bvi_rec: Batch Voice recording utility

Purpose

Record an incoming analog audio signal into a flat (continuous, unstructured) AIX file.

Description

The audio adapter converts the analog audio signal into digital sampled data that can be manipulated and stored within the pSeries computer.

The sampling rate used in the conversion process is fixed at 8 kHz, corresponding to that used for digital telephony.

Recording is done at a 16-bit linear resolution, with samples being stored to disk in big-endian format (that is, most significant byte first).

bvi_rec stores the input audio data in a flat (continuous, unstructured) AIX file.

Note: Although there is no theoretical limit on the number of segments that can be processed in one batch, we recommend an upper limit of 5 minutes and 200 segments.

This diagram shows an analog audio signal being input to a audio adapter from a microphone or DAT tape and being converted into digital sampled data by the bvi_rec utility before being saved to file.

Control parameters

The bvi_rec utility uses the following control parameters from the bvi.control file:

UMS_INPUT_SELECT_ LINE_IN (supplied with LINE_1)
Specifies which audio adapter input is to be used for recording. Possible values are:
LINE_1
Use line input socket 1
LINE_2
Use line input socket 2
LINES_1AND2
Use line input sockets 1 and 2
HIGH_GAIN_MIKE
Use the microphone input with a high sensitivity.
LOW_GAIN_MIKE
Use the microphone input with a low sensitivity.
VOICE_FILE_NAME (supplied value: Specifies the name of the voice file created by bvi_rec and used by the other utilities.)
bvi.voice You can specify any valid AIX file name. If you do not specify a path, the voice file is created in the BVI custom server directory ($CUR_DIR/ca/BVI_dir).

The file format is 8 kHz, 16-bit linear big-endian (that is, the most significant byte for each sample is written before the least significant byte).

After opening the BVI Custom Server Import window, (see Starting the BVI custom server), the current working directory should be $CUR_DIR/ca/BVI_dir.

  1. Check free space: Check that you have enough free space by typing the following command on the command line and pressing Enter:
    df .
  2. Check that you have enough free space for both the source audio signal you are going to record and the voice segments that you are going to store in the Blueworx Voice Response voice segment database. You require 16 000 bytes per second for the source audio signal. See Checking disk space for more information.
  3. Prepare to record: Start recording by typing bvi_rec on the command line and pressing Enter.

    The utility reads the control parameters from the bvi.control file and then waits.

  4. Start source: Start playing your input on the audio equipment.

    The utility starts recording. Immediately after starting, bvi_rec pauses before starting to record to disk. In this state, the audio adapter is in monitor mode (that is, the selected input is fed back to the audio adapter line and headphone outputs). This allows you to check that all is working correctly before starting the recording proper.

  5. Start recording: When you are satisfied, press Enter to start recording.

    The audio adapter remains in monitor mode but bvi_rec starts to record the digitized audio data to disk.

    While recording, bvi_rec displays three pieces of information each second:

    • An energy value. This is useful as an indication of whether bvi_rec is correctly receiving voice data from the external source.
    • A value indicating the minimum signal level (in dBm) over the last second.
    • A value indicating the maximum signal level (in dBm) over the last second.

    The dBm levels are useful in determining how to set the silence threshold level for a following segmentation (bvi_seg) operation.

    bvi_rec continually monitors the level of the input signal and displays a warning message if the signal amplitude is too great. Note that too high a signal input to the audio adapter causes clipping distortion which, in turn, reduces the quality of your voice segments. If you see the signal level warning message, you should attempt to reduce the level of the input audio signal. See The Batch Voice Import control file (bvi.control) for a discussion of input signal levels.

  6. Stop recording: To stop recording, press Enter.

    The audio adapter leaves monitor mode, the file is closed, and bvi_rec terminates.

  7. Show levels and amount recorded: To display how many bytes have been written to the voice file, together with the minimum and maximum dBm level values seen over the entire recording operation, press Ctrl C.

You can now use bvi_seg to find the start and end position of each voice segment.