Record

Author:         Jean Laroche
Platforms:      NeXT, ?
Prerequisites:  NS2.2 or higher?
Price:          free
Demo:           N/A
To acquire:     ftp://ftp.eecs.umich.edu/pub/nextmusic/Record.tar.Z (25k)
Information:    ?
Entry updated:  12 November 1996
Description:    A command-line utility that records soundfiles at any sampling rate from an A/D64X.



RECORD(1)           UNIX Programmer's Manual            RECORD(1)

NAME
     record - record sounds at any sampling rate from A/D64X.

SYNOPSIS
     record [ -UDSKtovsosibcr ] soundFile

DESCRIPTION
     record records from the A/D64X interface, simultaneously
     converting to any specified sampling rate.  record can
     either record from the digital input (for direct digital
     recording from a DAT machine or a CD player) or from the
     analog input (using the A/D64X analog-digital converters).
     When analog input is selected, record chooses the A/D64X
     quartz frequency (32, 44.1 or 48kHz) that makes the conver-
     sion to the desired sampling rate as easy as possible,
     unless the A/D64X sampling rate is also specified. When
     recording from digital input, you need to indicate the
     SPDIFF sampling rate (either 32, 44.1 or 48kHz). record can
     record in stereo, or in mono (selecting one of the two chan-
     nels, or mixing them), and write the output sound with or
     without a header. record can also swap the output sound's
     bytes, for compatibility with DEC machines, PCs etc...
     record can be forced to do or not do direct-to-disk record-
     ing.

OPTIONS
     -v   Tells record to shut up. By default, record gives all
          recording informations.

     -S sampling rate
          This specifies the sound's output sampling rate. If
          this sampling rate is not 32kHz, 44.1kHz nor 48kHz,
          record will perform a real-time sampling rate conver-
          sion to the specified sampling rate.

     -L length in seconds
          With this option, you can specify the duration in
          seconds of the signal that will be recorded. record
          then asks you to start the recording by hiting the
          "return" key. If this option is not used, record will
          ask you to start and to stop the recording.

     -a   (default) used to select the analog input (no sampling
          rate specified, record will choose the most appropriate
          one according to the requested output sampling rate).

     -dh  used to select digital input, indicating that the sig-
          nal coming from the digital input is in 48kHz.

     -dm  used to select digital input, indicating that the sig-
          nal coming from the digital input is in 44.1kHz.

     -dl  used to select digital input, indicating that the sig-
          nal coming from the digital input is in 32kHz.

     -b   Soundfile output. With this option, the converted sound
          is written in a soundfile, with a 28 byte header. If -b
          is not specified, the output is written in a raw,
          headerless file (short ints).

     -cs  records the signal in stereo mode. Both channels are
          interleaved, starting with the left channel.

     -cl  record only the signal's left channel.

     -cr  record only the signal's right channel.

     -c+  mixes the signal's two channels, and records in mono.

     -w   This option makes it possible to rescale the recorded
          sound. Use this if the recorded sound clicks. The
          default rescaling value is 0.9. This option is disabled
          when bit faithful recording is performed.

     -s   This can be used to swap the output sound bytes. Use
          this if you want the recorded sound to be in DEC or in
          PC format. This option is valid only with the -b
          option.

     -A   Disables direct-to-disk recording. In default mode, the
          sound is saved on the disk at the same time it's
          recorded (direct-to-disk recording), which limits the
          size of the virtual memory allocated by the record pro-
          cess. However, in some cases, direct-to-disk recording
          fails because the disk or the network cannot keep up
          with the rate at which samples are coming in. This typ-
          ically happens when recording at a high sampling rate
          in stereo mode and storing in a file located on another
          machine. In that case, the -A flag disables direct-to-
          disk recording: the sound is stored in memory, and save
          at the end of the recording. The danger with this is
          that the record process can become extremely fat
          memorywise.

     -al  This is used to specify analog input, with low A/D64X
          sampling rate (32kHz). You don't normally use this
          option.

     -am  This is used to specify analog input, with medium
          A/D64X sampling rate (44.1kHz). You don't normally use
          this option.

     -ah  This is used to specify analog input, with high A/D64X
          sampling rate (48kHz). You don't normally use this
          option.

     -r   forces the output header to contain the exact frequency
          you specified with the -S option. Due to the way record
          works, it sometimes can't convert exactly to the
          requested sampling rate, but only to an approximation.
          Normally, the approximated sampling rate is written in
          the output sound's header. If you use the -r option,
          the exact requested sampling rate is written instead.

     -K filter length
          This can be use to specify the filter length you want
          to use. This length is normally calculated
          automatically, so you shouldn't need to use this
          option.

     -B   This can be use to specify the type of filter windowing
          you want to use. By default, a Hanning window is used,
          giving a rejection of about -40dB to -50dB. It may be
          necessary in some cases to use a Blackman window to get
          more rejection (about -60dB), at the expense of a wider
          transition band. This can be done by use of the -B
          flag.

     -U upward conversion factor
          upward conversion factor. Normally, you don't need to
          use this option.

     -D downward conversion factor
          downward conversion factor. Normally, you don't need to
          use this option.

EXAMPLE
     "record -a -b -S16000 file" records a stereo soundfile at
     16kHz from analog input. The output soundfile has a 28 bit
     header.

     "record -dh -cl file" records a headerless mono file (left
     channel) from digital input (at 48kHz). The transfer is
     bit-faithful (no conversion, digital transfer).

     "record -dh -c+ -S16000 file" records a headerless mono file
     (both channels mixed) from digital input (48kHz) and down
     samples it to 16kHz.

SEE ALSO
     sndinfo(1), srconv, play, fromsnd, tosnd

AUTHOR
     Jean Laroche, June 1992, TELECOM PARIS.