FFMTIMIT

Acoustic-Phonetic Continuous Speech Corpus Far Field Microphone Recordings
Training and Test Data and Speech Header Software
NIST Speech Disc 21-1.1 August 1995

This CD-ROM contains the previously-unreleased secondary microphone waveforms for the TIMIT Acoustic-Phonetic Continuous Speech corpus. The primary microphone waveforms, which were recorded using a close-talking noise-cancelling head-mounted Sennheiser microphone (model HMD-414), are available on NIST Speech Disc 1-1.1. The secondary microphone used in the recording of the TIMIT corpus was a Breul & Kjaer 1/2" free-field microphone (model 4165).

While the Sennheiser microphone recordings are relatively "clean" with respect to non-speech noise, the FFMTIMIT recordings includes significant low frequency noise, which was due to the HVAC system and mechanical vibration transmitted through the floor of the double-walled sound booth used in recording. Because it is noiser than its TIMIT counterpart, the data of FFMTIMIT may be used in the development of more noise-robust speech recognition systems. Because the B&K microphone has extremely flat free-field frequency response and because calibration tones are provided, these data may be of value to researchers involved in vocal tract modeling. See the file "tones.doc" located in the "/doc" directory for further information about these calibration tones.

Note that the B&K Timit data contained with this release has not been processed through any highpass filter, (e.g., the 1581-point filter described in the paper "The DARPA Speech Recognition Research Database" by Fisher, Doddington and Goudie-Marshall in "DARPA TIMIT Acoustic-Phonetic Continuous Speech Corpus CD-ROM," NISTIR 4930 / NTIS Order No. PB93-173938.)

This CD-ROM is organized with an identical directory and file structure to the original TIMIT CD-ROM with the exception that the top-level directory is named, "ffmtimit", rather than "timit". All original online TIMIT documentation and time-aligned transcriptions have been included on this disc for your convenience.

Note that the B&K data for 17 TIMIT speakers which were included on the original TIMIT CD-ROM could not be retrieved from the tape archives (See chart below). 

        ===========================================
          DIALECT    MISSING
          REGION    SPEAKERS    SEX    TEST/TRAIN
        ===========================================
             1        DAC0       M        TRAIN
                      JEM0       F        TEST
                      SJS1       M        TEST
        -------------------------------------------
             2        AJW0       F        TRAIN
                      DWD0       M        TRAIN
                      JWB0       F        TEST
        -------------------------------------------
             3        CAL0       M        TRAIN
                      JJB0       M        TRAIN
                      SFV0       M        TRAIN 
                      GJF0       M        TEST
        -------------------------------------------
             4         --        -         ---
        -------------------------------------------
             5        BMH0       F        TRAIN
                      DAS0       M        TRAIN
                      RKM0       M        TRAIN
                      RVG0       M        TRAIN
                      CMB0       M        TEST
                      SFH1       M        TEST
        -------------------------------------------
             6         --        -         ---
        -------------------------------------------
             7        LET0       F        TRAIN
        -------------------------------------------
             8         --        -         ---
        ===========================================

Acknowledgment

Much of the data on this release was stored on 9-track digital magnetic tapes in NIST's archives, and these tapes proved to be largely unreadable on our current system. We would like to acknowledge the assistance of Jane Chang and Victor Zue, of the Spoken Language Systems Group at the MIT Laboratory for Computer Science, in recovering this data.

CD-ROM Contents

The following files and subdirectories are located in the top-level directory of this CD-ROM:

/readme.doc - this file

sphere/ - directory containing the NIST SPeech HEader REsources (SPHERE) software; SPHERE is a set of "C" library routines and programs for manipulating the NIST header structure prepended to the FFMTIMIT waveform files.

ffmtimit/ - directory containing the FFMTIMIT corpus as well as FFMTIMIT related documentation.

Each of these directories contains a "readme.doc" file which may be consulted for more detailed information.

The TIMIT speech corpus CD-ROM and a companion booklet can be purchased from either NTIS or the Linguistic Data Consortium. The CD-ROM was released in 1991 and is identified as NIST Speech Disc 1-1.1 / NTIS Order No. PB91-505065. The companion booklet, "DARPA TIMIT Acoustic-Phonetic Continuous Speech Corpus CD-ROM", was released in 1993 and is identified as NISTIR 4930 / NTIS Order No. PB93-173938.

NOTE

Certain commercial products are identified in various files throughout this CD-ROM in order to adequately specify procedures described. In no case does such identification imply recommendation or endorsement by the National Institute of Standards and Technology, the U.S. Department of Commerce, or the United States Federal Government, nor does it imply that the material identified is necessarily the best for the purpose.