Speech and Audio Utilities Contents: Basic audio functions: Recording rec grecord audiotool sound monitor mixer Playback play mpg123 ogg123 CD player X multimedia system (XMMS) File conversion audioconvert sox grip Visualisation: Speech filing system (SFS) Praat Snack (not yet available) Extace Lyre Waveform editing: CoolEdit Processing: Matlab Pitch-scaled harmonic filter (PSHF) Synthesis: Festival (unavailable) Recognition: HMM Toolkit (HTK) Segmental Viterbi decoder (SEGVit) Other: AVPlay for audio-visual playback from digital video Actra for computing vocal-tract transfer functions AFsp for operations on audio files file formats other utilities

Basic audio functions

  Recording

Mic. in:  When setting up a microphone for recording, you should check it is plugged into the microphone input on the soundcard (third socket from the end, as opposed to the line in, which is for signals that have already been pre-amplified). Next, use the mixer to check that the recording box is checked under "Mic" and that the gains for "Vol" and "IGain" are both non-zero (it's a good idea to turn them full on).

Line in:  Signals that have been pre-amplified, such as those from the output of a soundcard or the Sony mixing desk, should be fed into the line input on the soundcard (second from the end, after digital output). With the mixer you should check that the recording box is checked under "Line" and that the gains for "Vol" and "IGain" set.

• rec - a linux command-line utility for capturing the soundcard input from a mic or a line into a standard format file,
  e.g.,

  rec -r48000 -sl -c1 file.wav

  to produce a mono wave file of 16-bit integer samples at 48 kHz. See the man pages for further details.
   
• grecord - a simple sound recording program for Gnome. A bug has been reported for mono playback: if you select mono from the menu [Settings > Preferences > Sound], the resultant file plays back at half speed. The problem is avoided by using the default stereo.

 
• audiotool - for playing/recording.

 
• sound monitor

 
• gmix - audio mixer which can be used to set the software gain of your soundcard's inputs and outputs

  Playback

• play - a linux command-line utility for playing a file at the the soundcard output,
  e.g.,

  play file.wav

  See the man pages for further details.
   
• mpg123 - plays MPEG I, level 3 compressed audio files (i.e., MP3's)

 
• ogg123 - plays OGG compressed audio files

 
• CD player - plays audio CDs in the CD-Rom drive

 
• X multimedia system (XMMS) - plays various audio formats, including WAV and MP3

  Conversion

• audioconvert - for converting file formats.

 
• Sox - utility to convert between different audio formats. See also the practical guide to sox.

 
• grip - a Gnome utility for converting CD audio tracks to Mpeg3 format.

Visualisation

• SFS - a speech analysis utility which is currently maintained for Windows. We have installed an older version for Linux users, which takes command-line instructions.
  E.g., to produce a spectrogram from a 48 kHz wave file, type:

  hed -n newfile.sfs
  slink -i1.01 -tWAV -f48000 file.wav newfile.sfs
  Es -i1.01 -g1.01 newfile.sfs &

  See the online help for further details.
   
• Praat - a speech analysis utility

 
• Snack (not yet available)
• Extace
• lyre

Waveform editing

• CoolEdit Pro - a utility for recording, playing and editing audio waveform files, which is currently only installed on the Vampire project's laptop. Speak to Andrew Birt for further details.

Processing

• matlab - a versatile numerical calculation and visualisation utility, available on all Linux/Unix workstations. See its help pages for further details.
   
• PSHF
- see the Columbo project for further details.

Synthesis

• Festival - speech synthesizer, developed at Edinburgh University

Recognition

• HTK - the Hidden Markov Model toolkit, speech recognition software developed at Cambridge University

 
• SEGVit - Segmental Viterbi decoder, developed with the University of Birmingham

Other

• AVPlay - a RAVL utility for playing both the audio and video parts of a digital video (DV) formatted file,
  e.g.,

  AVPlay file.dv

  Type "AVPlay -help" for usage, or see the RAVL pages for further details of Audio classes and Audio IO devices.
   
• Actra for computing vocal-tract transfer functions
• Speech Acquisition system
• Audiotool for playing/recording
• Lyre for visualization
• AFsp some more utility for audio files. The path is /vol/vssp/m2vsoft/speech/AFsp/AFsp-V3R2/.