DAT Recording and Saving to Files

1.Recording with the DAT Recorder

  1. POWER: Plug in the power adapter and set the DC IN swith to OPERATE.
  2. MIC INPUT: For mono recordings, plug the microphone into the left channel.
  3. MIC POSITIONING: If you are using the head-mounted microphone headset, adjust the microphone so that it is placed at the corner of your mouth. This is done to avoid clipping of high amplitude sounds (e.g. stop bursts like /p/). You may want to ask someone else to help you adjust the microphone, as it may be difficult to position it while you're wearing the headset. In order to maintain the same amplitude ratio across all utterances in the recording, the microphone should not be moved at all during recording. So, if you have to stop recording briefly to go do something else (like go to the bathroom), but you're not yet finished recording, just unplug the headset and then get up and walk around  -- leave the headset as is on your head until you have finished all recordings. Also, take care not to bump/move the microphone when taking a drink of water.
  4. RECORDING LEVEL: Adjust the recording volume on the left channel of the DAT Recorder (the right channel level should be set to zero).  Check the meter on the Recorder's display -- the amplitude should be at about the 1/2-way mark, or slightly greater. To prevent clipping (of high amplitude sounds), make sure that the reading is not too high on the meter. Test this by saying some of your experiment sentences into the microphone at the same loudness you intend to record at. You can also test by saying a sustained /aaaahhh/ -- if the level looks OK, you should be fine, as this is the loudest vowel.  Also, set the Limiter switch to ON to prevent clipping (if the meter does for some reason end up reaching max amplitude).
  5. FREQUENCY: Choose either 44.1KHz or 48KHz sampling frequency on the top of the deck.
  6. RECORD: Press "record" then "play" at the same time. The red light will come on. When you press the record button to start recording, note that there is a significant delay before the recorder actually begins the recording. Take this lag into consideration or you will end up with the initial portion of your utterances chopped off! (When going from pause to record, the recording starts immedately -- the delay only occurs when going from stop to record).
  7. COUNTER: Press the counter mode button a few times to get the counter mode you want.

2. Saving the Recordings to Files

  1. CONNECTION: It's best to use a mono cable to connect the LINE OUT on the DAT to the LINE IN on the back of the Sun Sparc station (i.e. the circle with the arrow pointing in). If you do not have a mono cable, a stereo cable is OK, but only plug in the left side (white) of the cable into the LINE OUT on the DAT Recorder (plugging the right channel in will create noise).
  2. AUDIO INTERFACE: At the command prompt run the Audio Control Panel:
    gaintool &
    Here you can adjust the recording volume and select the input type.
  3. RECORD: To record the files, run sl6record at the command prompt. Example:
    4. s16record -f 48000 -s 1800 filename.d
    The options are: -f for the sampling frequency, -s is for the duration of the record session (in seconds -- you want to make this quite big, then just kill the record session with Ctrl-C when you are done), and filename.d is the name of the output file. Save audio files in the /man-ray/local/ directory, or make a subdirectory there for your files.
  4. PLAY FILE: If you want to listen to the file from the prompt, type: s16play filename.d at the command prompt.
  5. VIEW and EDIT FILE: To view the waveform and chop up the recording into individual subfiles, run xwaves. Example:
    6. xwaves filename.d
    You want to check to make sure that the amplitude of the waveform is OK. If the grid to the left in the display ranges within about + to - 20K, you're fine. Anything much above this is probably too loud and runs the risk of being clipped at some point. Anything much less than this, say +/-100, is much too soft. If necessary, adjust the record volume in the Audio Control Panel and record in again using

3. Using xwaves

  1. MOUSE BUTTONS: left= set mark, click and drag to "highlight". middle=play within marks. right=pull-down menu and choose option in menu.
  2. PLAY REGION: With the left mouse button, "highlight" the chunk of the waveform that you want to play. Click on the highlighted region with the middle mouse button to play.
  3. SAVING REGION: With the left mouse button, "highlight" the chunk of the waveform that you want to save. In the OUTPUT FILE blank in the top main menu, erase the "foo" filename and enter the name for the file you want to save (without the .d extension -- xwaves will add the .d extension automatically). Be sure to press return after entering the filename to make sure waves has acknowledged the name. Move cursor into the selected region and click the right mouse button to open the pull-down menu. Select "Save segment in file" from the menu. If your filename contains numbers, you will notice that the rightmost number will be automatically incremented when you save.

4. Converting .d (Unix) files to .wav files for use on PC's

  1. REMOVE "WAVES" (i.e. ESPS format) HEADER: Behead the file and redirect the output to a .raw file:
    2. bhd file.d > file.raw
  2. ADD "WAV" HEADER: Use sox, the "sound exchange" software, to put on a .wav header:
    3. sox -r 48000 -t raw -w -s file.raw file.wav

    The options are:

    -r is the frequency of the original file,

    -t is the typeof the original file, file.raw is the name of the original file, and file.wav is the name of the .wav output file.

    -w and -s are necessary options concerning the data formatting.

    Apparently, there is no need to byte-swap (the -x option) when going from UNIX to PC (though byte-swapping is crucial when going from PC to UNIX). If you want to down-sample the sampling frequency of the file, you can do that in this step too.


--Jennifer Venditti & Paul Tepper, 7/10/2001