Ericsson AVP 1000 1RU, Reference Manual, Page: 449 / 508

Audio Coding
EN/LZT 790 0023/1 R16A
C-5
AAC has specifically been designed to leave behind the baggage of Layers I and II
so that it can be more efficient. The time it takes to encode audio differs between the
profiles, MP being the most intensive and SSR the least.
C.1.7 HE AAC
This standard builds on the AAC. Some of the existing tools have been improved
and new ones added. All AAC objects (including the error robust modes) contain the
Perceptual Noise Substitution (PNS) tool. PNS detects noise in the signal and in the
final bit-stream a noise flag is raised and the power level of the noise stored. The
decoder generates a random noise of the level indicated, and inserts this into the
audio signal.
HE-AAC Formats
HE-AAC audio may be stored and streamed in many formats, which are described
in the standard. These include:
•
ADTS - This format places a header at the start of every audio frame, and
provides enough information for each frame to be independently decoded. It
contains a fixed and a variable header.
•
LATM/LOAS - As for ADTS this also places a header at the front of each audio
frame in such a way that each audio frame can be independently decoded. This
is the required format for compliant DVB streams (ETSI TS 101-154 v1.7.7) and
is available on the Encoder.
C.2 Audio Coding Modes
C.2.1 E.2.1 Mono
This mode has a single audio channel that is encoded independently. It is seldom
used in broadcast as most viewing devices now have stereo speakers or
headphones.
C.2.2 E.2.2 Stereo
This treats the incoming signal as separate left and right signals. Some coding
algorithms will look at the left and right signals and, depending on the correlation of
these will combine the signal into a mid and side channel to save on bits.
C.2.3 E.2.3 Joint Stereo
This option is available for MPEG 1 layer 2 only. This applies a technique called
intensity coding. The human ear is not as good at locating higher frequencies as it is
lower ones. The use of this mode may introduce more artefacts than stereo.