|
|
|
|
|
Philips Research - Technologies
|
Parametric coding for high-quality audio
|
|
 |
Audio compression techniques have revolutionized
the distribution and accessibility of music by
making audio content easily and readily available to
anyone with Internet connectivity. The most popular
format: MP3, a coder of the so-called ‘waveform
coding’ type, achieves its high compression ratio by
exploiting both the irrelevancy and redundancy
available in any audio signal. This is realized by
combining perceptual encoding, i.e. the reduction of
information based on the limited audio perception
capabilities of the human hearing system, with data
compression algorithms. In effect it hides the
quantization noise introduced by compression in
areas where the ear is less sensitive to produce a
perceived audio quality that is close to that of the
original .wav files. With this it can realize
compression ratios of around 15:1.
For even higher compression ratios, parametric
coding schemes are gaining in popularity. A
parametric coder not only exploits the limitations
of the human hearing system, but also attempts to
model the incoming audio signal.
 In response to a call for proposals by MPEG, Philips
has developed a new parametric coding scheme that
provides medium audio quality currently for a
compression ratio around 60:1. The scheme relies on
dissecting the incoming audio signal into three
‘objects’ that are a common constituent of all audio
signals: transients, sinusoids and noise The stereo
information is efficiently parameterized in a fourth
object employing the OCS (One Channel Stereo)
algorithm also developed at Research. Each of these
objects is encoded separately, taking account of
their distinctive characteristics.
A new compression standard based on the Philips
parametric coding scheme is currently under
consideration by MPEG and towards the end of 2003 it
is expected to be incorporated in the MPEG-4
extension 2 standard.
For more information:
E-mail:
bert.den.brinker@philips.com
|
|
|