MIT OpenCourseWare


» 進階搜尋
 課程首頁
 教學大綱
 教學時程
 相關閱讀資料
 專題
 Lab Software
 Lab Database

Lab Database

This page contains digitized media files for use in the labs. Most of the files are audio data, but there are also image files and other types of data. For some of the labs, you are asked to record your own voice. If you do not have the tools for making such recordings, example utterances are provided as spoken by both a female speaker and a male speaker. The audio files are given below under the heading of the relevant lab.


Lab Data: 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15
Lab 1: Recording Speech in a Sound-Treated Room, and Spectral Analysis and Waveform Editing by Computer

In Lab 1, you will record a database of speech utterances for use throughout the semester. Below are links to sample databases recorded by two speakers, one female (HH) and one male (KS).

Note that although these files have a .wav suffix, they are in Klatt format, for use with the Klatt speech-analysis tools (xkl); they are NOT Microsoft .wav format. (If there is a desire to use these files with analysis tools that require MS .wav format, we may consider providing both formats in the future.)

You can download tar archives of an entire database here:

female sample database (4MB tar file)
male sample database (4MB tar file)

  1. English vowels
    Female Male
    1 beat bit beat bit
    2 bait bet bait bet
    3 pot but pot but
    4 boat bought boat bought
    5 put boot put boot
    6 bat Bert bat Bert
    7 buy boy buy boy
    8 about about


  2. English consonants
    Female Male
    9 the putt the bug the putt the bug
    10 the tug the duck the tug the duck
    11 the cut the gut the cut the gut
    12 the mug the nut the mug the nut
    13 the fuss the vanilla the fuss the vanilla
    14 the thug the "the" the thug the "the"
    15 the supper the shutter the supper the shutter
    16 the rug the lug the rug the lug
    17 the wonder the young the wonder the young
    18 the chump the jump the chump the jump
    19 the hut sung a song the hut sung a song
    20 the /zah/ the /zhah/ the /zah/ the /zhah/


  3. Sentences
    Female Male
    21 teacher teacher
    22 He wanted to rebel He wanted to rebel
    23 school school
    24 high school teacher high school teacher
    25 He stopped the rebel He stopped the rebel
    26 school teacher school teacher
    27 tall school teacher tall school teacher
    28 Was he a high school teacher? Was he a high school teacher?
    29 The tall school teacher left early. The tall school teacher left early.
    30 teach teach
    31 We live in Cambridge and we study at MIT We live in Cambridge and we study at MIT
    32 He won those shoes. He won those shoes.


  4. Rainbow passage, female
    Rainbow passage, male


  5. Spontaneous speech, female
    Spontaneous speech, male

[back to top]

Lab 2: Spectrographic and Spectrum Analysis of the English Vowel System

You will use the following utterances recorded during Lab 1:

  • The word "shutter"
  • The simple vowel pairs (Items 1-7)

[back to top]


Lab 3: Broadband Spectral Analysis of Sonorant Consonants

You will use at least one of the following subsets of utterances recorded during Lab 1:

  • The words "mug", "nut", and "sung"
  • The words "rug" and "lug"
  • The words "wonder" and "young"

[back to top]

Lab 4: Spectral Analysis of Stop and Fricative Consonants: Sound Generation from Turbulence in the Vocal Tract

You will use at least one of the following subsets of utterances recorded during Lab 1:

  • The words "bug", "duck", "gut", "putt", "tug", "cut", "hut"
  • The words "fuss", "thug", "supper", "shutter", "vanilla", "the", /zschwa/, /zhschwa/
  • The words "chump", "jump", "tug", "duck", "shutter", /zhschwa/

[back to top]

Lab 5: Sound Generation at the Larynx: Characteristics of the Glottal Source

[back to top]

Lab 6: Formant Frequency Calculations from Area Function Data

There are no utterances required for this lab.

[back to top]


Lab 7: Sound Segment Reduction

You will use the spontaneous speech recorded during Lab 1.

If you do not have your own recordings, you can download speech samples that contain clear renditions of several words along with renditions of the same words as produced with reduced segments. The tar file will contain samples from the same male and female speakers (KS and HH) that recorded the speech samples for Lab 1.

Download the Lab 7 examples (xkl format) here. (560KB zip/tar file)

[back to top]

Lab 8: Prosody: Acoustic Evidence for Prosodic Constituent Structure

You will use the following utterances, produced by four speakers and stored in the labc account:

  • Part A: Please say Maine or Duke and Rice will play.
  • Part B: Please say Thomas or Kenneth and Lucas will stay.

Each sentence was produced with two different prosodic forms, and each prosodic form was produced twice by each speaker, for a total of 32 utterances. The utterances are available in two formats, Xwaves (.sd and .f0 files) and Xkl (Klatt .wav files).

Download a tar/zip archive of the utterances required for this lab (12MB tar file).

[back to top]

Lab 9: Intonation Labelling OR Factors Affecting Duration

You will use either the utterances for Part A or the utterances for Part B:

  • Part A:
    1) Utterances 3-10 in files lcorn.16k (Xwaves format), located in the labc account (Download files lcorn.16k and lcorn.f0 here)
    2) Utterances from Lab 8, Part A
  • Part B:
    The following words, recorded by two speakers: "fat", "fad", "fade", "cup", "cub", "coop", "pick", "pig", "peak", "men", "main', "mate", "tuck", "tug", "take". These files are in two formats, Xwaves (.sd files) and Xkl (Klatt .wav). (Download a tar/zip archive of the utterances here (15MB tar file).)

[back to top]

Lab 10: Speech Intelligibility and Confusion Matrices

In this lab, you will participate as a subject in two types of intelligibility tests. If you attend the lab, you will hear the stimuli on audiotape. If you do not attend class, you can download audio files of the tests below. In order to take run the tests on your own, you will need to download the response and answers sheets from the Lab 10 page.

A third type of test is available, although it is not presented in class. You will need to download the response and answers sheets from the Lab 10 page.

[back to top]

Lab 11: Introduction to Speech Disorders

You will use utterances produced by adults and children with several types of speech disorders, all in Xkl (Klatt .wav ) format:

[back to top]

Lab 12: Interpretation of Cineradiographic Motion Pictures of Selected Utterances

In this lab you will use .gif files of cineradiographic images. Because the full database of these images is rather large, it will only be available to students attending the lab. Sample .gif files can be downloaded here:

Download file KNS_6400995.gif, /a/ in "he'tat"
Download file KNS_6401423.gif, /a/ in "soggy"
Download file KNS_6400919.gif, /u/ in "he'tu"
Download file KNS_6401621.gif, /u/ in "shoes"

[back to top]

Lab 13: Speech Synthesis Using a Formant Synthesizer

For this lab, students will record new utterances. If you are unable to record new data, you can use the sample utterances provided for Lab 5, part B (either "pat" or "bad").

[back to top]

Lab 14: Measuring Speech Movements Using Data from X-ray Microbeam System

In this lab, you will use data recorded by an x-ray microbeam system. There are data for two speakers and they are in MATLAB format (.mat files):

Microbeam data for speaker JW16 (four .mat files in 1.7MB zip archive )
Microbeam data for speaker JW18 (four .mat files in 1.8MB zip archive)

[back to top]

Lab 15: Higher-Level Synthesis with a Formant Synthesizer, Using Quasi-Articulatory Parameters

There are no utterances required for this lab.

[back to top]


 
MIT Home
Massachusetts Institute of Technology Terms of Use Privacy