One set of processing instructions (its
config) and two folders, input/ and output/. Every
file in the input folder is processed with the project’s config. Projects.A project’s instructions: where the media comes from (
sourceMedia), the spoken language
(transcription), the target languages (translations), and where the results go and which
files to produce (outputs). The project config.Where a project’s media comes from, its
sourceMedia: the input folder (S3FolderSource), an
RSS feed (RssSource) or a Media RSS feed (MrssSource). Inputs.A video or audio file in a project’s
input/ folder, however it got there: uploaded, or
downloaded from a feed. Files.An entry of
outputs: the output folder (S3FolderOutput) or an output feed (RssOutput,
MrssOutput),
with the items written there for every input. Outputs.A file the project writes for an input, under
output/: a video, an audio file or a subtitle
file. One input has one output per item of every destination. Outputs.An RSS 2.0 document, plain (
<enclosure> per item) or Media RSS (<media:content>): the one a
project reads its media from (RssSource, MrssSource), or the copy of it the project publishes
in one language (RssOutput, MrssOutput), linking the dubbed files, its text translated if
asked. Feeds.What works through a running project’s input folder. It starts with the project and stops
itself when idle. The processing service.
One target language of a project, with the engine that translates into it. It feeds subtitle
outputs, and a dub when it has
ttsOptions. Translation.A translation spoken by a synthetic voice, timed to the original speech, and mixed with what is
kept of the original audio. Dubbing.
A dub in which the original voices are removed and the background (music, effects, ambience) is
kept, as in a studio dub. The default:
originalAudioSplitting: true.A dub over the whole original audio, played at a lower level, as in news and documentaries:
originalAudioSplitting: false.A dub whose voices are made from each speaker’s own voice (Deepdub
voice: "auto", Sarvam),
rather than chosen from a catalogue. Choosing a voice.One audio or subtitle stream inside a video or audio file: the input’s own audio, a dub, or
subtitles. Players let viewers switch between tracks.
A version of the processing software. Each build publishes the schema of the config it accepts,
and a project runs on one build until you change it.
Builds and the schema.
An AWS IAM role for your account, with which the AWS CLI or an SDK reads and writes your
projects’ folders. Upload with the AWS CLI.
The id of one processing run of a file. Quote it to support.