There are several types of eq. Let's have a look at each general type of filter which is commonly used in equalisation process.
High Pass Filter
A high pass filter does what it says on the tin. It allows high frequencies to pass through it and attenuates all others. On
analogue desks this type of filter would often have appeared as a switch with a predetermined roll-off frequency of
between 50-120Hz. This would mean that all frequencies below 120Hz would be attenuated at a particular rate. The rate is know as the slope of the filter and is defined in terms of dB per octave. Typically a high pass filter would attenuate at a rate of —6, -12, or —18dB per octave although in the digital domain —24dB per octave is not uncommon. We know that
every time we descend an octave the frequency is halved. This means that if the HPF has a slope of —6dB per octave and the filter starts at 120Hz then at 60Hz the level is 6db lower. On a technical note, any frequency that is attenuated by less than 3dB is said to be within the pass band. The cut-off frequency is where the filter has already attenuated by 3dB. Frequencies that are attenuated by more than 3dB are said to be in the stop band. A low cut is a HPF with a slope of -18 or —24dB per octave.
Low pass filter
A low pass filter works on exactly the same principles as a high pass filter except that it operates at the other end of the frequency spectrum. It lets all the low frequencies pass through and attenuates the high ones.
High and low pass filters are used for removing unwanted frequencies at either end of the spectrum where the wanted frequency will not be affected by the cut. For examplere moving low rumble from a vocal track or perhaps removing high hat spill from a snare track. In recent years they have also been used to great effect in dance music production. Drum tracks or even whole mixes are subjected to a low pass filter that gradually moves down the frequency spectrum cutting out almost all the high frequencies so that the music sounds as though it is being played in a next door room. The filter is then opened out again at a suitably dramatic point in the music for a great build up. High pass filters are often used on drum loops during a drop giving them a small and tinny sound. The filter is then opened out again suddenly or gradually at the end of the drop section giving a sharp or a smooth transition back into the main musical theme, as desired.
Shelving filters are a much more musical type of equalisation than HP or LP filters. They are designed more for gentle overall tonal shaping than for keeping out rogue frequencies. Rather than introducing an increasing gain change above or below a certain frequency, shelving filters apply an equal amount of cut or boost to all frequencies above or below the designated shelving frequency. Typically there would be two controls one for selecting the frequency and the other for selecting the amount of gain change. The transition from the OdB reference point to the designated cut or boost level of the shelf is not immediate and so this type of filter also has a slope. Whilst it would be theoretically possible, in the digital domain, to design filters that arrive at the shelving frequency almost immediately, these have been found to be less musically pleasing and thereby less useful for audio processing than those with a gentle slope.
Band pass filter
A band pass filter passes frequencies between two limits. It is effectively a low pas! and a high pass filter working together. Typically this filter would have controls for determining the central frequency and the bandwidth (the distance, in frequency terms, between the high and low cutoff points). This type of filter can be used for isolating a narrower band of frequencies in recordings that have too much high and low end or more typically it is used as an effect to create that nasally mid range typ of old radio voice sound that we all know and love! Band pass filters are also very widely used in synthesizer design and you can create interesting effects by using a narrow bandwidth and moving the central frequency around.
A notch filter does the opposite to a band bass filter. It is also know as a band reject filter. Frequencies above the upper cutoff and below the lower cutoff points are unaffected whereas frequencies in between are cut or boosted. A notch filter
enables you to zone in on a narrow frequency band and boost or cut at that band. Notch filters are generally used for eliminating rogue or unwanted frequencies that -20 fall in the mid range rather than at the extremities where you would use a low or high pass filter. This is great for such jobs as removing an unwanted "ringing" resonance from a snare drum or removing "muddiness" in the lower mid range from drum sounds. It's also good for creating frequency space in the mid range if you have two or more sounds with clashing or overlapping frequencies.
Peaking filters and Parametric Equalisers
With a peaking filter, the user is able to select a frequency and cut or boost the level of the sound around that frequency. A standard peaking filter (or semi-parametric EQ) will have controls for gain and frequency. A parametric EQ is a peaking filter with an additional control for bandwidth or Q. It is so-called because (dB) each of these parameters is adjustable. The bandwidth (see below) determines the spectral spread over which the EQ iseffective. As with other filter types the bandwidth is described from the point at which has been cut or boosted by 3dB from the unaffected level.
[box type="shadow"]Bandwidth or Q When boosting a frequency with a parametric EQ, the user is not just boosting a single frequency alone but also those immediately surrounding it. As with other filter types there is a slope. The gradient of this slope depends on two things, the amount of cut or boost and the width of the frequency band that is included. A wide bandwidth will affect a broader range of frequencies than a narrow one. Q and bandwidth describe the same thing but using inverse scales. Bandwidth describes the number of octaves (or fractions of an octave) that are included in the cut or boost range. Q is given as a number whereby a high Q value denotes a narrow bandwidth and a low one a large bandwidth. A Q value of 90 (approx.) means that the bandwidth is 1/60th of an octave. A Q value of 0.4 (approx) would denote a bandwidth of 3 octaves.[/box]
This is an extremely powerful type of equaliser with a huge variety of uses. A parametric EQ with a wide bandwidth can be used for general tonal shaping over the middle frequency range. Alternatively a very narrow bandwidth can be used to pick out and remove nasty resonance — performing a similar role to the notch filter. A moderate bandwidth setting could be used for emphasizing an instrument by tracking down the frequencies that give it presence and character and boosting them to taste. A narrow bandwidth, and a high boost can be used for creative special effects if the peak or resonant frequency is swept over time using automation or an LFO.
Most of us are familiar with the graphic equaliser. They are not commonly used in DAWs for mix purposes but deserve a mention since they are the type of EQ that is most commonly found in the outside world. They are widely used by hi-1 enthusiasts, in PA systems and studio control rooms and of course in cars. A graphic EQ has a number of clearly labelled frequency bands each with a dedicated fader and a detent point for zero boost or attenuation. There should be little or no interaction between the frequency bands and there is usually about 12 dBs of cut or boost. According to the ISO (International Standards Organisation) the stages (frequency bands) should be 1/3 of an octave apart. This means that when you reach the 4th stage it will be double the first.
Graphic EQs are used in PA rigs to deal with feedback and bad acoustic resonance. In control rooms they can help to correct monitoring anomalies.