": {                  // create multiple files, will be prefixed with `name`, store them in `dest`
        "name": "vid1_trims",
        "dest": "src/test/resources/outputs/trims/"
      },
      "source": "vid1_split",        // trim these source assets
      "start": "1",                  // cropped region starts here, default is zero
      "end": "6"                     // cropped region ends here, default is end of video
    },

    // overlay example
    {
      "comment": "overlay one video onto another",
      "operation": "overlay",          // name of the operation
      "creates": "overlay1",           // asset it creates
      "source": "combined_vid1",       // main video asset
      "start": "30",                   // when (on the main video timeline) to begin showing the overlay. default is 0 (beginning)
      "end": "60",                     // when (on the main video timeline) to stop showing the overlay. default is to play the entire overlay
      "overlay": {
        "source": "vid2",              // overlay this video on the main video
        "start": "0",                  // when (on the overlay video timeline) to begin playback on the overlay. default is 0 (beginning)
        "end": "overlay.duration",     // when (on the overlay video timeline) to end playback on the overlay. default is to play the entire overlay
        "width": "overlay.width / 2",  // how wide the overlay will be, in pixels. default is the full overlay width, or maintain aspect ratio if height was set
        "height": "source.height",     // how tall the overlay will be, in pixels. default is the full overlay height, or maintain aspect ratio if width was set
        "x": "source.width / 2",       // horizontal overlay position on main video. default is 0
        "y": "source.height / 2"       // vertical overlay position on main video. default is 0
      }
    },

    // ken-burns example
    {
      "comment": "apply zoom-pan effect to image, creates video",
      "operation": "ken-burns",        // name of the operation
      "creates": "ken1",               // asset it creates
      "source": "img1",                // source image
      "zoom": "1.3",                   // zoom level, from 1 to 10
      "duration": "5",                 // how long the resulting video will be
      "start": "0",                    // when to start zooming, default is 0
      "end": "duration",               // when to end zooming, default is duration
      "x": "source.width * 0.6",       // pan to this x-position
      "y": "source.height * 0.4",      // pan to this y-position
      "upscale": "8",                  // upscale factor. upscaling the image results in a smoother pan, but a longer encode, default is 8
      "width": "1024",                 // width of output video
      "height": "768"                  // height of output video
    },

    // letterbox example
    {
      "comment": "increase video size without scaling, add letterboxes as needed",
      "operation": "letterbox",        // name of the operation
      "creates": "boxed1",             // asset it creates
      "source": "ken1",                // source asset
      "width": "source.width * 1.5",   // make it wider
      "height": "source.height * 0.9", // and shorter
      "color": "AliceBlue"             // default is black. can be a hex value (0xff0000 for red) or a color name from here: https://ffmpeg.org/ffmpeg-utils.html#color-syntax
    },

    // remove-track examples
    {
      "comment": "remove all audio tracks",
      "operation": "remove-track",     // name of the operation
      "creates": "vid2_video_only",    // name of the output asset
      "source": "vid2",                // main video asset
      "track": "audio"                 // remove all audio tracks
    },
    {
      "comment": "remove all video tracks",
      "operation": "remove-track",     // name of the operation
      "creates": "vid2_audio_only",    // name of the output asset
      "source": "vid2",                // main video asset
      "track": "video"                 // remove all video tracks
    },
    {
      "comment": "remove a specific audio track",
      "operation": "remove-track",     // name of the operation
      "creates": "vid2_video_only2",   // name of the output asset
      "source": "vid2",                // main video asset
      "track": {
        // only remove the first audio track
        "type": "audio",               // track type to remove
        "number": "0"                  // track number to remove
      }
    }
  ]
}

What’s up with the name?

I dunno, a cross between a javelin and an icicle? does that have any positive connotations? ok then...