PSB4860HV4.1 Lantiq, PSB4860HV4.1 Datasheet - Page 60

• Manufacturer Part Number: PSB4860HV4.1
• Manufacturer: Lantiq

Specifications of PSB4860HV4.1

• Lead Free Status / Rohs Status: Supplier Unconfirmed

Data Sheet
2.1.13
The speech coder (figure 33) has two input signals I
the coder while the second signal (I
recording. The signal I
Figure 33 Speech Coder - Block Diagram
The data rates 10.3 kbit/s and 5.6 kbit/s are fixed rates. The data rate 3.3 kbit/s is an
average rate that can actually vary between 740 bit/s and 4.8 kbit/s. The rate currently
achieved heavily depends on the energy of the incoming signal. While silence is
encoded with 740 bit/s, high energy bursts require 4.8 kbit/s. This implies that for a voice
prompt consisting of only one word, the compression rate tends to be approximately
4.8 kbit/s. Furthermore if an Automatic Gain Control AGC is used, the AGC may increase
the signal power in such a way that silence is not recognized as silence. Then, the
compression rate also tends to approximate 4.8 kbit/s.
Data is written initially at the beginning of a file and the file pointer is advanced as
needed. In case of any memory error (e.g. memory full) a file error is indicated and the
coder is disabled. The controller must subsequently close the file. The file can be played
back, though.
The coder’s compression rate can be switched on the fly. However, it may take up to
60 ms until the switch is executed. No audio data is lost during switching.
The signal I
compared to a reference level MIN. If the filtered signal exceeds MIN, then the status bit
SD (table 31) is set immediately. If the filtered signal has been smaller than MIN for a
programmable time TIME then the status bit SD is reset.
The coder can be enabled in permanent mode or in voice recognition mode. In
permanent mode (bit VC is set to 0), the coder starts immediately and compresses all
input data continuously. The current state of the status bit SD does not affect the coder.
In voice recognition mode (bit VC is set to 1), the coder is automatically started on the
first transition of the status bit from 0 to 1. Once the coder has started it remains active
until disabled.
I
I
1
2
Speech Coder
2
is first filtered by a low pass LP with programmable time constant and then
MIN
LP
1
can be coded by either a 3.3 kbit/s, 5.6 kbit/s or 10.3 kbit/s coder.
GAP
2
10300 bit/s
) is used as a reference signal for voice controlled
3300 bit/s
5600 bit/s
60
1
VOX
Memory
and I
2
. The first signal (I
SCI
SCI
PSB 4860
2000-01-14
1
) is fed to