This Dolby system is one of the first surround systems to have been made and deserves to be analysed in detail for the sheer genius with which it was conceived. The system entails the encoding of 4 independent signals: Left (L), Central (C), Right (R) and Surround (S) onto just two signals (Lt and Rt, where t stands for Total). This way the encoded signal can be stored onto an appropriate stereo support (CD, Tape, DAT) if we are in the recording phase, or it can also be reproduced as a stereo television audio channel signal. When the two signals Lt and Rt are picked up (either from the media or from a receiver), they are sent to a decoder that recovers, virtually at least, the four initial signals. Let's now take a detailed look at the two processes: encoding and decoding.
As we have said there are 4 signals to encode: Left (L), Central (C), Right (R) and Surround (S). Let's take a look at the encoding process with reference to the following diagram:
We can see how signal C is at first attenuated by 3 dB and then sent to two channels Lt and Rt. This is done because by sending a copy of C to each channel we double it. To bring it back to its original volume, we have to lower it by 3 dB (remember how to add and subtract dB's? [Combination of sound sources ] ). As far as signal S is concerned, first of all it is lowered by 3 dB, for the same reason as previously, and then filtered with a pass-band filter [Filters ] which gets rid of the frequencies outside the interval [100 Hz- 7 KHz]. The signal then goes through a stage that applies a noise-reduction algorithm [Noise reduction systems ] and finally it is divided into two: one copy, anticipated by 90o, is sent to one of the channels, whereas a copy that is delayed by 90o is sent to the other one.
At this stage the four original signals have been encoded into two new signals and are ready to be sent on air along with a video signal, or recorded onto a stereo support. In order to refine sound S the following solutions are adopted:
Pass-band filter: this makes sure that the final S signal reproduces a "degraded" sound in order to not exceedingly distract the spectator and to act as an environment sound.
When the overall signal is received, the sound part is extracted and sent to the decoding stage whose scheme is illustrated in the following diagram:
The two signals Lt and Rt are forwarded directly to the frontal L and R channels. Since the original L and R had been sent directly to Lt and Rt, they are reproduced exactly as they have been encoded. Signal C has been added to both Lt and Rt and therefore loudspeakers L and R will reproduce its ghost image in the centre. It is a virtual reconstruction but nonetheless it is effective. Signal S on the other hand, is present on Lt and Rt but since it is out of phase on the two channels, it doesn't make us feel like it is coming from L and R but is perceived as an environment sound. This feeling is enhanced by the fact that, when extracted from Lt and Rt, it will be sent to the posterior loudspeakers. Let's see how: signals Lt and Rt are sent to a stage that returns the difference between the two. Thus calculating Lt-Rt, the two signals C, originally added in phase, are cancelled. Vice versa, signal S has been added out of phase to the two channels, therefore the substraction of signals Lt and Rt results in the amplification of signal S by 3 dB, and it is for this reason that during the encoding phase it was made to pass through a 3dB attenuator.
Let's summarize: we started off with 4 independent signals and we encoded them onto just 2 signals. In the decoding phase we tried to recover the 4 original signals and, even though we didn't manage to retain complete independence, we nonetheless managed to rebuild a configuration similar to the initial one. To refine sound S, the following actions take place:
Delay: by applying a delay of about 10ms and therefore within the Haas zone [Haas effect ] , signal S is used as a reinforcement and is thus doesn't distract the audience's attention.
Low-pass filter: it is used to attenuate the phase differences between the signals in front and behind the spectator (let's not forget that phase differences are mostly audible at high frequencies). In any case, high frequencies in signal S are unnecessary because, as we have said, it is an environment signal.
Decoding of the noise-reduction algorithm applied during the encoding phase.
An important characteristic of this system is that it is compatible with both mono and stereo systems. Indeed on a stereo television we'll have all of the signals that were present in the original encoding on its loudspeakers L and R. The surround signal S is present, and even if it won't come from behind the spectator, the fact that it is out of phase still guarantees "environment behaviour" characteristics. By adding the received L and R signals we'd end up in the mono television domain. In this case signal C returns to its original volume, having been attenuated by 3 dB and then added to itself. Signal S disappears, because it has been added to Lt and Rt out of phase. So, in mono the only information we lose for good is the information pertaining to surround. No big deal really... if you think about it a mono television with surround "sounds" rather ridiculous!