Concat Workflow Operation

ID: concat

The concat operation has been created to concatenate multiple video tracks into one video track.

For a concatenation of two video files to work, both files need to have the same format (timebase, resolution, codecs, frame rate, etc.). This workflow operation has two modes to deal with this restriction:

General Mode

No restriction on source tracks codecs

This will re-encode the videos first to the same format (framerate/timebase/codec, etc) before concatenation.

graph LR outro[outro: source-flavor-part-2] intro[intro: source-flavor-part-0] presenter[presenter: source-flavor-part-1] concat{Concat} outro --> concat intro --> concat presenter --> concat concat --> o_intro[intro] --> o_presenter[presenter] --> o_outro[outro]

The internal FFmpeg command for re-encoding is using the following filters: fps, scale, pad and setdar for scaling all videos to a similar size including letterboxing, aevalsrc for creating silent audio streams and of course the concat for the actual concatenation step.

This requires an output-resolution and an optional output-framerate for the pre-concatenation encode.

The automatically generated FFmpeg filter for this process does look like this:

-filter_complex '
  [b][0:a][c][silent][d][2:a]concat=n=3:v=1:a=1[v][a]' -map '[v]' -map '[a]'

Same Codec Mode

Requires the source tracks having the same format (same timebase/resolution/encoding, etc.)

If the same-codec option is specified to use this mode, the sources files can be arranged into one container losslessly without re-encoding first. This is often the case if the tracks came from the same camera/recorder for example.

This mode uses FFmpeg's concat demuxer, which puts all the video content into a single container without any re-encoding. The encoding profile then operates on the source in this container. If -c copy is used in the encoding profile, the complete concatenation is lossless.

The FFmpeg command for this is is:

-f concat -i videolist.txt

…where videolist.txt contains a line in the form file <path to video> for each source track.


This operation is quite similar to the compose operation. The only difference is that the input properties are not only limited to one source-flavor and source-tag. The operation supports multiple flavor and tags as input. To add multiple sources, add different keys with the prefix source-flavor-/source-tag- and an incremental number starting with 0. For example:

Alternatively, using the source-flavor-numbered-files option, the operation supports an undetermined number of ordered input files.

This is useful when the number of input files cannot be known in advance, such as chunked output files from some camera/recorders, and the names are ordered by number or timestamps and to be sorted lexicographically.

For example, the configuration could be source-flavor-numbered-files: multipart/part+source and the ordered input files:


Note that both methods of defining input files are mutually exclusive.

Configuration Keys

Key Required Description Default Example
source-flavor-part-X false An iterative list of part/flavor to use as input track. NULL presenter/trimmed
source-tag-part-X false An iterative list of part/tag to use as input track. NULL source-to-concate
source-flavor-part-X-mandatory false Define the flavor part-X as optional for concatenation. false true
source-tag-part-X-mandatory false Define the tag part-X as optional for concatenation. false true
encoding-profile true Encoding profile to use for the concatenation. NULL concat
target-flavor true Flavor(s) to add to the output track. NULL presenter/concat
target-tags false Tag(s) to add to the output track NULL engage-download
output-resolution true Output resolution in width, height or a source part NULL 1900x1080, part-1
output-framerate false Output frame rate in frames per second or a source part -1.0 25, 23.976, part-1
source-flavor-numbered-files false Files of this flavor are ordered lexicographically to use as input track. NULL multipart/sections
same-codec false All source files have identical formats. false true


Example of a concat operation in a workflow definition.

<!-- Add intro and outro part to the presenter track -->
    description="Concatenate the presenter track and the intro/outro videos.">
    <configuration key="source-flavor-part-0">intro/source</configuration>
    <configuration key="source-flavor-part-1">presenter/trimmed</configuration>
    <configuration key="source-flavor-part-1-mandatory">true</configuration>
    <configuration key="source-flavor-part-2">outro/source</configuration>
    <configuration key="target-flavor">presenter/concat</configuration>
    <configuration key="target-tags">engage-download,engage-streaming</configuration>
    <configuration key="encoding-profile">concat</configuration>
    <configuration key="output-resolution">1920x1080</configuration>
    <configuration key="output-framerate">part-1</configuration>

Encoding Profile

The encoding profile command must contain the #{concatCommand} parameter which will set all input and possibly filter commands required for this operation:

profile.concat.ffmpeg.command = #{concatCommand} \
  … #{out.dir}/#{}#{out.suffix}