AEQ
TALENT
30
NOTE:
in this mode the TALENT behaves in the exact same way as the Phoenix Mobile
unit when it is using a STUN server.
4.3.6. AUTO 4 (audio over internet).
This mode is equivalent to AUTO3 but it will be used the SIP server is not the one provided by
AEQ and there are problems with AUTO3 mode. The configuration parameters are the same as
for AUTO3 (STUN server specification).
4.4. FEC modes and reception buffer configuration.
-
FEC error correction mode.
Error correction is performed by sending redundant
information that allows the receiver to recompose the lost data in case of not-perfect
transmissions.
Forward error correction always generates a higher binary rate, and this in turn can
generate more and more losses in very narrow transmission channels, as well as delays. It
is recommended that the communication is started with no FEC (OFF) and, once
established, experiment with the different available modes in case of problems and check if
the results are better with some of them.
•
LOWEST: generates a 40% higher binary rate and produces a 575 ms additional
delay.
•
LOW: generates a 50% higher binary rate and produces a 375 ms additional delay.
•
MIDDLE: generates a 66% higher binary rate and produces 225 ms additional delay.
•
HIGH: duals the binary rate producing 125 ms additional delay.
NOTE: please activate FEC only when advised by AEQ Technical Support, as improper
usage may cause more trouble than simply not using it, depending on the network
characteristics.
-
Adaptive
/
Fixed:
you can set up the
reception buffer
as adaptive or fixed. In the first case,
its size will vary according to the network transmission quality. In fixed mode, its size will be
steady according to manual configuration. We always recommend starting with a fixed
configuration
-
Adaptive Buffer Max/Fixed buffer length:
this is the maximum size of the reception
buffer. When it is defined as adaptive, TALENT will start to shorten it from this value as the
network´s transmission quality allows. If it is defined as FIXED, this max value will remain,
as the buffer’s size won’t be varied during the connection. This value must be set in
milliseconds. The longer the buffer is, packet misses will be less likely, but base delay will
also be longer, especially if the buffer is set to FIXED mode.
SOME RECOMMENDATIONS:
In order to help you select the best option for each application, we recommend to use a Fixed
buffer, with a low value (around 100ms) in applications where optimal audio quality is the main
concern (mainly when using PCM modes in suitably sized networks). If the received audio
quality is as expected, and the network allows for it, you can continue adjusting the buffer to
lower values in order to minimize delay, until you find that audio is compromised (as the buffer
size reaches the network maximum jitter value). At this point, just increase the buffer a little bit
to have some margin.
In high-quality PCM connections, and if the network allows for it, you can start using highest
quality modes (48 KHz, 24 bits, mono or stereo only if required), and if you can't obtain the
desired quality and/or stability (no noises present) and good delay, you can lower quality
progressively until, for example, 16 bit (CD quality audio).
On the other hand, for applications where lowest possible delay is the main goal, but
transparent audio is not necessary (for example, in voice connections with commentators), it is
better to select the Adaptive Buffer mode, starting from a 1000 ms maximum size, approx. If the
network is not too bad, the unit won't increase the buffer to highest values from the network's
jitter value, and it will try to minimize delay continuously. Please not that if the network has very
variable delay, the adjustments required to increase or decrease the buffer size can produce
noticeable effects in the received audio, so this method is not usually recommended for a start.
