<%@LANGUAGE="JAVASCRIPT" CODEPAGE="65001"%> NIST Speech Group Website
Information Technology Lab, Information Access Division NIST: National Institute of Standards and Technology

  • Speech Group Home
  • Benchmark Tests
  • Tools
  • Test Beds
  • Publications
  • Links
  • Contacts
  • Source Data

    Audio data

    The audio data is available via the LDC as an ECorpus (audio data order number LDC2004E01, transcriptions order number LDC2004E02 ). Please contact the LDC to obtain the data.

    All commercial mics were collected at a 48Khz/24-bit sampling resolution. The data was SPHERE-encoded, down-sampled to 16KHz/16-bit and gain-normalized for distribution.

    Video data

    The video data is extracted from a NIST-internal format and is then encoded using the MPEG-2 standard in NTSC format using the following parameters:

    1. 720x480 resolution (NTSC based/29.97 fps)
    2. Main Profile / Main Level
    3. 4500 bits/seconds
    4. Default quantization tables as defined by ISO/IEC 13818-2.
    5. 4:2:0 chrominance format. Chrominance channels are each subsampled half in horizontal and vertical directions.
    6. Uses a GOP sequence of I, P and B frames of length 15 as follows: IBBPBBPBBPBBPBB IBBPBBPBBPBBPBB ...
    7. Two audio channels are added to help with understanding the data: the left channel is a gain-normalized mix of all the head microphones and the right channel is a gain-normalized mix of all the distant microphones. Both channels are encoded at 256 kbps, 44.1 kHz using the MPEG-1 layer II format.

    Data samples

    This section contains audio and video samples of actual meeting recordings. These samples are encoded for playback using the MPEG-4 format for the video and MPEG-1 Layer 3 (MP3) format for the audio. Note that the format used for research will be different and will be specified on this page once it is finalized.

    Audio Recording Samples

    Video Recording Samples

    You can play these samples with Quicktime 6 on Windows and Mac OS. On Linux, either VLC or MPlayer should play the video.

    camera panning still

    Excerpt showing the different camera views

    news gathering still

    Excerpt of a news gathering meeting


    office design still

    Excerpt of a meeting with an office design expert

    presentation review still

    Excerpt of a staff meeting

     

     

    Page Created: September 19, 2007r
    Last Updated: December 19, 2007

    Speech Group is part of IAD and ITL
    NIST is an agency of the U.S. Department of Commerce
    Privacy Policy | Security Notices|
    Accessibility Statement | Disclaimer | FOIA