Processor units
Graphic EQ - A graphic EQ unit is used to isolate certain bands of frequency of a sound or multiple sounds, filtering a tapped off section of a signal through a high or low pass filter, and then to be cut out or boosted depended on whether they feedback in a live sound setting for example, and this is done with a fader which controls an attenuator to cut or boost the signal, which then rejoins the other signals that have gone through the same process to the output. A Graphic EQ can have varying amounts of bands, the more bands there are the narrower the band of frequency each will process will be. So if an EQ has 32 bands, it will have 32 different ranges of signal to process, meaning a smaller band of frequency can be identified as being a problem, and be removed from the output sound, without taking the character of the sound being processed away. If an EQ had only 4 bands, and the two middle bands were feeding back through the front of house monitors for example, then taking both of those down in volume would mean when the sound went back out again it would have no mid range frequencies and would therefore sound terrible, but more bands allows problem frequencies to be targeted more directly with less effect on the overall sound being outputted which is better for the resulting sound quality. On the below diagram I have drawn only 4 bands, which really wouldn't adequate in the live sound setting where these are more often used, but it shows principally how it works. Each of the four bands targets a specific frequency range (eg. band 1 - 20Hz to 500Hz, band 2 - 500Hz to 2KHz, band 3 - 2KHz to 10KHz, band 4 10Khz to 18KHz) and these each have an attenuator to cut or boost the sound once you've determined what you need to do with it, and they have a 48dB per octave filter so that sounds don't overlap between frequency bands, and a frequency you've tried to eliminate occurs in another band that doesn't need altering. A less steep filter like 12dB/O may allow for sounds not to be completely cutoff as octaves relevant to frequency can cover more range than this would alleviate. When all individual signals have gone through this process in their frequency bands, they join together at the main output to the speakers.
Parametric EQ - With EQ units there are four types of filtering, which are Hi Pass, Lo Pass, Shelf and Peak. Hi pass allows signals above your threshold to pass without effect, removing the lower ones, lo pass does the opposite and then shelving which boosts signals by removing signals equally above and below the 0dB line within the threshold. All these EQ types can be implemented with parametric EQ because it allows you to control the three important parameters to make these EQ shapes, which are frequency, bandwidth and gain. The frequency lets you select how far through the spectrum you want your EQ'ing to begin at, for example a low pass filter may set the threshold at 1KHz, allowing anything below to pass and cutting above. Then there's width which is often referred to as Q value and controls how wide the band of cut you make is, because you may want to cut from 1KHz to 20KHz, or 1KHz to 2KHz, in which case you'd nead to narrow the width, so the width control allows you to do this, and works both ways to increase and decrease. Finally gain makes a peak or trough above the 0dB line using the width and frequencies you've selected, so if you put the gain to +12dB it will excentuate the sounds at the selected frequency, and -12dB will removed them from your audible sound.
As for the workings of the parametric EQ, the signal from the input is split 5 ways to Hi Shelf, Lo Shelf and three mid bands, which is where the sounds are attenuated specifically according to their frequency band, as if you're using hardware with pots, it's better to have 5 pots targetting their own band (eg. 20Hz to 200Hz) rather than trying to skip through to find the frequency like you could with a digital parametric EQ in a sequencer that may give you a grid and let you drag to any point. Then it has the possible frequency, width and gain parameters that could be used upon the signal, before going to the output. Here's a diagram to illustrate my explanation.
Variations of Parametric EQ are semi-parametric units, where an attenuator controls the width of a the pass, according to the parameters of the sound it is processing. A drawback of this maybe that if you don't like the alteration it's making to the sound then nothing can be done about it, but a plus point is that units doing this in their own unique way develop a signature style and sound making them desirable units. Pultec make EQ's in this way, and these have had software emulations made of them, proving their desirability.
Filter - A filter is a unit that cuts frequencies out of a sound to change the sounds characteristics. It cuts sounds at different frequencies and with varying severity, and can be used creatively with controls like the LFO, envelope shape, slope and resonance to alter. Firstly the signal from the input goes to the cutoff, which is the threshold or limit set at a high or low amount to allow frequencies below or above to pass through. If the filter is a high pass one, a limit for cutoff will be set and everything below will have the filter effect to it. The parameters of the cutoff are the resonance, which adds a peak or boost of signals in volume at the point of the cutoff which audibly sounds pleasant. Then the the slope is measured in dB and is the severity of the filtering from the cutoff point. A 48dB slope would take the sound away more severely than a 24dB or 12dB slope would, and this means less sound below the point gets back through to the main signal. There are finally modulation controls which are an envelope and LFO (low frequency oscillator). The envelope is what shapes the opening and closing of the filter over time, and has parameters of Attack, Decay, Sustain and Release which represent stages of the sound over time, which these parameters then define how the filter acts on these sections of the sound. A long attack means the filter won't act quickly on the sound, like a 12dB slope in a sense, but then a fast attack closes the filter quickly like a harsh slope, and then the sustain is the period in the middle where the filter can allow the sound to breathe or keep it pinned down at a lower volume, and then the release is for the filter to open or close again. The LFO is a frequency set too low to hear, but to control the opening and closing of the filter rhythmically. It could be set to a tempo on some digital versions, so it's in time with the click of a track, or set on older analog versions to operate at a frequency which could equate to a tempo but that would just open and close the filter at a rhythm that works with the sound going through it and the track playing. Below is a diagram showing all of this, and how the signal would flow between the parts of the filter.
