Offline media conversion and prepared versions

Offline media conversion prepares another media representation before playback, separating production time from the later session’s delivery deadlines.

Offline conversion produces another media representation before the viewer starts using it. The work can take longer than the programme's duration without interrupting a live viewing session.

A prepared version can then be selected as an existing source. That shifts when the conversion happens; it does not remove the later requirements for access, delivery, decoding and presentation.

Prepared and session-generated output

Question	Prepared representation	On-demand playback conversion
When is the encoding work done?	Before playback is requested	During the playback job
Which deadline governs production?	The preparation workflow	The session's consumption needs
How is the output retained?	As an intended resource with its own lifecycle	As generated session output under delivery policy
Can another session reuse it?	If made available and selected appropriately	Depends on implementation and representation context

A download of the original file is a different operation from converting it. A client can also store a converted version for offline use, but storage location and conversion are separate decisions.

A defined original conversion example

For an SDR, 16:9 source, this illustrative command requests 720p H.264 video and stereo AAC:

ffmpeg -i source.mkv -map 0:v:0 -map 0:a:0 -vf scale=1280:720 -c:v libx264 -b:v 6M -c:a aac -b:a 192k -ac 2 prepared.mp4

It was not executed for an actual media file. The named encoders and format must be available in the tool being used. The command also does not implement an HDR-to-SDR colour conversion.

The selected input streams become a new output set. Extra audio, subtitles, chapters and attachments require their own decisions rather than being assumed to survive every preparation command.

Conditional time and storage estimates

If a 108-minute programme is processed at an average 2x, the ideal processing-time estimate is 54 minutes.

If its average combined output rate is 6.192 Mbit/s over 6,480 seconds:

6,192,000 × 6,480 / 8 = 5,015,520,000 bytes

That is about 5.016 GB in decimal units before additional packaging effects. An encoder's bitrate target does not guarantee that exact average rate or file size, and conversion speed can vary across the content.

These estimates explain the dependencies without claiming an actual benchmark.

A prepared source still needs a meaning

The representation should retain which work, episode and edition it represents. A smaller encode of the same cut is different from an abridged or edited version.

Its own streams, indexes and timing need analysis. Selecting audio index 2 from the old source can be wrong if the prepared resource has another stream set.

A corresponding resume position also needs the represented timeline. Preserving scene timing is useful when an encode preserves the sequence; changes to introductions, cadence interpretation or editorial content can require another association.

Preparation changes the later trade-off

A compatible prepared resource can reduce per-session conversion work while adding retained storage and preparation time. A representation made for one target can still need another plan on a different player path.