Dolby S Explained
Dave Lockwood explains how Dolby S noise reduction increases the potential of analogue tape recorders.
Dolby S has been heralded as the saviour of analogue tape recording, but how does it work and why does it sound so good? Dave Lockwood reveals all.
Analogue tape recorders have been developed to a very high level of sophistication, but in order to achieve the kind of noise performance required for professional work, they invariably need to be run with some form of noise reduction system. This is particularly evident with the narrow or low speed tape formats used for home recording equipment. Several different noise reduction types are already in established use, the most popular being the Dolby and dbx systems, but while all reduce the level of tape hiss by a significant margin, all exhibit undesirable side effects to a greater or lesser degree. The more professional systems offer the best sound quality but are, by the same token, too costly to include in home recording equipment. Dolby S, however, is the latest and unarguably most effective consumer system available, from the viewpoint of minimising side-effects. Like its predecessors and rivals, Dolby S is an encode/decode system, which means that it has to be used both during recording and during playback in order to work.
Although Dolby S has been widely spoken of as the 'consumer version' of the professional Dolby SR system, it should perhaps more accurately be viewed as a further development of Dolby C, with an additional LF band, and a couple of the more sophisticated functions of SR thrown in. Although originally developed specifically to improve the performance of domestic cassette machines, the system, at present, has undoubtedly achieved greater prominence in the semi-pro recording-field, via its excellent performance on 1" 24-tracks and ½" 16-tracks. However, the terms of the licensing agreement for Dolby S actually prohibit manufacturers from using the system on equipment deemed to be 'professional'.
The unique feature of Dolby SR is its ability to make optimal use of some of the fundamental characteristics of the human ear, which determine the way we actually hear things, whilst simultaneously embracing the basic Dolby principle of 'minimum processing'; if the ear is not able to hear noise in a particular part of the spectrum, then signal in that area will not be processed. Conforming to the same fundamental principle that 'the least treatment is the best treatment', ensures that Dolby S suffers none of the modulation artifacts or potentially compromised transient performance of simple broadband compander systems.
Although inevitably somewhat less sophisticated than the professional SR system, Dolby S still consists of five active elements. These can be divided initially into high and low frequency stages, with a single fixed LF band operating in the less critical area below 200Hz, whilst the other four stages all operate from 400Hz upwards. Both fixed and sliding bands are utilised, facilitating optimal matching of the system characteristic to the exact spectral content of the incoming signal. The two separate HF bands are then further subdivided into high and low level stages, with two 12dB compander stages operating with staggered thresholds, facilitating the relatively high maximum of 24dB NR when both stages are fully activated, without the need for the unfavourably high compression ratio that would otherwise be required.
The familiar and widely used professional Dolby A system relied on dividing the signal into four discrete bands, employing compression and expansion in each one, to achieve between 10 and 15dB of noise reduction (dependent on frequency). One of the most significant improvements achieved by SR lay in its ability to conform much more closely to the actual signal spectrum than is possible with this type of fixed-band system. Dolby C and even B type systems both adapt their processing to the HF content of the incoming signal to some extent, so that noise reduction is only carried out on any area of the signal where the HF content is insufficient to mask tape noise. Dolby S takes this principle further, with the two sliding-band HF filters (separate High and Low level circuits with a threshold 25dB below Dolby level), but in addition to these sliding bands, fixed bands are also employed, combining with the sliding bands via a technique termed 'Action Substitution'.
It is an inevitable characteristic of a sliding-band system that the noise reduction effect is lessened at lower frequencies as the band slides towards the HF region; action at HF will, of course, be maintained. It is similarly inevitable that a fixed-band system will display loss of NR at frequencies above the frequency of a loud signal within its band, whilst showing significantly less loss of NR below. By employing both fixed and sliding bands, the boosting of low-level signals, and hence the NR effect, can be far more constant. Effectively, Action Substitution is able to take advantage of the beneficial characteristics of both band types, without suffering the disadvantages of either.
Like Dolby SR, a key factor in the subjective performance of Dolby S processing is not just that the noise floor is very low, but that it remains apparently unchanged whether there is signal present or not. Noise reduction in the absence of any signal is easy (a noise gate would suffice) but for there to be no perceptible change in the noise level when signal is present is a significantly harder task, requiring a rather more elaborate approach. Dolby S, like SR, has the apparent ability to discriminate between wanted signal and unwanted noise, thereby rendering recording artifacts such as modulation noise effectively inaudible, and going on to achieve considerable reductions in distortion and improvements in transient response.
The well known non-linearity of analogue tape at the frequency extremes, which progressively worsens with increasing level, is counteracted by applying a specific 'Anti-Saturation' characteristic. This consists of a simple shelving filter network that reduces the recording level of frequencies where the tape cannot handle high levels. A simple complementary shelving boost is then applied on replay, restoring the response. Furthermore, there are the fixed high and low frequency 'Spectral Skewing' filters first seen in the Dolby C system, again with complementary fixed replay tailoring. These are active at all levels, effectively modifying the record EQ characteristic to achieve the optimum compromise between frequency response and noise. There is a secondary function for these, reducing any tendency for response errors inherent in the record/replay system to cause mistracking of the encode/decode process. Unlike a wideband compander system, Dolby S is thus extremely tolerant of any HF response errors in the tape machine, such as HF droop through head wear, etc.
Finally, there is 'Modulation Control' which deals with the way in which the system reacts to very high level signals just outside the NR bands. It is possible, with a sliding-band system, for such a signal to cause the adjacent band to slide so far as to create a gap between its noise reduction and the masking action of the signal itself. Similarly, signals just outside a fixed band will inevitably cause some gain reduction within it, for the band-pass filters used to create the band will have a finite slope. Modulation Control, again a technique borrowed from SR, undertakes to mitigate this effect, and the subjective total absence of modulation noise would seem to indicate that it works extremely well indeed.
There it is, for those who are interested. That is why people got excited about the advent of Dolby S (and indeed SR). But you don't need to know any of the above to appreciate the system; simply go and listen to, and if possible perform a few tests on, a Dolby S-equipped machine, and I think you'll get the point soon enough.