Dynamic Enhancement

Compressor Uses and Abuses

Compressors were originally devised to reduce the dynamic range of audio signals. To do this they use a gain cell that makes adjustments to signal levels automatically dependant on the level and dynamic of the signal itself, and depending on the character and control settings of the particular compressor unit.

Over the years many different brands of compressors have been produced using many different types of gain cell technology; each with a distinctive sound. Users have found applications that suite many of these taking the use of compressors far beyond simple dynamic range reduction, sometimes even generating the complete reverse; dynamic enhancement.

Applications include:-

• Protecting systems from amplitude overload
• Improving the power of signals when an artist performance is inconsistent
• Maintaining a stable signal level to help an instrument sit better within a mix
• Changing the percussive timbre of an instrument to produce more punch
• Compressing a mix so that it maintains a stable signal level (typically trying to make it sound louder)
• Changing the percussive timbre of a mix to make it sound more whole and add punch
• Frequency conscious compression to remove unwanted sibilance or popping

Very often the process of compression produces unwanted side effects especially if the compressor chosen does not match the application it is being put to very well.

These can include:-

• Breathing (when high frequency environmental or processing noise levels jump up and down sounding similar to someone breathing in the background)
• Pumping (when level modulation can be heard on an instrument making it sound unnatural)
• Amplitude flutter (a faster oscillating modulation that makes sounds “gritty”)
• Source inter-modulation (when one sound source within a mix causes amplitude changes to another; for example, the natural decay of cymbals can be modulated by the crack of a snare drum in a drum mix)
• Transient enhancement (an unwanted accentuation clicks at the start of a sound)
• Dull sound (when all the natural attack and brilliance is stripped away from the sound source)

Providing controls for threshold, attack, release, ratio, and providing responsive metering can go along way toward helping with suitable parameter set up for any signal source such that most of the negative effects above can be reduced or eliminated. However, this is a complex and critical task which often involves a degree of compromise between the desired compression effects and the unwanted side effects, which can be time consuming.

Simple Compression

For many every day compression jobs an RMS compressor with automatic time constant set up (attack and release) is the best solution providing quick and easy set up. The RMS averaging process slows the time constants on relatively steady state signals reducing distortion and pumping; and when large signal changes occur they automatically speed up capturing and constraining the bulk of any large sound level variations.

Very often a Ratio control and Threshold control combined with the automatic time constant adjustment just described is all you need to set up a good sounding compression. The DN540 auto compressor operates in exactly this way providing simple fast set up on straight forward compression jobs.

Advanced Compression

RMS compressors are not fast enough to capture everything on transient material because the RMS averaging process always adds some delay; thus they are not suitable for ultimate protection against system overloads etc. Also their creative use to tailor percussive instruments is very threshold dependant and often results in attacks that are either too fast or too slow for the desired effect (unless the sound source is extremely regular, which is not typical with most musicians).

For more difficult compression duties a compressor with fully adjustable attack and release is a better choice. With this style of compressor there is no averaging process delay (RMS detector) so the action of attack or release can start the instant there is a change in signal amplitude that requires it. The user must define the rate of response and can adjust this precisely to match the sonic effect required on the source material. Normally this also results in high distortion on constant signal levels because the compressor attacks and releases on every cycle but advanced compressors utilise windowing methods whereby the time constants set are greatly increased on steady state material. Another technique used is to automatically apply a little hold before any release commences.

The above techniques are integral to the semi-linear attack and second order release characteristics of the DN540 “normal mode” compressor (i.e. when auto is off).

Linear attack provides a constant rate of attack (in dB against time) such that large changes in program signal level take a little longer to compress than smaller ones. However, on material with more constant signal levels the attack rate of the DN540 “normal mode” compressor automatically reduces. This appears as a curvature in the linear attack rate characteristic as it nears completion, hence the term semi-linear.

This makes the compression very transparent providing some dynamic control but without unduly effecting the intentional dynamic content of the source material.

It can be used on difficult instruments like acoustic guitar with slower attack time settings and relatively fast release to keep equal perceived loudness within a mix without producing excessive amplitude flutter or distortion.

It can also be used with faster attack times to capture dynamic instruments like electric bass guitar without adding excessive distortion on constantly compressed passages.

