dmFS1016Decode(3dm)                                        dmFS1016Decode(3dm)



NAME
     dmFS1016Decode - implements the US Federal Standard 1016 4800/7200/9600
     bits/s CELP (Code Excited Linear Predictive) Voice Coder.


SYNOPSIS
     #include <dmedia/dmedia.h>
     #include <dmedia/dmparams.h>
     #include <dmedia/dmaudioutil.h>

     DMstatus dmFS1016Decode(DMFS1016decoder handle,
                          unsigned char *ibuf, short *obuf, int nsamples)


PARAMETER
     handle    DMFS1016decoder structure, created by
               dmFS1016DecoderCreate(3dm), specifies the signal processing
               parameters.


     ibuf      pointer to input compressed data buffer. The data format is
               unsigned char (8-bit).  This bitstream buffer must consist of a
               multiple of 18 bytes.


     obuf      pointer to output sample data buffer.  The data format is short
               (16-bit).  The samples are assumed to be two's complement.  The
               sampling rate must be 8 kHz.


     nsamples  number of sample in the output buffer to be generated.  The
               value of nsamples passed to the FS1016 decode/decode routines
               must always be a multiple of 240, 160, or 120 for operations of
               4800, 7200, or 9600 bits/s respectively.


RETURNED VALUE
     Returns DMFAILURE or DMSUCCESS.


DESCRIPTION
     dmFS1016Decode(3dm) implements the Federal Standard 1016 decompression.
     The uncompressed data sampling rate is 8000 Hz. The compression bitrate
     and modes for decoding operation can be

     DMFS10164800BITS
     DMFS10167200BITS
     DMFS10169600BITS

     DMFS1016POSTFILTERINGYES
     DMFS1016POSTFILTERINGNO




                                                                        Page 1

dmFS1016Decode(3dm)                                        dmFS1016Decode(3dm)



     and

     DMFS1016HIGHPASSYES
     DMFS1016HIGHPASSNO



     that select a decoder with or without postfiltering and highpass
     filtering.  Cautious application of postfiltering at the decoder's output
     is recommended.  The human ear's masking properties are exploited to
     trade off speech distortion vs. quantizing noise.  In tandem coding
     scenarios, only one stage of postfiltering is recommended and multiple
     stages should be avoided.  The bitrate and decode mode are set by a call
     to dmFS1016DecoderSetParams(3dm). Without calling
     dmFS1016DecoderSetParams(3dm), decoder sets defaults to
     DMFS10164800BITS, DMFS1016POSTFILTERINGYES, and
     DMFS1016HIGHPASSNO.


     see DESCRIPTION in dmFS1016Encode(3dm) for the discussion of
     decompression of a part of compressed bit stream beginning in the middle
     of a compressed data file.


NOTE
     FS1016 algorithm was developed for sampling rate of 8 kHz.


SEE ALSO
     dmFS1016DecoderCreate(3dm), dmFS1016DecoderDestroy(3dm),
     dmFS1016Encode(3dm), dmFS1016DecoderSetParams(3dm), Federal Standard
     1016.























                                                                        Page 2