Audition

This repo contains the scripts and data related to the Audition dataset.

The dataset comprises of six JSON Lines files. One for visual objects data/VO.jsonl and five for textual annotation coming from different sources:

• Original audio transcriptions: data/OA.jsonl
• Audio descriptions: data/AD.jsonl
• Subtitles: data/SUB.jsonl
• Text overlays: data/TXT.jsonl
• Closed captions: data/CC.jsonl

For information and download of the short-films, refer to this table:

Film	Source	Audio Descriptor
A Árvore do Dinheiro	http://cinematecapernambucana.com.br/filme/?id=3264	Marcos Buccini
As aventuras de Pety_animação	https://www.youtube.com/watch?v=h0TbaPIDkFI	Filmes que voam
Ave Maria ou Mãe dos Sertanejos	http://cinematecapernambucana.com.br/filme/?id=3266	Projeto Alumiar Liliane Tavares.
Cinema Gloria	http://cinematecapernambucana.com.br/filme/?id=3267	Projeto Alumiar Liliane Tavares.
Glênio	https://vimeo.com/722339213/d76443858e	Marilaine Castro da Costa
O cuidado vem da terra	https://youtu.be/BsQnR4CkJcI?si=fwVxLuLTVwG1rTig	Luis dos Santos Miguel

Visual annotations

The visual annotations file consists of a list of objects similar to this:

{
  "episode":"As aventuras de Pety",
  "genre": "animation",
  "objectId":9524,
  "objectTimespan":[402.16,407.24],
  "frame":"Natural_phenomena",
  "frameElement":"Natural_phenomenon",
  "boundingBoxes":[
    [402.16,244.0,1.0,351.0,472.0],
    [402.2,244.0,1.0,346.0,473.0],
    [402.24,244.0,1.0,346.0,473.0],
    [402.28,244.0,1.0,346.0,473.0],
    [402.32,244.0,0.0,341.0,477.0],[
    ...
  ]
}

Transcription Annotations

The textual annotation follow the format:

{
  "episode": "As aventuras de Pety",
  "genre": "animation",
  "sentenceId": 216566,
  "sentenceTimespan": [686.6, 688.919],
  "sentence": "Enfim, vamos embora.",
  "tokens": ["Enfim", ",", "vamos", "embora", "."],
  "frames": [
    {"span": [0, 0], "id": "Time_vector", "children": [{"span": [0, 0], "label": "Distance"}, {"span": [2, 3], "id": "Event"}]},
    {"span": [2, 3], "id": "Departing", "children": []}
  ]
}

The episode and genre fields identify the the short film where the visual object or text comes from and the films genre. The objectTimespan and sentenceTimeSpan fields are a tuple that representing the start and end miliseconds of the video where that object or text appears/is spoken. frame and frameElement are the actual FrameNet entities that the visual object represents. The frames field in text annotation represents a list of all frames evoked by that sentence and their frame elements. Their labels are identified by the id field and the span field informs the tokens that evoked the frame or are the frame elements. Finally, boundingBoxes is an array of variable size (with at least one element). Each element is a 5-tuple representing a fixed time point where that bounding box appears and the four other numbers to represent the box itself. It's a tuple of (milisecond, x, y, width, heigh).

License

This dataset is shared under a CC BY-NC 4.0 DEED license. Requests for commercial use should be directed to [email protected].