Perhaps it's endemic to the society in which we live that nothing, no matter how good, is ever quite good enough. It seems that we have a fixation with wanting everything to be larger and more impressive than life -- witness the extravagant special effects used in contemporary films and the expansion of the cosmetic surgery industry. When it comes to audio, we're also tempted to seek out silicone enhancement to give our mixes more lift and forward projection. Accepting that there's a demand for this, what can you do to a perfectly good mix to make it sound even better?
Perhaps the first real enhancer was the equaliser, a device that works by cutting or boosting a part of the audio spectrum to alter the overall spectral balance. EQ can help brighten sounds or change the relative balance of the bass and mid-range, but it can only work with what's already there -- if the frequencies you want to hear never existed in the first place, no amount of EQ will bring them out. Another limitation is that when you add EQ boost, it's there all the time -- you can't, for example, decide to add bass boost only to the kick-drum beats or top boost only to the snare hits. However, this is possible using dynamic equalisation, where the amount of tonal boost varies according to the dynamics and spectral content of the signal being processed. Properly applied, dynamic equalisation effects are quite dramatic -- they can actually increase the tonal contrast within the music rather than applying a blanket treatment to the overall sound or mix. Dynamic equalisation is used in several enhancement devices, described later in this article.
Several companies, the most famous of which are probably BBE, Aphex and SPL, build enhancers of one kind or another, and some combine elements of dynamic EQ with other processes, including harmonic synthesis and phase manipulation. Not all manufacturers use the same combination of techniques, the outcome of which is that each type of enhancer has its own characteristic sound.
BBE's Sonic Maximizer system is based on the premise that most audio systems introduce unwanted phase shifts into the audio signal, resulting in a lack of clarity. To remedy this, the BBE processor splits the audio signal into three frequency bands, then applies different delays to each band to restore the original phase relationship of the harmonics making up the sound. On the models I've used, frequencies below 150Hz are delayed by around 2.5ms, while those between 150Hz and 1200Hz are delayed by about 0.5ms. Frequencies above 1200Hz are not delayed, but do have some form of dynamic control (compression/expansion) applied to them so as to enhance transients. Unlike the Aphex process described later, the BBE principle does not add new harmonics, but rather attempts to realign the relative phase of existing harmonics, as well as using dynamic processing to emphasise transients.
Essentially, the controls comprise a low-end EQ control to help balance the bass end against the enhanced top end, and a Definition control. LED metering shows the extent of the dynamic processing applied to the HF band and, on the models I've used, there has also been an Auto/Manual button. In Auto mode, the dynamic treatment of the HF band is determined by the dynamics of the mid-band signal; in Manual mode, the high band is subjected to a fixed degree of dynamic processing. The Definition control affects only the dynamic, high-band processing level -- the inter-band delays are always active when the process is engaged.
Setting up a BBE Sonic Maximizer usually entails selecting Auto mode, adjusting the Definition control to bring up the high end, then using the Lo-Contour control to make up for any lost bass end. Overall subjective brightness is increased, but the sound lacks the incisive (and sometimes fatiguing) edge of the harmonic exciter. On the benefits side, the process generally sounds very smooth and natural.
The dynamic nature of the high-band processing is a real advantage when you're dealing with noisy material, as boost is only applied when the signal level is high enough to mask any noise that might be brought up by the dynamic control. On some material, the Manual setting produces a more even result than the Auto setting, which varies more noticeably as the input material dynamics change. Which setting to choose is down to personal taste.
Another way to enhance sound is to create new high-frequency harmonics not present in the original recording. This principle was original discovered by Aphex, who found that adding a small amount of carefully controlled distortion to the original signal could actually make the sound quality appear cleaner and more detailed than it was before processing. The way it works is that some of the input signal is diverted, via a side-chain and a high-pass filter, into a harmonics-generating circuit. A high-pass filter is necessary to remove unwanted low frequencies so that the newly generated harmonics exist only at the top end of the audio spectrum, and this filtered signal is then processed dynamically to add phase shift and to create synthesised harmonics related to the original signal. The secret of the harmonic enhancement principle is that only a very small amount of the processed signal needs to be mixed back in with the original, and when this is done properly, music appears more detailed, with better separation between individual sounds.
Both harmonic enhancement and dynamic equalisation processes are useful for treating production master tapes for use in cassette duplication, where the processing can help to compensate for the loss of clarity inherent in mass cassette duplication. However, enhancers are also extensively used in mixing, to make specific sounds stand out more in a mix. While it's possible to enhance a complete mix, doing this means that you lose a valuable opportunity to create contrast by enhancing just those elements of the mix that you want to stand out. As I've said before, enhancement has the effect of putting sounds 'up front', and if you put everything up front, there's nothing left out back to create the illusion of depth and perspective that a good mix needs. Harmonic enhancement has been used on many occasions to make a lead vocal more intimate, but you do need to take care to avoid sibilance, as any form of high-frequency enhancement tends to exacerbate it.
In recent years, SPL have acquired a certain amount of mystique with their Vitalizer process, not the least because nobody outside the company knows exactly how it works. Their process involves mainly equalisation, though the secret of their sound apparently hinges on an interdependent network of dynamic filters. The process works by combining a side-chain signal with an unprocessed main signal, in such a way that the original signal is modified both additively and subtractively. Subjectively, the process creates the impression of an increase in both bass and brightness, while the mid-range is brought into sharper focus, increasing the sense of transparency. Apparently the spectral shaping is closely related to the way the response of the human hearing system changes at different listening levels.
Though the SPL process is quite complex, one of its tricks is to add low-frequency equalisation in such a way that phase cancellation occurs in the lower mid-range, thus preventing the effect of the bass enhancement from spilling over into the mid-range. At the high end of the spectrum, a circuit employing fourth-order filters is used to pull out transient detail; again, some form of dynamic processing may be involved, but, wisely, SPL have kept the exact process secret in order to avoid the most sincere form of flattery.
On the original Vitalizer, a Process Depth control determined how much of the output from the Sub-Bass and Mid-High filters was added back into the dry signal. The Harmonics control operated independently, though a tuning control defined the area of the mid-range to be processed and also affected the feed to the Harmonics circuit. Apparently this derived its input partly from the untreated signal and partly from the output of the Mid-High filter. Though the term Harmonics is used, I'm told that this circuit emphasises existing harmonics rather than creating new ones.
SPL's Bass Process control is interesting: the control has a centre-off position and produces two different bass characters when moved right or left from centre. Clockwise, the bass becomes tight and well-controlled; counter-clockwise, the bass end becomes very full and deep -- ideal for rave music. Because of the dynamic nature of the process, it isn't possible to duplicate these effects using conventional equalisers.
Subjectively, the SPL process seems to add more life to the whole audio spectrum, increasing the impression of loudness, detail and space. As it would for any enhancer, the subjective result has a lot to do with psychoacoustics, but this is a great tool for use in pop music production.
Dolby also have an enhancer based on dynamic filtering; like the SPL systems, its exact filter configuration is unknown. The Dolby approach relies on treating the side-chain signal via a bank of complex filters which modify their characteristics according to the nature of the input signal. The filtered signal is then compressed before being added back into the main signal path. A number of companies, including BSS, also produce dynamic equalisers that combine elements of EQ and multi-band compression. These may be used to create a number of different enhancement effects, as all the relevant parameters are accessible to the user.
Which type of enhancer you use is a matter of personal preference, but I can offer my impressions of what the different types do best. Without a doubt, harmonic enhancers are best at replacing missing high-end detail -- old guitar strings can be made to sound (relatively) new, dull electric pianos can be made to sparkle, and tired old analogue recordings that have been under the bed for the past 10 years can be given a new lease of life. Because harmonic enhancers process the high end of the spectrum, and because some degree of level compression is used in the side-chain, any noise present in the original signal will also be 'enhanced'. For this reason, it's sensible to use as little enhancement as will do the job. Also, make frequent checks between the bypassed and enhanced sound, as it's easy to get used to over-enhancement -- when you come to listen to the mix in a day or two's time, you may find you've gone too far and made it fatiguingly bright. As a rule, the overall tonality of the music stays much the same, but high end is enhanced.
Phase-shift enhancers, such as those built by BBE, tend to be rather less effective at rescuing desperately dull mixes -- but, on the other side of the coin, they tend to be kinder to well-recorded material, and it's possible to add quite a lot of processing without the sound becoming harsh. Again, there should be little overall change in tonality, but the sense of detail and separation should be improved.
Those systems that rely on dynamic filtering can be used just to enhance top-end detail, but units such as the SPL Vitalizer can also be used to radically change the character of the whole audio spectrum if you so desire. At the top end, the result is similar to the other enhancer types -- detail is lifted out and instruments are easier to separate -- but as you come down the spectrum, it's also possible to tighten up the mid-range and to add a generous amount of low-end clout. Hit the bypass button and you'll notice the tonality of the whole mix change; if the controls are set up properly, it should change in a natural and beneficial way. The control system of SPL's devices is slightly less intuitive than that of the other main enhancer types, but perseverance pays off with results that combine the attributes of a good EQ with a very effective enhancer.
As well as the risk of enhancing noise, users should be alert to the risk of sibilance due to processing vocals that are already bright, and also to harshness due to overprocessing. In the world of enhancement, the saying 'Less is more' really is valid. Where possible, use the process to create contrast within a mix, and make sure that your processed mix isn't fatiguing to listen to. The idea is to introduce a little air and space into your music, not to make your ears bleed!
One important psychoacoustic principle is based on the fact that our perception of the audio spectrum changes as sounds become louder -- those familiar with the textbook Fletcher Munsen curves will already know that at high listening levels, low and high frequencies tend to be perceived as being louder; at lower listening levels, the mid-range is more dominant. Using an equaliser to cut the mid-range or to boost the high and low extremes can make music sound louder, which is how the loudness button on a hi-fi system works.
The Aphex harmonic principle is also based in psychoacoustics. Whenever an audio signal is subjected to distortion, intentional or otherwise, high-frequency harmonics are produced. Normally these sound pretty unpleasant, but by using filters to confine the distortion to the upper reaches of the audio spectrum, it's possible to 'fill in' missing or weak HF detail in a manner that the human brain will accept as natural. The dynamic element of the process applies the most harmonic enhancement to transient sounds, producing an apparent increase in detail, presence and loudness.
Devices that affect phase shift trade on the fact that, in real life, high-frequency sounds travel slightly faster than low-frequency sounds, so the further you are from a sound source, the more the high frequencies lag behind. Delaying the bass and mid-range to compensate for this effect makes sounds seem nearer and more immediate.