66
What is Sensus
?
First we will present some important information about our unique
Sensus
technology that will allow you to
maximize the audio quality and minimize the audible artifacts of the codec process:
Overview
Until now, digital signal processing has been a more precise numeric implementation of well
-
known analog methods. Even
relatively recently designed digital audio processors couldn’t veer too far from the comparatively simplistic concepts that
analog dynamics processing had utilized… until now!
Our
new
Sensus
technology takes digital dynamics processing into a completely new frontier. Instead of the two
-
dimensional static processing architecture of the past,
Sensus
enables the audio processor to modify its own architecture
in real time and in response to ever-
changing program conten
t.
Simply stated,
Sensus
has the ability to “sense” what must be done to a signal in order to best tailor it for the following
codec. As program content changes, it “rearranges the algorithms” to accomplish this goal. The uniqueness of the
Sensus
technology makes it highly suitable not only for codec pre
-
conditioning (or provisioning), but also for a range of other
highly specialized signal processing challenges. The following is a discussion of how
Sensus
technology can be applied
to a coded audio envi
ronment.
Codec Provisioning
The codec is now a common denominator in the world of audio and broadcasting. Digital broadcasting (HDTV, HD
-
Radio
R
, DAB, DRM), podcasting, webcasting, cellcasting, and downloadable music files all employ a form of codec
-based
data compression in order to minimize the bandwidth required to transmit data. The necessarily low bitrates utilized by
these mediums presents a tough challenge for any audio processor used prior to a codec.
Traditional dynamics processors were designed to fulfill the requirements of a medium where the functions were generally
static.
T
hat is, they were well suited to the rather simplistic peak control and bandwidth limiting methods that were
required for analog broadcasting, as well as for the signal normalization techniques used in recording and mastering.
Audio codecs on the other hand are moving targets
-
each codec algorithm has its own set of artifacts. So not only does the
sonic quality vary depending on the algorithm and bitrate used, but more importantly they vary in their ability to mask their
own coding action. This is why we call it a ‘moving target’, and is why conventional audio processors fall short in a coded
audio environment and can actually make coding artifacts worse due to their inabil
ity to adapt appropriately to the
changing operation of the codec as the program content changes.
Prior art in audio dynamics processing could only address
some
of the challenges of provisioning audio for coding. This
hurdle existed because the codec adapts to the incoming program (so as to generate the least amount of output data
representing the input audio) causing the sonic artifacts generated by the process to continually change. Unless the audio
processor can predict these changing characteristics of
the codec, it can’t possibly create output audio that is perfectly
tailored for the coding process.
Conventional processors utilize rather simplistic high frequency limiters and fixed low pass filtering that does not change
with the program material. When these less intelligent processors feed a codec the audio might sound acceptable one
moment and offensive the next. Because they cannot “know” what the codec will do next, the result is over
-
compensated,
dull and lifeless audio… audio that
still
contains objectionable codec
-
generated artifacts!
HD Radio
The advent of HD Radio
R
has introduced the capability to transmit multiple program streams, or “Multicast”, within a
single 96kbps digital broadcast data channel. To facilitate this, multicast relies on the use of codecs with comparatively low
bitrates. A broadcaster can choose to transmit a number of multicast channels and select the bitrate for each one. However,
the more multicast channels there are, the lower the bitrate each channel must have in order for them to all fit within the
total available bandwidth.
