Skip to content

Latest commit

 

History

History
40 lines (25 loc) · 1.58 KB

README.md

File metadata and controls

40 lines (25 loc) · 1.58 KB
Raw

VoiceCoach

This is the TED Talk dataset for VoiceCoach, CHI 2020, https://arxiv.org/abs/2001.07876.

TED_dataset

The datasets contain meta information of 2623 TED Talks in official TED.com website until Jun 7th, 2019.

The meta information includes fields: 'author', 'datefilmed', 'totalviews', 'comments', 'language', 'downloadlink', 'vidlen', 'aws-transcripts', 'datecrawled', 'datepublished', 'title', 'id', 'url', 'keywords', 'videoname', 'ratings', and complete information is stored in the field 'alldata_JSON'.

Fileds

  • url: original video link
  • aws-transcripts: Each video in the dataset is transcribed by AWS. It has two fileds, including:
    • transcript: all words in the video
    • words: an array containing detailed information about all words. e.g.,
      • "start_time": "12.94",
      • "end_time": "13.25",
      • "alternatives": [{"confidence": "0.9097", "content": "we"}], "type": "pronunciation"}]

Video Downloading

tedvideo_download.py contains the code for downloading ted videos from TED.com

Video2mp3/wav

You can use ffmpeg to convert .mp4 to other audio formats (e.g., mp3, wav, etc.)

(updating)

Notice

If you use this dataset, please cite our paper

VoiceCoach: Interactive Evidence-based Training for Voice Modulation Skills in Public Speaking

Preprint: https://arxiv.org/abs/2001.07876

Authors: Xingbo Wang, Haipeng Zeng, Yong Wang, Aoyu Wu, Zhida Sun, Xiaojuan Ma, Huamin Qu

Acknowledgements

The dataset is shared under the Creative Commons license.