RFX144V24-S23 and RFX96V24-S23 Modem Designer’s Guide
1070
8-1
8. VOICE CODEC AND AUDIO CODEC MODES WITH ROOM MONITOR
The voice codec compresses voice with near toll quality playback at an average rate of 2.9 kbps (CONF=90h) or at a fixed
rate of 4.7 kbps (CONF=98h). An average rate of 2.9 kbps provides 24 minutes of stored voice messages in 4 Mbits of
memory. Optional error correction coding allows voice message storage in audio-grade random access memories (ARAM) at
an average rate of less than 3.15 kbps or at a fixed rate of 5.0 kbps.
The audio codec compresses audio (e.g., music/voice) using an ADPCM algorithm at 24 kbps (CONF=92h) or 32 kbps
(CONF=94h) for highest fidelity reproduction.
Available DTMF detect, tone detect, and tone transmit functions provide for a complete digital telephone answering machine
(DTAM) implementation. DTMF detect (see Section 3.2.5) and three tone detectors operate continuously. A local line echo
canceller (LEC) is used for improved DTMF and tone detection during voice and audio decoder operations. Dual/single tone
transmission is permitted when the decoder is disabled (see parameters 6,7,8, and 9 in Section 4.2 and control bit RTSP in
Table 3-1)
The input source for the voice encoder, audio encoder, DTMF detect, and tone detect may be either the internal integrated
analog codec (IIA) or the external integrated analog codec (XIA). The output destination for the voice decoder, audio
decoder, and dual/single tone transmit may be either the IIA or both the IIA and the XIA (see control bit CODECS in Table 3-
1).
Room Monitor allows the remote-end user to monitor the local room activity by listening to audio captured by the microphone
connected to either the IIA or XIA input. The voice codec, audio codec, DTMF detect, tone detect, and tone transmit
functions are also available during room monitor operation.
8.1 VOICE ENCODER AND AUDIO ENCODER
Encoder error correction coding must be enabled for voice encoder compression of messages with error correction and
disabled for messages compressed without error correction (see control bit HDLC in Table 3-1).
The encoder is enabled by setting control bit CDEN (1A:4). The encoder output data is transferred from the encoder to the
host one byte at a time. Each byte is read from DBUFF (see Table 3-1) in response to a modem generated interrupt (see
status bit B2A in Table 3-1). The maximum time interval between interrupts is seven sample periods for the voice encoder
and one sample period for the audio encoder. The voice codec time interval is programmable having a default value of three
sample periods (see parameter 66 in Section 4.2). Since this parameter is used for both encoder and decoder, the host must
rewrite the desired interrupt time interval between encoder and decoder operations if this value is different from the default
value of three sample periods.
The encoder output is organized into data blocks. The variable rate voice encoder’s output is organized into variable length
data blocks from 4 to 38 bytes. The fixed rate voice encoder’s output is organized into fixed length data blocks. Without error
correction coding the repeated data block sequence is 35, 36, 36, 36 bytes. With error correction coding the repeated
sequence is 37, 38, 37, 38, 38, 37, 38, 38 bytes. The audio encoder’s output is organized into fixed length blocks of 120
bytes at 32k bps and 90 bytes at 24 kbps.
A data overrun condition will occur if the host does not complete the transfer of the current data block when the next data
block is ready for transfer. When an overrun occurs the encoder sets status bit VOVUN (17:0). The encoder input samples
may be discarded and encoding suspended. Encoding continues after data transfer resumes and the encoder resets
VOVUN.
For an encoder implementation example refer to the flowchart shown in Figure 8-1.
Summary of Contents for RFX144V24-S23
Page 197: ......
