DOLBY DECODERS
A simple passive decoder, in effect the inverse of the block diagram of Fig. 21.2, can be used to give a surround effect. It is based on a simple L-R difference amplifier. The centre signal is reproduced in ‘phantom’ form between the L and R main speakers. The surround signal is also present in the L and R speakers, but out of phase between them, and thus diffused. Although this form of passive decoder is capable of good signal separation (greater than 40 dB between S and C, and L and R channels) it is not used; decoding is carried out in an active system called Pro-Logic, for which purpose- designed IC packages are available. Before we examine this it is worth looking into the purpose and effect of the blocks in the S-signal processing chain of the encoder in Fig. 21.2, and in the decoder used in the receiving or playback equipment.
While crosstalk between front and rear channels is tolerable to some degree, its effects are minimised as far as possible: in the electrical system, in the acoustic link, and in the listener’s ear and brain. The 7 kHz filter has two main benefits: in the presence of azimuth error between the two main channels, crosstalk increases with signal frequency, and reducing h.f. content in the S channel mitigates its effect. Secondly (and less important in the home than in the cinema) the suppression of high frequencies and thus transients in the rear loudspeakers has the effect of making them sound more distant and their sound more diffused, important to viewers who are not seated near front centre of the viewing area. The noise-reduction system, apart from its usual function of reducing hiss and noise, helps to reduce front channel signal leakage, though the amount of NR processing is limited by the need to preserve the nature of the L and R front-channel signals.
The time-delay inserted in the S-channel of the decoder is not there to provide any kind of echo or spatial effect, at least within the terms of reference of the decoding process, though in commercial equip- ment designs it is used to help simulate the acoustic effects of different types of building, and the ‘Hall’, ‘Theatre’, ‘Club’ and other settings provided are mainly dependent on changing attack/decay times and the delay period. The function of the delay in its Dolby decoder role is to sharpen and focus the front-channel sounds, and prevent them from being ‘fuzzed’ by rear-channel sound. The delay period is adjusted according to the distance between front and rear speakers in order to compensate for the propagation time of sound through the air, about 3 ms per metre; the Haas or precedence effect then ensures that the front-channel sound hits the listener before that from the rear/effects speakers. An adjustable delay range of 15–30 ms caters for all normal domestic situations.
Pro-Logic decoder
Fig. 21.3 represents the heart of a Dolby Pro-Logic decoder, an adap- tive matrix. The Lt and Rt signals pass straight through for the L and R channels, and initially into a pair of bandpass filters to derive control signals for surround processing. The filters strip out low- frequency signal components, which convey no directional indications; and high-frequency components, whose phase and amplitude characteristics cannot be relied upon. Now the signals are sampled, amplitude-wise, in a series of full-wave rectifiers: one for Lt, one for Rt, one for Lt + Rt and one for Rt – Lt. The resulting d.c. voltages are log-converted and then compared to produce difference signals at points A and B. Each of them varies in amplitude and polarity according to the dominance axis of the soundfield from moment to moment. Control signal A, derived from L and R, indicates left/right dominance vector, while signal B conveys the front/rear dominance vector. For small dominance factors a relatively long time-constant is used in the follow- ing filters; when a large dominance factor is detected by the threshold switches, a shorter time constant is invoked. Now the control signals are resolved into positive and negative components in a pair of polar- ity splitters to produce four unipolar ‘steering’ voltages EL, ER, EC, and ES, each representing directional sound dominance. They are applied to a bank of eight VCAs (Voltage-Controlled Attenuators) through which the Lt and Rt signals are passing. The resulting eight outputs, together with the ‘pure’ Lt and Rt signals, now enter a combining network in which portions of the ten available signals are added or subtracted according to a characteristic weighting factor to produce left, right, centre and surround output signals in correct proportions, and always with a total acoustic power corresponding to the original ‘studio’ conditions.
Fig. 21.4 shows the other main components of the Pro-Logic
decoder. The noise sequencer is a local source of test signals for setting-up purposes, switched off during normal listening. The L, R and C signals from the adaptive matrix pass through volume-, balance- and trim-control stages on the way to their separate power amplifiers and loudspeakers. The surround signal undergoes conditioning before it is applied to these: first comes an anti-alias filter to prevent sampling errors in the following digital time-delay block, whose storage period is typically 20 ms. The 7 kHz low-pass
filter next downstream has the main function of suppressing spurious switching spikes from the delay section and smoothing the output from its D−A converter. Finally comes the noise-reduction section, working in complement to the NR encoder of Fig. 21.2.
While the 100 Hz–7 kHz limitation on the surround signal fed to the rear speakers may appear restrictive, the facts that all the front speakers have full frequency response, and all the sound channels are allied, means that the overall effect is perfectly acceptable. It would be quite otherwise if the rear speakers were carrying a totally differ- ent programme signal – only then would any shortcoming become discernible. The same remarks apply to the separation/crosstalk characteristics of a Dolby Surround system, whose practical (as opposed to theoretical) signal-separation map is given in Fig. 21.5. An internal diagram of a dedicated Dolby Pro-Logic decoder chip is shown in Fig. 21.6. Lt and Rt signals enter the IC at pins 9 and 10, and L, R, C and S leave on pins 37, 36, 34 and 35 respectively. Pins 1–8, 15–21, and 22–28 of the IC in Fig. 21.6 are occupied only by capacitors which provide reservoirs and time-constants for the chip- internal circuits. This IC works in conjunction with a digital delay chip for rear/surround channel processing, filtering and effects.
Alternatively the whole Dolby Pro-Logic decoding process can be carried out in the digital realm, with an A−D converter working on the Lt/Rt signals at the input of the DPL processor chip, and D−A converters at each of the four outputs. This is common in late decoder designs.
Surround amplifiers and processors
Many large-screen TV sets have Dolby Pro-Logic decoders and amplifiers built into them, with output sockets for the rear speakers. It is difficult to provide the drive power and to accommodate suit- able speakers within the confines of a TV cabinet, however, and a
better alternative is a separate, dedicated surround decoder/amplifier: there is a wide range commercially available. They typically incorporate many stereo input sockets to cater for all the possible signal sources (Dolby surround sound can be conveyed by video- tape, disc, terrestrial, satellite and cable transmissions in analogue or digital form); on-screen set-up, adjustment and indication by virtue of a character/graphics generator and Scart-to-TV hook-up; very often a radio tuner; a variety of effects and simulations of different buildings and venues; and so on. Output powers range up to many hundreds of watts.