Compression - Compression is an effect commonly used in post production to remove louder peaks of audio signal, and then when dynamics are more compact, boost the whole band to bring up the quiet and louder sounds creating a fuller sound with less dynamic variation and often having a certain characteristic added to it depending on the compressor being used. So there are three key stages in compression, from the input a threshold needs to be set to determine the highest volume a sound can be before it needs to be acted upon to bring it back below the threshold. Signals peaking the threshold are acted upon by a volume reduction stage which has parameters like knee, envelope and ratio which all define the behaviour of the volume reduction, before they rejoin the untouched quieter signals. Firstly though it's important to explain how the knee, ratio and enevelope work. The knee is a setting that controls how harsh the compressor acts as the signal excedes the threshold. A knee setting of zero means the signal goes from unaffected to being immediately compressed at the set rate without being eased into the compression, which may sound slightly harsh. A knee setting increase will smooth the peak of the line entering the threshold, rounding it off and making a more subtle transition into compressing the signal. This isn't the most important control, and it's setting isn't always the most obvious, but it's useful to really refine the effect of the compressor. Secondly, the ratio is set at a 1:(x) and is what defines how severe the sound above the threshold is flattened to return beneath the threshold again, which isn't massively different to knee, except knee is for entering the threshold rather than once it's exceeded, and the ratio can be set to be far more obvious. By this I mean a ratio of 1:2 will only half the volume above the threshold from the volume the peak originally reached, which may not have much effect, however double that to a ratio of 1:4 will be double the reduction of 1:2 making the reduction more severe, and this can go on theoretically to a ratio of 1: infinity which would be such a severe reduction that it would be as if the signal never exceeded the threshold and would therefore sound too severe, so a happy medium needs to be found, which depends what instrument you're compressing. Compression on vocals would probably suit a quite severe ratio as the dynamics of human vocal can vary between low notes and high notes, meaning high notes will peak far higher than others, and a severe ratio will act heavily upon the high notes, and bring them neatly in line with the others, for a more streamline vocal sound. Finally, the envelope involves the attack and release settings of the reduction, which work alongside ratio as they define how quickly the reduction comes into action after a sound exceeds the threshold. A fast attack means that as soon as the signal exceeds the threshold, the reduction (defined by the ratio) will immediately act upon the signal to reduce it quickly, and the release then is what sends it back to beneath the theshold in terms of speed of release of the reduction once it's approaching the threshold going back down. All these parameters combined, are how you shape the sound and effect a compressor has on a signal, allowing for creative manipulation of sounds, and simultaneously ensuring no sound is massively jumping out during mixing, though this isn't the primary reason for using compression, and automation would come into play for problems of that nature. Finally as the processed signal meets the unprocessed signal, the result is that the low levels are still low, but the higher levels have no been reduced, so altogether the signal goes to a gain section, which boosts the overall loudness up depending on how much is needed, which is defined by our perception of how much of an extra boost is needed. This means that the low signals now sound louder than they did to begin with as the noise floor has been lifted, but also the higher volumes have been lifted to which means dynamics are preserved, and the louder sounds are still nice sounding.
Below is a diagram showing the signal input being split to sounds above and below the threshold and then those above having those effects put upon it before returning to the untouched signals, both going through the gain and then to the output.
Limiter - Limiting is a very similar process to compressing, as it essentially does what the 1:infinity ratio compressor would do, and stops the sound at the threshold from exceeding it at all, so creates a barrier at a set volume that no signal can exceed. This justifies the term "brick wall limiting" often given to the process. It differs from a compressor though as the gain stage is at the beginning of the process as before the gain is added, the signal may not be getting to the threshold at all, making the affect inaudible, but once the gain is added and the volume is going above where your limiting will be added, then the threshold can be added then. As with a compressor, when the the signal that would have peaked the threshold is reduced, they go back together, but they go straight to output in limiting as the gain was added at the start. Below is a diagram to show this limiting.
Noise Gate - A noise gate is a unit that stops sound below a set volume threshold to open a gate, and for it to be audible, and is put to use in live sound and is an alternative to strip silence in mixing multitrack recordings. It works by having a threshold set, according to how loud we expect the instrument it's being used on to be, therefore meaning that only a noise as loud as the instrument produces would open the gate and allow what the microphone has picked up to be heard. Say for example with a kick drum, every time we hear the kick drum play we get a volume spike of 6dB, but the spill from the other drums produced a volume change of 3dB, we would set the threshold of the noise gate at 5 or 6dB, which would guarantee that the gate would only be opened by the kick drum itself, and no spill or background noise would be able to open that gate. This would help to us to isolate each instrument, and not have bleed muddying the overall drum sound in-between hits of the individual drums which would give us an overall cleaner sound.