Adding “soft knee” noticeably delays the onset of attacks, which can be particularly useful on drums where compression can be applied to emphasise transients giving more punch while retaining a good deal of artistic dynamic from the drummer.

Thus when suitably adjusted the normal compression mode is suitable for any task from capturing fast transients in order to provide system protection; producing subtle compression of dynamic range without changing timbre or removing intentional accents made by the artist; to deliberate thickening of transient sounds.

Soft Knee Compression

Many vintage compressors exhibit a soft knee transition between linear gain transfer, on signals below threshold, to compression on signals that are above threshold. This normally occurs because of non linear errors in the gain cell. Modern compressors normally have much better gain cell linearity so any soft knee has to be generated in a separate process that operates outside the gain cell.

The DN540 soft knee generation is set up so that it does effects the gain cell like vintage compressors. This means it bends the compression ratio at the onset of compression as you would expect but it also bends the attack and release character. This provides very natural sounding compression as found on many vintage compressors and generates more full bodied punch and definition when attach times are deliberately set very slow.

Dynamic Enhancement

Many instruments have a percussive start to notes that are played. These contain the bulk of the signal harmonics that are recognisable and that we use to distinguish one instrument from another. Without this initial attack most instruments sound quite similar - and very dull! Unfortunately this is what tends to happens on compressed sounds sources.

Compressors capture much of the percussive start and reduce it in level more than they reduce the remainder of the sound. It’s not as extreme as totally removing the start of notes but it still strips much of the harmonic content and removes presence from the sound.

This can be corrected by using equalisation to boost the upper frequencies but this is dangerous in sound reinforcement because when the instrument is silent and the compressor relaxes there is no gain reduction; but the upper frequency boost remains increasing noise and making microphonic feedback much more likely.

The Dynamic Enhancement in the DN540 was developed to correct the tendency for dull sound during compression without changing the sound of uncompressed signals. It works by reducing the ratio of the compressor in a relatively broad range of frequencies centred on 5kHz and the effect is continuously variable so as much presence accentuation can be added as needed or to suite taste.

When Dynamic Enhancement is used, any applied compression acts on high frequencies to a lesser extent than low frequencies and by a predetermined amount set by the user. The transient start to sounds are be greatly reduced in the low frequency content, which is where all the power is (that needs to be controlled), but the harmonic content is preserved in a more natural and dynamic state.

The presence band is an area where we are most perceptive to sounds and reducing compression ratios in this area allows much greater compression to be applied at the other frequencies for what ever reason without sounding unnatural.

Dynamic Enhancement compression can produce results that are very similar to multiband compression but with only one addition control required (as opposed to many) it is much more straightforward to set up and uses far less rack space.

Some additional corrective benefits from Dynamic Enhancement are listed below:-

Reduction of source inter-modulation – it is very common for the pop/rock singers in to stand in front of drum kits and unless suitable screens are placed between them there will be a lot of spill from the kit. Often you can hear the compressor on the vocalist modulating the spill from cymbals which sounds very unnatural. This can be eliminated by using the Dynamic Enhancement to stabilise the higher frequencies; and if that makes the vocals sound too bright the high frequencies can be reduced a little using EQ reducing the spill and reducing the chances of high frequency howl round.

• Reduction of breathing – most processing noise occurs in the presence band and it is much more noticeable when it is modulated by the compression of a sound source. Just as above presence masking can eliminate this effect.
• Reduction of pumping – because human hearing is so sensitive to presence band frequencies, stabilising their dynamic response with presence accenting can mask the pumping effects at lower frequencies.
• Reduction of dull sound and increased brilliance – it is often tempting to boost high frequencies to get sounds to cut through a mix but in sound reinforcement this increases the likelihood of howl round. Dynamic Enhancement can help increase of the brilliance of compressed sounds by correcting their tendency to sound dull, and with increased use it can be used creatively to add even more high frequency energy without increasing the risk of howl round. This can be used to great effect on vocal both sweetening and improving their ability to cut through a crowded mix.

The Dynamic Enhancement feature contained in the DN540 adds another tool for the sound engineer that can be used to enhance and compliment the application of many types of compression as outlined earlier in this paper.