AMBE2000 Datasheet, PDF (33/47 Pages) List of Unclassifed Manufacturers – VOCODER CHIP

AMBE-2000™ Vocoder Chip
User’s Manual Version 3.0
Echo Canceller (EC) :
A 1 in this output field indicates that the echo canceller has been enabled.
Sleep (SL) :
A 1 in the this field indicates the device has been put into sleep mode.
Decoder Output Volume Control : Indicates the current decoder volume.
5.3.8 Formatted Output : Words 12-23 :Channel Data
This is the field that contains the actual coded bits. Output of the data begins with the MSB of the first word in this field and
continues through with the final bit output being the LSB of the final word. If the data rate selected is less than 9600bps then
the unused bits in each frame are zero and populate the end of the field. As is noted in the Channel Interface definitions, these
unused bits must still be clocked out of the AMBE-2000™. The packet must always consist of 24 words.
5.4 Unformatted Serial Format
The Unformatted Format for the channel data is useful for applications which desire minimal glue logic between the AMBE-
2000™ 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. Using this data format, the system designer need
only transfer the 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 Unformatted mode will sound equivalent to one running at 2400 bps in Formatted
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. Systems which are attempting to save
power by 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 Formatted mode with a more sophisticated framing
method.
The 16 bit per word format, pictured in Section 5.2, is maintained in this 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 pins
BAUD_SEL[0:1] and Table 4-A. IMPORTANT : The voice coding data rate selected must be evenly divisible by the number
of voice data bits per word selected.
5.4.1 Unformatted Serial Output Format
The Unformatted output format contains 1 to 4 bits within each 16 bits serial output word. The formats which contain more
than one bit each word the MSB of the data bits is considered first in the transmission. In Unformatted mode, only the coded
voice data bits are output. None of the superfluous information that exists in formatted 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 5-P Unformatted Serial Output Data Format
Bits per Word
See Table 4-A
1 bit per Word Format
2 bits per Word Format
3 bits per Word Format
4 bits per Word Format
Data
Unused
15
14 13 12 11 10
9
8
7
6
5
4
3
2
1
0
D msb 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
D msb D 0 0 0 0 0 0 0 0 0 0 0 0 0 0
D msb D D 0 0 0 0 0 0 0 0 0 0 0 0 0
D msb D D D 0 0 0 0 0 0 0 0 0 0 0 0
page 33
Visit us at www.dvsinc.com
Confidential Proprietary, Subject to Change