Specifications

AMBE-3000™ Vocoder Chip Users Manual
Version 2.8, August, 2011
In
itial Design Considerations
4.4.3 DTMF Dual Tone Multiple Frequency, Detection and Generation
The AMBE-3000™ Vocoder Chip is capable of detecting, transmitting, and synthesizing DTMF tones. When the encoder
detects DTMF tones the voice data field will contain the DTMF tone data. Additionally, the encoder passes the DTMF data in-
band (within the regular voice data bits) so that normal DTMF tones pass seamlessly from the encoder to the decoder for
synthesis. The decoder synthesizes a DTMF tone in response to reception of an in-band DTMF tone frame or reception of a
control packet with the DTMF word set. When this voice data is received by an AMBE-3000™ Vocoder Chip decoder, it will
regenerate the inband tone. The AMBE-3000™ Vocoder Chip can also generate “Dual Tones” at many different frequencies.
Each tone packet generates 20 milliseconds of output tones. The length of the output tones can be extended by repeating the
tone packet. DTMF may be enabled or disabled through a control packet. DTMF is enabled by default.
4.4.4 Soft Decision Error Correction
Significant improvement in FEC performance can be added by setting up a receiver so that the demodulator is making a finer
estimation of the received energy prior to sending it to the decoder, this is called soft-decision decoding. To use Soft Decision
Error Correction use the CHAND4 (ID 0x17) field in the channel packet. The AMBE-3000™ Vocoder Chip utilizes a 4-bit
soft decision decoder. The bits are defined as follows:
Decision Value (Binary) Interpretation
0000 Most confident 0
0111 …
1000 …
1111 Most confident 1
Table 10 Soft Decision Error Correction
The user must implement circuitry at the receive end of the channel for making a finer (4 bit) estimation of the received energy.
The AMBE-3000™ Vocoder Chip uses a different channel data field (CHAND4) to specify channel data represented by 4 soft
decision (SD) bits. The decoder will make the decision of whether or not a 1 or a 0 is represented by the SD bits.
4.4.5 Skew Control (SK_ENABLE TQFP pin 6 BGA pin D3)
The AMBE-3000™ Vocoder Chip processes speech in voice frames that are approximately 20 ms in duration. Skew control
can provide the designer with flexibility in dealing with clock drift. The AMBE-3000™ Vocoder Chip skew control feature
allows the vocoder chip to compensate for drift between the frame and sample rate clocks.
Codec Mode
When skew control is enabled, the AMBE-3000™ Vocoder Chip adjusts the frame boundaries so that they occur on the rising
edge of the TX_RQST signal. The user must generate the TX_RQST signal such that the frame size varies between 156 and
164 samples.
Packet Mode Skew Control Enable
In packet mode the normal length of the input speech packets is 160 samples. However this can vary between 156 and 164
samples in length. Output speech packets can also vary in length from 156 to 164 samples.
4.4.6 Noise Suppressor (NS_ENABLE TQFP pin 7 BGA pin D2)
The integrated Noise suppressor feature of the AMBE-3000™ Vocoder Chip is used to reduce the effect of background noise
in the encoder input signal. The Noise suppressor is applied to both silence frames and voice frames, but not tone frames.
When the noise suppressor is started it may take up to a few seconds to converge allowing for it to begin fully working.
(Subject to Change) Page 30
DVSI CONFIDENTIAL PROPRIETARY