语音数据压缩问题

在STM32上采集了有声音的和无声音的数据，经过压缩发现数据完全一样，没有压缩之前采集的数据不一样，用的是G711的压缩协议，有什么问题吗?程序有点看不懂，可否帮我解释一下，做好每句附上注释，嘿嘿！void            ulaw_compress(lseg, linbuf, logbuf)
  long            lseg;
  short          *linbuf;
  short          *logbuf;
{
  long            n;            /* samples's count */
  short           i;                /* aux.var. */
  short           absno;        /* absolute value of linear (input) sample */
  short           segno;        /* segment (Table 2/G711, column 1) */
  short           low_nibble;        /* low  nibble of log companded sample */
  short           high_nibble;        /* high nibble of log companded sample */

  for (n = 0; n < lseg; n++)
  {
    /* -------------------------------------------------------------------- */
    /* Change from 14 bit left justified to 14 bit right justified */
    /* Compute absolute value; adjust for easy processing */
    /* -------------------------------------------------------------------- */
    absno = linbuf[n] < 0        /* compute 1's complement in case of  */
      ? ((~linbuf[n]) >> 2) + 33/* negative samples */
      : ((linbuf[n]) >> 2) + 33;/* NB: 33 is the difference value */
    /* between the thresholds for */
    /* A-law and u-law. */
    if (absno > (0x1FFF))        /* limitation to "absno" < 8192 */
      absno = (0x1FFF);

    /* Determination of sample's segment */
    i = absno >> 6;
    segno = 1;
    while (i != 0)
    {
      segno++;
      i >>= 1;
    }

    /* Mounting the high-nibble of the log-PCM sample */
    high_nibble = (0x0008) - segno;

    /* Mounting the low-nibble of the log PCM sample */
    low_nibble = (absno >> segno)        /* right shift of mantissa and */
      & (0x000F);                /* masking away leading '1' */
    low_nibble = (0x000F) - low_nibble;

    /* Joining the high-nibble and the low-nibble of the log PCM sample */
    logbuf[n] = (high_nibble << 4) | low_nibble;

    /* Add sign bit */
    if (linbuf[n] >= 0)
      logbuf[n] = logbuf[n] | (0x0080);
  }
}



第二个问题是如何去判断语音的停顿呢？有声音和无声音如何区别，通过程序如何实现，查的一些资料说是通过声音的能量来判断，在数据里又是如何显示的呢？