The way a noise gate works is that the signal goes straight from the input to a threshold set at at a certain level as I just explained. It has an additional side chain control, which the threshold controls the amount of, for example the side chain could be used on a synth sound and set to receive signal from a percussive instrumental, which would rhythmically control the gate acting upon the synth, causing a sort of tremolo effect. Then from the threshold it's split to above or below the threshold. Below nothing happens to the gate and therefore the signal doesn't go to the output, but above the threshold it goes onto the set attack and release settings, which control the speed of the opening and closing of the gate, which changes the perceived harshness of the effect. Then the signal is outputted, and here's a diagram to show this.
Delay - A delay unit creates duplicate versions of the sound input, and staggers the release of the duplicated sound until rhythmical intervals after the original is released. There are parameters to control the speed, intensity and rhythm of these repeated sounds, meaning the delay has many uses, and could be used creatively on any instrument or sound, as it changes the characteristics of it so much that it can sound like a completely different sound altogether, meaning it has use in all sorts of musical genres. The way analog delay works is that inside the unit are reels of tape, with play heads that read any sound recorded onto them. The playhead of the second tape is set to be behind the first, so when the two reach the output they are staggered. Also however, as this would only create the one repetition, there's a deliberate feedback loop where the signal travels back to through the circuitry and out again multiple times, according to the intensity of the feedback setting which creates the looping, and increases the speed as the repetitions begin to create repetitions of the initial repetition, until it all amounts to a massive feedback loop. The dry signal is never effected, so the instruments dry signal goes straight to the output. When a signal is sent back to the input to be repeated again this is referred to as a tapped signal, of which there can be three or more, to create the repetitions. Creative controls on modern delay unit plug-ins are things like a groove setting, which uses musical notation to represent where the repeats will be positioned in musical terms, so a quaver would repeat a quarter of the way through a bar, or on the off beat which would create an interesting groove to the delay. Also there are high and low cuts, like a filter for the delays, and dry/wet blends which control the difference in volume between the original dry signal and the effected delay along with it. So if you took the dry down completely, and the wet all the way up you would hear only the effect of the delay on the sound, and none of the original sound that was inputted. Here's a diagram showing the delay setup and tapping of signals. In the diagram, related to the old tape method of creating delays, the taps represent different play heads taking the signal from the tape and sending it to the input to create the loop, then eventually reaching the output depending on the settings of feedback etc.
A classic example of a tape delay unit is the Roland Space echo, used by Jamaican producers in studio production to create the Genre Dub Reggae, with the delay applied to snare drums and guitar stabs creating the spaced out instrumental sounds which have become synonymous with the genre.
Phaser & Flanger - These two effects are similar to each other in how they work so can be explained together, and create their effect in a way not massively dissimilar to delay, with a few different stages along the way. Like delay, the dry signal from the input immediately goes to the output, so a mix of dry and wet is created when they reconvene at the output, and a second signal is tapped off to be effected. This signal in both of these units has a delay controlled by an LFO applied to it, set at a shorter time than delay so to not sound like a delay, but in Phaser the sound goes through different band pass filters which effect different frequencies across the sound at different speeds, meaning that coming out the other end the sound has some sharper increases in applied effect and some decreases too, meaning that the effect doesn't move equally to the original wave, and over the duration of time the note is played, creating the wave, the phase relationship of the dry and wet signal is not the same, creating the whooshing sound. However with flanger, this set of filters isn't present, so instead the LFO delaying the tapped signal makes this essentially a very fast delay of one tap rather than multiple, but as there is a slight delay, this again creates some changes in phase relationship which makes the wobbly sort of sound of the flanger. The signal flow of both is very similar, but with the differences in stages along the path of the wet signal creating the different effects.
Flanger diagram with LFO only filters:
Phaser diagram with filters and LFO:
Tremolo - A tremolo is effect is a rhythmical reduction in an otherwise constant sound like strumming guitar chords or a synth drone, which gives the sound a pulsing effect. The effect can work in mono or stereo, in mono you hear the same signal left and right ducking in and out at the same time, meaning that it's in both speakers at a peak, and not at all at a trough, but in stereo, 180degrees out of phase, it pans from Left to Right, so the sound never completely goes away.
To begin with the signal is split into two from the input, to create the stereo sound if it's isn't already a stereo file, and then one of them is modulated by an LFO, the rate of which dictates the speed of the dipping or oscillation if you will, and the other signal is phase inverted 180degrees so that it doesn't mirror the other, but does the opposite creating this effect of bouncing from one speaker the other without panning or joining in the middle at any point. These signals then meet up again at the output. If the signals didn't have one inverted 180degrees, they would give the mono effect I mentioned, which you hear from some effects pedals, or in built tremolo effects like on guitar amplifiers that just modulate the one sound in and out. On tremolo plug-ins like the one in Logic you can change the phase to be different degrees, like 90 for example which will make the sound effect stagger, as it won't appear as intensely in either speaker which will sound interesting. Also you can control the depth of the signal, meaning when the LFO dips the whole signal isn't removed and therefore the effect is more subtle. Here's a diagram showing this effects.
