AMBE-1000 Digital Voice Systems, Inc., AMBE-1000 Datasheet - Page 45

• Manufacturer Part Number: AMBE-1000
• Manufacturer: Digital Voice Systems, Inc.

4.2
4.2.1
The AMBE-1000 is not recommended for new designs.
2 bits per Word Format
3 bits per Word Format
4 bits per Word Format
1 bit per Word Format
Please refer to the AMBE-2000/2020 product line.
See Table 3-A
Bits per Word
Unframed Serial Output Format
Unframed Serial Format
The Unframed Format for the channel data is useful for applications which desire
minimal glue logic between the AMBE-1000™ and the channel hardware. The use of
minimal hardware in place of a microcontroller can be realized using this data format.
Another distinct difference in this data format is that framing information (data which
carries the positional information relating to the coded bits) is embedded into the data
stream itself.
coded data itself. A single bit each frame is ‘borrowed’ from the voice data to embed
the framing information. Keep in mind that this ‘borrowed’ bit reduces the effective
voice coding rate quality by 50 bits per second. For example, a system with no FEC
running at 2450 bps in Unframed mode will sound equivalent to one running at 2400
bps in Framed mode.
The designer should also be aware that it takes approximately 15 frames (300
milliseconds) for the decoder to attain synchronization with the incoming stream before
it can output synthesized speech.
shutting down transmission during periods of silence, and then resuming during periods
of speech can not handle this 300 millisecond delay for each synchronization, and thus
should use Framed mode with a more sophisticated framing method.
Unframed Format maintains the ability to input the Control Frames described in
sections 4.1.14.2 through 4.1.14.7.
Unframed format only exists in passive serial mode which inputs and output data in 16
bit words.
mode but only a fraction of the full 16 bits is used to transfer the coded data. The user
selects whether 1, 2, 3 or 4 bits will be transferred in each word based on the selection
in Table 3-A. Important : The voice coding data rate selected must be evenly divisible
by the number of voice data bits per word selected.
The Unframed output format contains 1 to 4 bits within each 16 bits serial output word.
For the formats which contain more than one bit each word the MSB of the data bits is
considered first in the transmission. In Unframed mode, only the coded voice data bits
are output. None of the superfluous information that exists in framed mode is available
in this mode. The number of words that need to be transferred out of the encoder for
each 20 millisecond frame will be the number of bits per frame divided by the number
of bits per word. So a system coding at 4800 bps with 3 bits per word will need to read
32 ( [4800 50] 3 = 32) words each frame.
Table 4-Y Unframed Serial Output Data Format
The AMBE-1000 is not recommended for new designs.
The 16 bit per word format, pictured in Figure 3-L, is maintained in this
Please refer to the AMBE-2000/2020 product line.
Using this data format, the system designer need only transfer the
D
D
D
D
15
msb
msb
msb
msb
Visit us at www.dvsinc.com
Data
14
D
D
D
0
13
D
D
0
0
page 45
12
D
0
0
0
Systems which are attempting to save power by
11
0
0
0
0
10
0
0
0
0
9
0
0
0
0
0
0
0
0
8
0
0
0
0
7
Unused
0
0
0
0
6
5
0
0
0
0
4
0
0
0
0
0
0
0
0
3
0
0
0
0
2
0
0
0
0
1
0
0
0
0
0