Digital Speech Processing

LPC at 2.4kbps

In their basic form, Linear Predictive Coding (LPC) algorithms for speech achieve high compression ratios by developing short-term, steady-state models of the vocal tract and transmitting only the quantized and encoded parameters of these models. The speech production process is modeled by a flat spectrum excitation source representing the glottal movement, which is filtered by an all-pole short-term stationary digital filter representing the shaping due to the response characteristics of the vocal tract. The latest version of the LPC voice coding algorithm officially tested by the DDVPC is LPC-10e version 52. It conforms to the requirements of the DoD Standard for operation at 2.4k bps (FED-STD-1015 dated 28 November 1984). The characteristics of LPC-10e are described below: ^[1]

• Sampling Rate: 8 kHz
• Frame: 22.5ms, 54 bits per frame
• Analyzer: Semi-pitch synchronous
  • Linear Prediction Analysis: 10^th Order
  • Low Pass Filter: 19 tap tranversal
  • Pitch: AMDF with dynamic pitch tracking (50Hz to 400Hz; 60 Values, 20/octave)

  • Voicing: 2 decisions/frame based on low band energy, zero crossing count, spectral shape, and periodicity measures

  • Preemphasis: Single zero low frequency cut +6dB/octave high frequency boost
  • Matrix Load: Covariance
  • Matrix Invert: Truncated Cholesky decomposition
  • Reflection Coefficient (RC) Coding: Log area ratio for RC1 and RC2, linear for others
  • Transmission Error Protection: Hamming codes on selected bits during unvoiced and transition frames
• Synthesizer: Pitch Synchronous
  • Error Correction/Detection: On selected bits during unvoiced and transition frames
  • Parameter Smoothing: Pitch, RMS, RC1 - RC6 during voiced frames, smoothing thrshold varies with error rate
  • Interpolation:
    • Log Area ratio for RC1, RC2
    • Linear for RC3-RC10
    • Log for RMS
    • Linear for pitch period
  • Deemphasis: 200Hz high pass + single pole low frequency boost.

[1] J.P. Campbell Jr., T.E. Tremain, "Voiced/Unvoiced Classification of Speech with Applications to the U.S. Government LPC-10E Algorithm," IEEE International Conference on Acoustics, Speech, and Signal Processing, Tokyo, 1986, pp. 473-476.

Performance Measures - Evaluation results of the 2.4kbps LPC algorithm were compared to 2.4kbps MELP and other DoD Voice Processors in a poster session presented at the IEEE International Conference on Acoustics, Speech, and Signal Processing in Munich Germany, 1997 (ICASSP '97) entitled: A Comparison of the New 2400 BPS MELP Federal Standard with Other Standard Coders

Sound Samples - The accompanying table entries are links to sound samples of the 2.4kbps LPC algorithm. These are 8KHz sampled, 16 bit linear PCM files in WAV format.

Source Code - ARCON Corporation, as part of it's technical support of the DDVPC, has provided an ftp site for the housing and distribution of narrowband digital voice compression algorithms. An implementation of FS1015 LPC is available. Simply point and click on the button below to begin the ftp transfer. The Fortran source code, a C language translation, input and processed speech files, and some utilities are provided. This package is available in one compressed tar file (lpc10-1_0_tar.gz, 549KB) via ftp. A copy of the included "README" file is available to view before download: LPC README file

Download 2.4kbps LPC Source Code (549KB)

Please note the size of the file (549KB) and the filename extension (.gz). Make sure your browser is setup to save *.gz files to disk and be aware of the time it will take to download.

Algorithmic Delay - The 2.4kbps LPC-10e algorithm, implemented with minimal buffering, has a speech throughput delay of 90.0msecs.