
Preliminary
DSP for Digital Answering phone with FLASH interface
KS16121
The speed of playback is adjustable by the host, from 0.5 to 2 times the normal speed in an increment of 0.25,
without changing the voice characteristics. The host can specify an offset in the playback command. The offset,
given in seconds, instructs the KS16121 to skip a specified length of the message, from the beginning, before
the playback starts.
During the playback operation, the host can issue a pause command to stop the playback momentarily and
resume the operation from the same point.
VOICE PROMPTS
Certain applications require to play a pre-stored speech segments, including fixed messages, voice prompts
and voice guides. The user can be store voice prompts in an external ROM/EPROM or in a part of flash
memory , but not both. Voice prompts stored in either memories is in a compressed form ( 5.15 Kbps ) to
allow for an increased number of phrases and efficient memory usage.
The KS16121 has a memory interface that can read these phrases from each memories , without interfering the
messages stored in flash memory. The memory space for voice prompts is limited to 64Kbytes when we use
external ROM/EPROM as a voice prompts storage, but is not limited when a part of flash memory is used.
The maximum number of voice prompts which KS16121 supports are 128 . The KS16121
S
oftware
S
upport
T
ools
enable users to generate user-specified speech samples using an external ROM/EPROM or a flash memory.
It is often important to control the time delay or gap between phrases properly when a multiple number of
phrases are played in sequence. The KS16121 host interface enables users to control the gap between the
current and next prompts accurately in 20 - msec resolution.
DTMF / CALL PROGRESS TONE DETECTION
During the record, playback and voice prompts modes as well as the tone detection / generation modes, the
KS16121 monitors the in-coming signal for the presence of DTMF and call progress tones. The monitoring
result is returned to the host as a part of the 16-bit status word.
The KS16121 employs a
N
ear-end
E
cho
C
ancellation ( NEC ) algorithm in its tone detection subsystem.
The NEC removes the echo signal, returned by the codec, from the in -coming signal and helps the detector
perform reliably in the presence of a strong local signal.
The result of DTMF detection is expressed as a 4-bit number as follows: