Abstract
Compressors were once only used as a protective barrier for radio broadcasts, using basic limiting techniques, designed to protect the transmitters from over-modulation. This developed over time, to radio stations that used compressors as a tool that could be employed to increase the perceived loudness of music played, over preference to other stations – a competition if you will, between rival stations to see who could be ‘louder’ and gain more listeners. Because of the limited bandwidth of radio, and the limitations of recording to analogue tape, compressors were also used in the recording studios of the time. But it was pioneers such as Joe Meek that introduced ‘creative’ compression techniques, such as ‘pumping’ and squashing the sound. In other words, he was using techniques that flipped all the conventions of the time upside-down.
Jumping forward nearly four decades, we are now in a transitional period production-wise. Artists no longer need the use of large expensive studios to produce broadcast quality music, and with analogue recording almost extinct, 24bit digital audio has now become the industry standard, with it’s increased dynamic range over 16bit digital audio. It still has some way to go to compete with analogue two inch tape, in terms of 'feel', but due to audio now being kept in the digital domain, analogue noise and hiss (as we once knew it) has become a thing of the past. 16bit ‘CD quality’ audio was the industry standard for about 15 years, during the reign of the CD, but as we near the DVD age (it is still not entirely established just yet – almost there!) and 24bit audio becoming the norm, manufacturers will be busy devising yet another format they can ‘flog to the masses’. It could see a doubling of bit-depth, or, more realistically from a marketing point of view, it will be 32 bit audio, as companies would never pass up the opportunity of missing any steps on the path to achieving their ultimate aim – making money.
As the bit-depth of audio increases, so will the dynamic range that accompanies the format. With this, the inherent need to compress audio into a limited audio bandwidth will diminish. And with the demise of analogue radio into a vanishing blip on a distant planetary system’s radar pattern, the use of compressors will change forever, as no transmitter would be used. Broadcasts would be over a hybrid internet-type line where compression would be of data, not audio.
With the introduction of higher dynamic ranges into systems, it may encourage the invention, and discovery of whole new forms of music, that would have been impossible to record using the old formats. The increased frequency range would also allow instrumentation to be free of the old constraints of 20hz-20khz. Nuances previously only calculated on graph paper would be heard and felt for the first time, creating the most life-like recordings ever possible, as it is now widely acknowledged that instruments contain harmonics well above human hearing; this may well affect and modulate harmonics within the standard range of 20hz/20khz in a way that is not fully understood yet. Harmonics outside our hearing must affect us as humans, it’s just that we cannot detect them individually – only their interaction with other frequencies within our range. Not only are we limiting our capabilities by using mainly 12-tone equal temperament in our music, but also by limiting the frequency and dynamic range of music we listen to, we close our ears to possibilities that are screaming out to be musically discovered. We are on the brink of something new and exciting, which is beginning to show itself in the tale chasing antics of current popular music and culture. Will compressors be a part of this change? Maybe, but they are not used in the same way they were originally intended, and in this article I will discuss the implications that technological advancement will have on the use of compressors in the future.
Introduction
The compressor has become the essential tool in both music production and in broadcast. What would the world be like without it? Well, television programmes would always be either too loud or too quiet. Viewers would get extremely annoyed and frustrated with the constant level changes, and the tin(n)y TV. speakers that have been commonplace up until now would seem even more muffled and distorted. Music on radio would also sound dire – am broadcasts are bad enough, but FM would sound almost as bad were it not for the compressor. Stations wouldn’t dare risk blowing up their transmitter transformers, so they would broadcast at very safe and extremely quiet levels. Rock and dance music would not be the same – there would be a distinct lack of power in recordings – especially in contemporary dance music, where squeezing the last drop of dynamic range out of every instrument has become common practice. Classical music, on the other hand, exists at all levels (dynamically). If a piece of classical music were squeezed as much as pop music, it would lose all its character, and in turn would sound flat and uninteresting. The whole ethic of classical music relies on the basis that one can have very quiet passages, building up to crescendos of power, which then transcend into the depths of serenity – all in one piece of music. Power and Loudness seem to be the demigods of modern music, but with it they have killed off a lot of the true feeling that is present in the real world. Engineers seem to worship these two gods (and so do the listening public) and are ever seeking to squeeze bigger music into ever-smaller boxes.
Need for compression
A compressor’s basic concept is relatively straightforward. It works by adjusting the gain in relation to the input signal level – keeping peaks under control, whilst making softer passages louder. Compression and limiting are both the same process, but the term ‘compression’ is used for ratios that are below 10:1 (a change of 10dB or less of input signal level results in an overall change in output level of 1dB). Above 10:1, compressors will more abruptly prevent signals above the threshold level from exceeding a certain maximum value – this is known as limiting.
There are several uses for compressors, or dynamics processing, as it is also known; one of which is applied to individual or groups of instruments in multi-track recording. For radio and television broadcasting and mastering, entire songs are compressed as a whole. It is a current necessity in broadcasting (to protect from distortion or other unpleasant effects, and also to make things seem louder), but in electronic music, it is more often than not used for a more creative purpose.
Why is ‘louder’ better?
For some reason, although the life has been squeezed out of audio, the listening public seem to love this type of sound – ‘loud at whatever cost’. Maybe this could be compared to adding sugar to food. At first people don’t notice - they love the taste, but are not consciously aware of the reasons ‘why’. An example of this is Walkers crisps. Don’t they taste great? But on the same level, don’t they taste sweet? In fact, they have lost almost their entire savoury ‘crispness’, in favour of saccharin ‘sweetness’. Most of the general public would neither care nor notice such a seemingly trivial thing, but when you’re a perennial crisp lover, it becomes torturous on the taste buds. Maybe they have gone too far with the ‘sweetener’, and Smiths crisps will pick up on it and cry ‘cheat!’ but more than likely it will continue un-noticed. This very same principle can be put to the use of ‘loudness-maximisation’ on music. People adore that sound, but why is this? It may have more physiological reasons than we realise.
Maybe our hearing process favours sounds that are more-or-less at the same constant level, making interpretation easier for the brain. A little like data compression makes reading files easier for computers. There are the obvious reasons for compressing audio to protect sensitive equipment, and to stop transients from clipping. But that doesn’t link logically to why we prefer that type of sound. It may only be a current state of collective mind, which musical fashion is depicting at this present time, but ever since the dawn of recording, the human race has sought ever louder artificial audio reproduction.
When A&R people receive demos, they generally file the dats or CDs to the filing cabinet under the desk if the loudness does not meet their expectations. Even if the general mix is ok - if it hasn’t been ‘loudness maximised’, the artist can kiss goodbye to ever getting a contract, unless they are exceptionally talented at harassing people. This shows that it isn’t just the general public at street-level that love things to be ‘loud and proud’ – people of all level seek satisfaction from the dynamic demon. This isn’t necessarily a good thing though.
History of recording
The increase in dynamic range has inherently affected the signal-to-noise ratio in recordings made over the last century. Thomas Edison invented the phonograph in 1877 Shellac disks made before 1925 had a signal-to-noise ratio of 30dB. This was due to a weak analogue signal together with the noise caused by abrasive fillers in the shellac compound. These compounds consisted of shellac (actually less than 20%), vinsol moulding modifier, Congo gum binder, white Indiana limestone filler, red Pennsylvania slate filler, carbon black colouring, and zinc stearate lubricant. The shellac presses were extremely hard wearing, and due to the abrasiveness of the materials, the needles needed for playback could only be used once, before being rendered useless. After 1925, new microphones, amplifiers and speakers were being introduced, which made possible the boost in amplitude of sound waves by electrical means. Shellac records ceased production during the Second World War, but there were exceptions (if it was deemed as ‘supporting the war effort’). During the war, the U.S. developed a shellac substitute, called ‘vinylite’, which is now the substance of choice in the records of today. This polyvinyl chloride plastic used after the war did not need abrasive fillers, meaning there was less noise produced - the signal-to-noise ratio was rated at 60dB.
Moving onto the digital age, noise became a thing of the past, where sound reproduction was free from any surface contact associated with previous noise creation. The compact disc (invented in 1980) saw audiophiles biting at the backside of companies such as Sony and Philips, whilst claiming that the quality of CD didn’t match up to that of good quality vinyl. These protagonists gradually quietened down, as CD began to override the vinyl market in the commercial mainstream. Vinyl continued to flourish in the underground counterculture (now moulding into the mainstream), but that’s a different story for another day.
It seems though that these audiophiles were listened to, because we are now seeing yet another advancement in audio recording – that of DVD-Audio. Although it utilises the same medium as DVD-Video, the two should not be confused. DVD-Video, and the audio that accompanies the picture, uses an entirely different codec to DVD-Audio. Companies are currently using hard-sell tactics to convince the public of the advantages of the new video medium, but the picture quality is not what some might expect – it has been thoroughly compressed (using data compression). To put things into perspective, an uncompressed feature-length movie would take up 40 DVD discs, which shows the extent of the compression used. The audio version of the disc, however, has been optimised so that the music will be uncompressed. This is not to be confused with the compression being discussed in this article, but it is worth mentioning, as using MPEG data compression severely compromises quality, both of audio and video.
DVD-Audio
When DVD-video was released in 1996, it far exceeded the quality of VHS, and the audio capabilities were better than CD – even though the medium was not optimised for audio yet. An industry group called the ‘DVD Forum’ was responsible for establishing an industry standard optimisation for audio DVDs, and they released the final version 1.0 of the spec in 1998.
One of the advantages of DVD-Audio over DVD-video and CD, is the increased quality of Pulse Code Modulation (PCM) used. The range of frequencies are four times that of a CD, giving instruments the most live feeling yet in a recording. The dynamic range is also far greater than is possible on CD, which is ideal for classical music, where the loud parts are louder and the quiets are a lot quieter, without noise quantization. The signal-to noise ratio is also increased.
If you compare the 20hz-20khz range of CDs, to the massive 0-96khz frequency range of DVD-audio, and 144dB dynamic range of DVD-audio to the comparatively whimsical 96dB of CD, it is a giant leap forward.
As this new format starts to replace CD albums in the shops, it will push some artists to utilise the new-found dynamic and frequency responses in their music. Although for the time being, because of the limited speaker technology (which no doubt will step onto the digital podium sooner or later), the music will at first be enhanced only by the modulating properties of incidental harmonics (to coin a term).
And where do compressors fit into the equation? Because up until now, it’s almost become a standard ‘quick fix’ for making audio seems as loud and ‘up-front’ as possible, within the technological limitations of the past.
Good question indeed. Of course they will still be used, but this step up in technology is the first nail in the compressor’s coffin. Neve (stalwarts of monstrous analogue consoles), have already released a floating-point 32bit digital desk, which (theoretically) has an unlimited dynamic range. Now all manufacturers need to do is develop another medium (and speaker) with the same response!
This will also be a money-spinner for the music industry, as they have lost control of the CD market, due to piracy and ease of duplication. With the new format, they will be able to re-release a lot of their back catalogue, as ‘re-mastered’. They won’t necessarily use the 5.1 system available to DVD either, as this limits the length to 74 minutes. If only a stereo mix is recorded, it can hold about 600 minutes of 24bit 96khz audio, which is more than enough. Not that most artists would utilise the space, either. It would come into it’s own for concert recordings and compilations, and used as a showcase for such items – almost as a novelty at first. At this moment in time, there are very few recordings by mainstream artists, that have utilised the DV format, but the trickle has begun.
Our hearing dynamic range
The human ear is far less sensitive to changes in amplitude than frequency, but the hearing response also changes with frequency.
The loudness curve seems to vary dramatically throughout the frequency range heard by humans. The threshold of pain is generally understood to be somewhere in the region of 120-130dB. The inner ear has an inbuilt ‘compressor’ to combat loud sounds from damaging the membrane and hair cells. The tensor muscle and stapedius muscle automatically reflex to tighten the tension of the eardrum. This acts as a natural dampener to the sound, causing the tympanic membrane to not move as much, which in turn reduces the air pressure behind the skin of the drum.
Digital Broadcast
Radio has been broadcast since the 1920s, starting with ‘Amplitude Modulation’ (AM), which was widely used up until the 1950s. ‘’Frequency Modulation’ (FM) was then introduced, and has become the main standard used for high quality music radio. Both of these standards have required immense audio compression. AM needs more surprisingly enough, as it has a very limited bandwidth compared to FM. For FM transmissions, broadcasters now use multi-band compressors, which tame the audio into the lowest common denomination (in-car and transistor radios).
The new medium being touted is D.A.B. (Digital Audio Broadcasting). There are many differing standards currently being used throughout the world – the US uses I.B.O.C. (In-Band On-Channel), which utilises the same bandwidth as the old analogue transmissions. This means that traditional radios can still pick up the signal, but there is a more sinister reason in that the US military have outlawed the use of L-band transmissions for radio, as it is the frequency in use by their military. Europe, however, are utilising the L-Band for their transmissions. The European ‘Eureka-147’ system transmits at a rate of 1.5Ghz, which means that old analogue radios cannot pick up the digital signal. This seems to have opened up a whole new future bandwidth for radio, now that the old analogue airwaves have become saturated. The much higher frequency will also mean that there will be logarithmically more space for more stations.
Vintage Compressors – why are they sought after?
One of the best-known vintage compressors was the LA-2A levelling amplifier. They were hand-wired, and originally patented by Jim Lawrence. Produced by Teletronix in Pasadena, California, they became a division of Babcock Electronics Corporation in 1965. Studio Electronics Corporation acquired Babcock’s broadcast division two years later. Shortly afterwards, the C.O. Bill Putnam changed the company name to UREI. Three versions of the LA-2A were produced under these different companies before production ceased in 1969.
Because of the rise in popularity of transistor technology, valves took a back seat in the world of consumer audio. The development of solid-state technology decimated the old companies, who could not compete with the new manufacturing methods. This may have been due to the economic downturn of the early seventies, or the unwillingness of such established companies to change with the times. Whichever of these it was (or a combination of the two, plus other unmentioned factors), It caused the downfall, and extinction of many manufacturers that had dominated the market for decades.
The success of the transistor may have been it’s own shortcoming, however. Companies tried desperately to keep up with demand, and due to the high volume required, quality was not high on the agenda. Producers noticed that equipment that ran using transistor technology sounded crisper than the equivalent valve products, but something was missing in the blueprint of producing a great sound. There were new distortions to contend with, which sounded ‘harsh’ and unpleasant. A similar argument ensued as between digital and analogue recordings today. There were some people that argued that transistor-based equipment was clean and crystal clear, compared to the valve gear that was beginning to age. It has been said by Manley Labs, that -
“..the early designs used noisy carbon resistors, puny power supplies and the transformers were far from flat. Some coloration crept in with time. Capacitors dried out, connectors and switches corroded and pots became scratchy. Much of what people attributed to tubes (valves) was actually everything but the tubes.”
What Manley were saying, was that a lot of people who, at the time, said that old valves sounded worse than transistors, were actually hearing the attributes of ageing pots, dodgy connectors and defunct capacitors. If these had been replaced, the valve gear would have sounded far superior to the new transistor gear that was being championed. Why would it have sounded better? This is the ‘million dollar’ question. Many listening tests were conducted, but the finger could not be placed on a rational reason what sounded better. They had all the answers as to why. Such as manufacturers knowing that - (ctd. next page)
“..the common transistor design practice using large amounts of negative feedback to correct linearity damaged the transients. We did not have to resort to feedback to produce a good sounding circuit, but instead we were able to use much less feedback and in better ways.”
Virtually Vintage – back to the future?
A prime example of manufacturers succumbing to the ‘if you can’t beat 'em, join em’ school of thought, Universal Audio have put their vintage expertise into a native DSP card (UAD-1DSP) for use in programs such as Steinberg’s ‘Cubase’ and Emagic’s ‘Logic Audio’ series. The card contains a single ‘Super-DSP’ chip, which, according to the literature, outperforms similar cards with multiple inferior chipsets. The literature also claims that their calculated algorithms perfectly replicate the circuit designs of the hardware it seeks to emulate.
The original Teletronix LA-2A (which is one of the supplied software plugins) worked using a controlled attenuator. This consisted of an electro-luminescent panel, driven by the tube peak detector, shone onto a light-sensitive cadmium-sulphide resistor.
This really does have serious implications on the future validity of using classic outboard gear, as the previous excuse was that digital was ‘too cold’, or harmonically ‘stark’. With the power of cards such as this now a reality, and a reasonable price range to boot, there really seems to be no reason why one should even think about spending $4000 on a compressor, however good it sounds.
Fashion
A popular buzzword recently has been the ‘Warmth’. Everybody wants it in his or her music, but what exactly is it? There have been numerous tests, which show that, in general, tube-driven gear can give a recording a certain type of distortion that is almost unattainable using solid-state transistors. This could actually be attributed to using electrolytic coupling capacitors and inexpensive op-amp integrated circuits, rather than the solid-state devices themselves. However, the argument still rages in this area. An article on the Internet, quoted John Atwood (consulting engineer and owner of One Electron Company in California), as saying –
“Some of the differences in audio qualities between tubes and transistors have to do with the inherent physical properties of the devices and with the circuit topologies and components used with each type of device. There is no way around it: linear [triode] vacuum tubes have lower overall distortion that bipolar transistors or FETs, and the distortion products are primarily lower-order… the clipping characteristics of tubes is actually not much softer than transistors, but feedback tends to ‘square up’ the clipping. Thus, the heavy feedback in most solid-state designs gives them worse overload performance.”
He continued by saying –
“A low- or no-feedback design can be driven harder without audible distortion. High feedback also can lead to transient intermodulation distortion (TIM), caused by clipping or slew-rate limiting within the feedback loop”.
In the same article, it puts forward the argument I have already mentioned – that the inherent ‘warmth’ attributed to the valves, is actually the transformers that accompany them. -
“The warmth is created by a large component of second-order distortion, and the slow rise time of the output transformer causes a coloration that I would describe as a smoothing effect.. the transformer is a non-linear element that causes alterations of the signal in the time and frequency domains, thereby altering the sound.” (From audio expert Bruce Rozenblit, owner of Transcendent Sound Company, Kansas City).
Manley Labs, a high-end professional vintage gear manufacturer based in California, mirrors the argument about the transformers being the real source of ‘warmth’.
“…a simple passive component like a transformer falls into the category of ‘complex distortion’. Saturation first. In a reasonable transformer the saturation is dependent on signal level (power) and frequency – The lower the freq, the easier it will saturate. The saturation itself is mostly a function of the magnetic core size, core material and number of turns on the primary windings as well as the circuitry around the transformer… the biggest part of reason people describe vintage gear as warm sounding is due to the transformer saturation.”
It goes on to explain that a good transformer will have a flat response down to 20hz, and although most speakers do not go down that low, there are odd harmonics from the low frequency saturation, that the listener will still hear (even though the fundamental will be inaudible). This, according to Manley, is why the bottom end becomes bigger and ‘phatter’.
If this is all that is happening, there is absolutely no reason, in this new digital age, why this cannot be replicated by DSP.
And why do we still use an ageing technology such as the current transistor? Surely the music equipment industry has grown to such a point where they would refuse to rely on something that only serves as a ‘quick fix’? It seems that companies are resting on their laurels when it comes to advancing the technology of amplifying audio.
Conclusion
Audio compression has been omnipresent since the 1940s, and will probably still be around in 2040. Although the technology is emerging that allows less compression to be used (DVD-Audio), there are always the trail of former technologies that are revitalised and adapted for use in areas previously unheard of. A non-musical example of this is computer games and their parent consoles. In 1980, Nintendo released the entertainment system, with its 8bit technology at the forefront of its time. 20 years later, and Nintendo introduce the Gameboy Colour, utilising graphics of a similar (if not better) quality, but downsized to the pocket. Mobile phone games are following a similar, if not simpler path, with ‘snake’ games appearing on the Nokia phones only a couple of years ago.
These were originally games of the seventies, but the advancement in mobile phone technology will very soon allow games of 16bit quality, as well as high quality audio and video on a handheld phone. It seems the decades are being re-run in some sort of compressed format, maybe at a ratio of 5:1. And so it leads to the use of digital audio, and the need for digitally compressing the data so that it will useable for a variety of formats, from streaming on the Internet, to D.A.B. radio transmissions, to mp3 players the size of a credit card. All of these formats will have a reduced dynamic and frequency range contrary to the increase with DVD-Audio and its future siblings. Therefore, compressors are here to stay for the foreseeable future, even if their analogue forebears are gradually phased out of the ‘real world’. They will exist more as algorithms on computers, and this will no doubt be the catalyst for integration into digital hardware compressors using the same calculations as current DSP cards have put forward, for stability that is not as guaranteed in software-only based applications.
So to sum up, compressors will still be used in mediums that need to limit the dynamics of audio for radio broadcasts, due to receiver limitations. Audiophile music will begin to flourish, and there is beginning to be a two or three tier system in audio quality. Audiophiles still prefer vinyl, but DVD-Audio will gain a sizeable following – especially amongst classical music fans. A movement of ‘Anti-compressionists’ will possibly form amongst studio engineers of various types of hi-fidelity musical forms. That may be a long way off yet though.
This was written back in 2002. (c)2005 vexin (K.Roland) All Rights Reserved.
