
PSB 7280
Semiconductor Group
175
Data Sheet 1998-07-01
The timing of the JADE firmware is controlled by the video processor, which generates
an interrupt every 10 ms at the SIO line. The JADE then starts generating a number of
frame sync signals at RFS and TFS, depending on the length of the data packet that has
to be exchanged. The RFS and TFS bursts are asynchronously, i.e. the RFS burst starts
about 16 frame syncs before the TFS. After data packet transfer the JADE waits for the
next SIO interrupt.
During startup procedure the uncompressed interface (IOM) must be setup before the
Serial Audio Interface is started, i.e. the FSC and DCL signals must be stable before the
first 10 ms interrupt is generated by the video processor.
Due to small differences in the clock of the video processor and the audio output, the
JADE is able to add two uncompressed audio samples every 10 ms. That means, a skew
of about 2.5% (
f
S
= 8 kHz) or 1.25% (
f
S
= 16 kHz) between the communication board’s
clock and the audio codec’s clock is acceptable to the JADE and should be aurally
imperceptible. In the following this will be called the long term skew.
In addition to the long term skew, the JADE can correct for short term variances using
an internal buffer mechanism. This allows single SIO periods to be 10 ms
±
15%.
The full definition is as follows:
Long term SIO period
T
L
:
T
L
10 ms
=
Short term SIO period
T
S
:
Duration of n consecutive SIO periods:
The basic clock for the definition of [ms] is the frame sync signal of the uncompressed
audio interface.
Note: For maximum audio quality it is recommended to keep the skew between the
IOM-2 and the SIO time base as small as possible, i.e. to adjust
T
L
in the above
definition as close to 10 ms as possible. In an application with the VCP from 8
×
8
(formerly IIT) like in the Siemens/8
×
8 demonstration board design, the SIO
interrupt period is locked to the IOM-2 time base after a call is setup, so no
compensation on the uncompressed audio needs to be done by the JADE any
more. This ensures the maximum possible audio quality.
0.25 ms
±
T
S
T
L
1
×
15%
±
=
T
i
i
1
=
n
∑
n
1
–
T
L
T
S
+
×
=