Page 26
SOLIDYNE 462
dsp - broadcast digital processor
This technique, based in a discovery of the Dr
Leonard Kahn, acquires
international validity
with the work of Oscar Bonello, published at the
Journal of AES, Vol.24,5 in which it is described,
for the first time, the theory of its operation.
The peak asymmetry canceller is in essence an
all-pass network, a class of not minimum phase
network. That is: a network whose transference
function has zeros in the right semi plane. This
network has a full flat response to frequency;
only its phase response is function of the
frequency. This phase rotation, which must
compliment a very particular condition, is
responsible of the peak symmetry of the audio
signals. Those signals that by their nature, are
totally symmetrical (like most of the musical
instruments), are not modified by this processor.
This processor, by itself, allows to increase
between 3 and 5 dB the final power broadcast by
your transmitter (it is to say that it increases by a
factor of TWO the average power transmitted).
Numerous tests have been made in different
countries, to verify, in real conditions, these
results.
6.2.4 SUBSONIC
PHILTER
Is a Chebyshev type high-pass filter, with low
ripple. Its cut-off frequency, of 20 Hz, eliminates
the signals of audio below that frequency. These
subsonic components do not contribute anything
to music, since they aren’t listened by the ear.
Nevertheless, they have a pernicious effect that
produces a disagreeable sensation: the saturation
of the amplifiers and the loudspeakers (by
excessive excursion of the cone). The sound that
is obtained with the connected filter is pure and
clean.
6.2.5 LINEAR
EXPANDER
The expansion, previous to the compression
process, is an excellent resource to increase the
signal/noise ratio of the original program. This is
advisable, since the compression process, when
reducing the high level passages, consequently
increases the relative level of the passages of low
level, and therefore the noise. This is a forced
consequence of the compression process that has
particular effect in the increasing of the ambient
noise of the microphones. In order to avoid that,
the Solidyne processors incorporate a linear
expander, previous to the compressor stage.
The expander works within a very wide range of
signals, below a threshold value. That means it
always expands within that range, for any level of
signal. The curve of transference, based on the
input level, is a straight line (from there the "linear"
name). While the signal stays below the threshold;
by each 10 dB that the input level reduces, the
expander will reduce, for example, 3 dB additional
(1,3:1 ratio), that is to say that the output will be
reduced in 13 dB. This happens for any input
value, below the threshold. Then if the input is
reduced in 30 dB, the output will do it in 39 dB;
that is to say that the background noise has been
reduced in 9 dB. This way, the expander
compensates the increase of the background
noise that the compressor, like undesired effect,
will increase.
At this point, maybe you will be thinking that
DOES NOT HAVE SENSE to make an expander
of the signal and soon to compress it. You will
think, perhaps, that an effect cancels to the other.
But it’s not true for two reasons. First: the different
attack and recovery times. Second: multiband
compressors have elevated threshold, whereas
the linear expander has a very low threshold and a
linear behavior below the threshold. It means that
the actions do not cancel, because both
processes are not complementary.
The linear expander, to optimize its behavior, has
instantaneous attack and a fast recovery times.
Here is where the psychoacoustic concept “post-
pulse hearing inhibition” is used. This allows using
an expander with a quick recovery time, so that it’s
not perceived by the ear. The broadband
compressor that follows the expander has a very
slow recovery time. Therefore, with impulsive
signals, as the audio program, any cancellation
effect occurs.
Another advantage of using a linear expander
previous to the processing is that an excellent
audible sensation of dynamic range is obtained. In
fact, recent studies have demonstrated that the
audible sensation produced by the level variations
of an audio signal, is related to the changes
happened in the first 50 milliseconds, and is little
dependent of the reached final value. This implies
that an expander in the short term is perceived
like a great dynamic range, whereas the power
sensation (and even the coverage area of the
radio transmitter) is related to the AVERAGE
ENERGY, which depends of the compression of
the energy level.
You can see that they are two concepts different.
With audio processors of conventional design, the
expander and the compression were antagonistic
concepts. This does not happen in the field of the
psychoacoustics processors.
