Media Metadata
Each uploaded Media asset is analyzed automatically and its Metadata
is extracted.
This starts with basic Metadata
like file size, creation date etc. and goes far beyond.
It includes descriptive Metadata
like transcripts
; speakers
diarization and recognition; keywords
and entities
; and a
scene recognition, which is obtained applying various machine learning techniques.
Transcription in multiple languages is supported, you can retrieve a list of available languages
for your tenant and request multiple language transcripts
for an asset via extract metadata.
In particular, we also support the automatic cross-language transcription from any source language to English.
The speakers
in videos and audio assetss are automatically assigned representations, that can be labeled.
Then they are named automatically in the transcripts
and Subtitles if they are recognized
again in subsquently uploaded Media
assets.
The Metadata
is indexed and constitutes the basis for the deep-search functionality accross all
Media
types.
The Metadata
can be retrieved as structured-data for SEO enhancement of embedded Media
assets.
๐๏ธ Retrieve metadata
This endpoint returns all derived `Metadata` for a given `Media` asset.
๐๏ธ Extract metadata
Extracts metadata (with optional presets) from the `Media` asset.
๐๏ธ Delete metadata
Deletes a metadata for a given asset.
๐๏ธ List transcripts
The transcripts endpoint returns a list of transcripts associated with the given `Media` asset.
๐๏ธ Retrieve a transcript
Retrieves a transcript of a `Media` asset by its identifier. An ID is needed since there can be multiple `Transcripts` for a single `Media` asset.
๐๏ธ Update a transcript
Updates a particular `Transcript` of a `Media` asset. Partial updates are supported based on a subset of the `Transcript` paragraphs.
๐๏ธ Delete a transcript
Deletes a `Transcript` of a `Media` asset.
๐๏ธ Update the transcript time alignment
Realign the word timestamps of a `Transcript` and update the `Transcript` with the word timestamps.
๐๏ธ Activate a transcript
Adds the `isActive` attribute to the specified `Transcript` and removes it from all other `Transcripts` of the given `Media` asset.
๐๏ธ Update a speaker in a transcript
Update a speaker in a `Transcript` by assigning a different speaker to the `paragraph`. This informs the system that the recognition is incorrect and that the representation of the speaker should be updated.
๐๏ธ Assign a transcript as subtitle
Derives subtitles from the `Transcript` and assignes them. The ISO-639 `languageModel` code of the `Transcript` will be automatically mapped to the language name.
๐๏ธ Download a transcript
Download a transcript in the specified format. Available formats are `JSON`, `TXT`, `SRT`, `WebVTT` and `DOCX` (*Microsoft Word*). A [template can be created](#tag/Design-Settings/operation/createDocxExportDesignSettings) and applied to the `DOCX` format export.