Channel coupling
In digital sound processing , channel coupling enables dependencies between multi-channel sound signals to be exploited in order to obtain a more compact representation (digital audio data compression ). The repeated description of common contents is avoided by only writing information about the difference to an existing channel or to a new center channel in a channel .
The two-channel configuration, i.e. in the field of stereophony , is the most common with the increasing spread of multi-channel systems . Accordingly, the English joint stereo (short: JS) and its varieties such as level difference stereophony , mid / side stereo (English) and parametric stereo are often spoken of . These varieties have their counterparts in traditional sound engineering, for example as intensity stereophony and MS stereophony . They are sometimes associated with sound losses of varying degrees. Many users only know the lossless MS method.
The difference signals can
- stored losslessly ( lossless coupling ) or
- quantized and coded accordingly lossy; the quantization noise falls into the center of the stereo sound image, where it is well masked by the rest of the signal .
Mid / side stereo
Mid / side stereo is a technique that takes advantage of the fact that the left and right channels of a stereo sound recording are mostly very similar; H. the recording has a high middle or mono content. The channels on the left and right are no longer saved, but a middle and a side channel are formed from them.
The method is widely used in audio data compression with newer lossless audio codecs such as FLAC and also with lossy ones such as LAME . It can already be found in the original MP3 codec of the Fraunhofer Institute for Integrated Circuits and also in radio with FM stereo.
Parametric multi-channel sound
In the case of heavily lossy parametric stereo (“intensity stereo”), the signal is analyzed and parametric descriptions are stored, from which the decoder can then, for example, restore an approximate impression by arranging certain signal components accordingly on a virtual sound stage.
Only a mono channel and direction vectors for reconstructing the stereo data are saved.
Level difference stereophony
The level difference stereophony (English "intensity stereo" ) is a simple case of parametric stereo. It only records the level difference between the channels; the delay-related delays of certain signal components are ignored and phase information is lost. The phase shift that may occur is, however, no longer perceptible above a certain frequency . However, the original stereo information is lost; the method is lossy and is generally only suitable for low data rates, where one would prefer to have some stereo information instead of mono coding.
In digital audio data compression, it is usually only used in encoders at very low data rates. A large number of encoders decide on the basis of the selected data rate whether they want to use intensity stereophony , but there are some, mostly older codecs that rely exclusively on this technology.
rating
Especially in the early days of this compression technique, many advocates of “pure” stereo argued that channel coupling destroyed the stereo feeling and thus impaired the listening experience.
Often, however, in such arguments, MS stereo with level difference stereophony (or MS stereo with data reduction of the difference signal) were lumped into one pot, which is partly due to the unfortunate combination of the two technologies under one name.
Meanwhile, the number of strict rejecters of the algorithm seems to be on the decline; this is also due to more clever implementations in modern audio codecs. The technology is practically established, especially in lossless algorithms and at lower data rates.
In the case of a data reduction of the difference channel and the center signal, there are various sonic advantages and disadvantages compared to a separate quantization of the channels. The absolute advantage is the constant directional dominance, which often turns out to be a problem when reducing data from two separate stereo channels, especially at low data rates, because there it is noticeable by the stereo image moving from left and right. However, since joint stereo is encoded in the middle and side signals, this effect occurs here as a pulsation of the stereo width from the middle to the side, which is perceived as much more beneficial in terms of sound psychology or as significantly less disturbing than the stereo middle, the in the media mainstream area contains most of the important signal components, cannot be clearly localized or changes position. The disadvantage is that in the case of a quiet side signal, such as in reverberation rooms, the depth of field on the sides suffers. If the left and right channels differ greatly, there is also a deterioration in MS stereo compared to the separation according to the conventional channels. Since the decision as to whether stereo is coded as left / right or middle / side signal can be made in many audio formats for individual blocks (sections over a few hundred milliseconds), encoders choose the representation for such formats that is easier to encode unless the user has specified a specific display.
literature
- Michael Dickreiter: Handbook of the recording studio technology. 6th edition, KG Saur Verlag KG, Munich, 1997, ISBN 3-598-11320-X .
- Thomas Görne: Sound engineering. 1st edition, Carl Hanser Verlag, Leipzig, 2006, ISBN 3-446-40198-9 .
- Hubert Henle: The recording studio manual. 5th edition, GC Carstensen Verlag, Munich, 2001, ISBN 3-910098-19-3 .