MPEG AUDIO
The very flexible MPEG-2 DTV system provides a wide choice of audio standards and modes, again with a trade-off of bit-rate against quality and the number of audio channels available. The two most useful for domestic transmissions are 2-channel stereo (upon which a matrixed surround system like the Dolby one described in Chapter 21 can be imposed) and MPEG 5.1, in which five separate sound channels (R, L, centre, rear R, rear L) can be sent to provide a home- cinema surround sound field. Normal DTV transmissions carry 2-channel stereo signals.
PEG layer 2 audio uses the MUSICAM (Masking pattern adapted Universal Sub-band Integrated Coding And Multiplexing) system, in which bit-rate reduction is achieved by taking advantage of the characteristics of the human ear. The basic encoder set-up is shown in Fig. 12.9. The multi-filter block converts incoming audio samples from the time domain to the frequency domain, producing quantised output samples for passage to the noise allocation block. Meanwhile the psycho-acoustic model (based on the human ear/ brain response) calculates, from moment to moment, a noise- hearing threshold level which, passed to the noise allocation block, sets the quantisation characteristic and the bit-rate. The MPEG algorithm governs, for each separate frequency band, which parts of the signal are below human audibility at any given moment, and sup- presses their passage to economise on bit-rate.
Finally the bistream formatter groups, formats and encodes the data into audio packets for combination with the video bitstream.