|
To generate stereo at the receiver the designers decided to send the difference between the two channels as a double sideband suppressed carrier centred as 38 kHz. The pilot at 19 kHz serves two functions: it tells the receiver that a stereo signal is being sent and by doubling it a carrier can be generated to coherently demodulate the L-R signal at 38 kHz. The third harmonic can also be used to demodulate the Radio Data System (RDS) sub-carrier at 57 kHz.
The digital systems are not very different from the analogue one.
Mid/Side or (M/S) stereo is the sum and difference signals defined as follows :
Mid Stereo = Left + Right
Side Stereo = Left – Right
The use of M/S stereo can come about as a technique to overcome the problem in encoders of a high level low frequency signal causing a huge change in the Masking Threshold and generating unwanted digital artefacts that consume more data capacity.
This effect can be overcome and requires fewer bits using an M/S solution than in the L/R model. M/S and L/R can be used together where the coding strategy is of M/S being selectively applied to various audio sub-bands.
M/S stereo plays a part in Digital Radio Mondiale when low bit rates are used in conjunction with Spectral Band Replication and Parametric Stereo.
JOINT STEREO
The first concept we need to look at is the idea of Joint Stereo or more precisely the Joint Coding of Stereo or Multi Channel signals. The coding of audio relies on reducing the overall bit streams down to levels that can be transmitted in the available bandwidth of the channel. All encoding processes apply two techniques to achieve this :
·Using redundancy within the signal itself that can be compacted into a more efficient format without an irreversible loss of information about the signal.
·Exploiting features of the human auditory system that mask certain portions of the signal. The human ear just doesn’t’ hear certain sound and distortions in the presence of others.
The objective of joint stereo coding is find ways of efficiently encoding a pair of related channels such that the human ears still hear the spatial aspects of the signal when they are decoded at the receiver. The spatial aspects will be both the left to right and the front-to-back differences that our ears would distinguish if one was actually sitting in the auditorium.
Intensity Stereo
Intensity stereo is a type of joint stereo encoding and is used extensively in DAB to reduce the bit rates. The reduction can be of the order of 10 to 35 kbps, which is significant. In the UK the highest DAB bit rate is the 192 kbps used by BBC Radio 3. A handful of stations that you would expect to use the same bit rate have opted instead to use a joint stereo mode at 160 kbps. Many UK DAB stations use J-Stereo at 128 kbps.
Intensity stereo works by reducing what is termed as perceptually irrelevant information between audio channels. This means that if the human ear won’t hear it then there is no point in encoding that particular sound into digital form. Like all of these techniques the audio spectrum is split into a number of sub-bands with 32 being a very common number. Don’t forget that at this stage the audio is in digital form and represented by many thousands of data samples. In the Eureka-147 DAB system the first 4 sub-bands (numbered 0 to 3) are kept in true stereo mode. Depending on the degree of the bit-reduction required then bands 16-31, 12-31, 8-31 or 4-31 are converted to the intensity stereo mode.
Now that the audio samples for both the Left and Right channels are in digital form and described by a number of parameters called spectral co-efficients they can be manipulated mathematically to represent the audio channels by a SUM signal the preserves the amplitude (level) information and a VECTOR signal to position the stereo image. These SUM and VECTOR numbers are generated for each sub-band and are therefore providing a very detailed description of the audio signal.
The Joint-Stereo encoding comes about as samples from the TWO input channels are jointly analysed and reduced down to produce a SINGLE set of samples to represent the signal.
|
