Tutorial 1: Compression and limiting

Tutorial 1: Compression and limiting

A compressor is a device to automatically control the dynamic range of a signal, with a limiter being essentially a compressor with an infinite ratio. In their simplest terms these are devices which reduce the volume of the signal passing through them according to that signal and the chosen settings of the device.

The compressor in the Duende channel has not only the legendary SSL sound which has been sought after by DAW users for years, but many functions making it extremely versatile and easy to use.


The Theory

Refer to the diagram above; the red line shows what happens to the signal level when it is passed through a compressor. As you can see when it hits the threshold, the output no longer equals the input; it is lower.

If we pick an arbitrary point ‘A’, where Ai is an input level of say -20db. Ao is the output level of the compressor for that input level. Here, Ai is equal to Ao because the compressor is not changing the gain of the signal as the input level is below the threshold. If we now pick another point of a higher level, ‘B’, you can see that the output level Bo, is lower than the input level Bi. This is because the compressor is reducing the gain as the signal is above the threshold point.

Soft/Hard Knee

The channel compressor in its default setting is of the soft-knee type which means that gain reduction won't just suddenly start happening at the threshold. Instead this point is rounded off so just below the threshold some gain reduction will take place and increase up to the threshold point where the reduction will be according to the ratio. This would be shown in the above diagram as a curve on the red line, instead of the abrupt change in angle. A soft knee compressor will generally sound more gentle and musical, where as in hard knee mode a more aggressive sound can be achieved.

Make Up Gain

Because a compressor lowers the level of the signal passing through it, it is necessary to 'make up' for this attenuation at the output. Many compressors will have a make-up gain control to manually compensate for the gain reduction carried out by the compressor. The Duende channel however does this make-up automatically, leaving the user with one less control to think about. As you vary the amount of compression applied to a signal the output level will stay roughly constant, which means in use the compressor is easier to operate as you don't need to be constantly adjusting the make up gain.

Time Constants

The Duende Channel compressor features a switchable attack time and constantly variable release time. The attack time is the time taken for the compressor to carry its gain reduction when it is presented with a level above it's threshold. The release time is the time taken for the gain to return to normal once the signal has dropped below the threshold.
Bear in mind that due to the speed of the attack and release on the Duende channel that it is possible to distort the audio when these are set too fast with large amounts of gain reduction. In this scenario it is generally worth increasing the release time slightly, decreasing the ratio or using the long attack setting - unless the distortion is a desirable effect!


Leveling out vocals

  • Solo the vocal track and open the corresponding Duende channel.
  • Play the track through with the Channel in BYPASS ALL. Notice how some of the words are higher in level than others. This may not seem a problem when the vocal is in solo because we can easily hear all of what is sung. However, when the vocal is against music and backing vocals for instance, controlling the level can really help the vocal sit in the mix and ensure all of it is heard over the music.
  • Now press the BYPASS ALL button to bring the processing into place. Notice that the vocals now sound more even and controlled. The compressor is bringing down the level when the signal is above the threshold, and leaving it alone when it is below. The longer attack setting has been used to allow the initial transients through.
  • Try pressing the FAST ATT button and notice how more gain reduction happens and the vocal tends to sink back slightly. The release is set to 300ms. This is short enough to catch the peaks and long enough to avoid pumping. The threshold is at -4db and the ratio at 7:1. These make the compressor apply a small amount of gain reduction on the quieter words and about 6db on the louder parts.

Changing the character of drums

Load tutorial 1, solo the room mic track and open the corresponding Duende channel.

Listen to the dynamics in the room track, noting the difference in volume between the low level and high level content. Now press the BYPASS ALL button to kick in the compressor and notice immediately the increase in energy, loudness and punch!


  • You can see that the threshold is set to about -12db, which means any part of the signal above this level will get reduced in gain by the amount set by the ratio control. Any part of the signal below this threshold will be left untouched.
  • The ratio is set at 3:1 which means anything above the threshold will get reduced in volume by a factor 3. For example, if the signal is 12db above the threshold it will be reduced to 4db. Now increase the ratio and you can see the yellow gain reduction meter going into the red indicating the compressor is working harder as it is reducing the gain of the signal by increasingly large amounts.
  • The FAST ATT button is off so the compressor lets through the initial attack of the drums before working. If the FAST ATT button is now depressed the compressor can be heard starting to work almost straight away at the beginning of the transient, which kills the beginning of the hits somewhat.
  • The release is set to its fastest setting meaning the compressor brings the signal back up to its normal level very quickly and is ready to catch the next transient. Try increasing the release time and you can hear that the compressor doesn't cause the signal to jump up in level between transients and a smoother sound is achieved. But if the release time is set too long the compressor does not recover quickly enough to catch the next transient and the process does not work properly. A longer release will generally give a smoother sound, where as a short release will give more intensity and energy to a signal.
  • The PK (peak) button puts the compressor into a hard-knee mode, whilst also forcing the compressor to react with more attention paid to the peak level rather than average level (rms).


Limiting can be achieved on the Duende channel by setting the ratio to 8 and depressing the PK button. Do this and you can hear that anything above the threshold gets reduced in gain to a fixed level. This can be useful for making a signal generally louder, but the transients are sacrificed. Try bringing the threshold down with a high ratio and fast attack and notice how that all important information at the beginning of a drum hit is compromised.

If too much compression or limiting is used it can make the sound lifeless and dull as it kills the transients. It’s a process which should be applied with discretion, and beware of the trade-off between loudness and transients.


Dynamic Control

In the DAW project for tutorial one, the purple kick, snare, hats and toms-oh tracks have all been dynamically altered to simulate what it might sound like if the material was badly recorded. A signal may appear something like this if the compression has been set wrong on the input, the drummer did not play the kit with even and controlled dynamics or the microphone stands were not steady. In these cases the Duende channel compressor can be used to compensate for these shortcomings.

Mute the room track for the time being and solo each of these four tracks in turn. Listen to how uneven the levels are and the way the drums tend to jump up or down in volume.

Let's try and level out the kick track first.

Solo the kick track and open the corresponding Duende channel. Use the BYPASS ALL button to place the compressor in circuit and notice how much more even the kick drum sounds.

  • The attack has been set fast to catch the peaks as they are quite apparent and the PK button is depressed so the compressor responds to the peaks rather that the average signal level.
  • A ratio of 1:8 is used as there is quite a large swing in the dynamics so a fair amount of gain reduction is required to even out the peaks.
  • The release is set to 0.2s so the compressor recovers quickly enough to catch the next peak yet doesn't come back too quickly to cause a pumping effect.
  • Finally the threshold at 2.5db is low enough to catch the peaks yet leave the lower end of the dynamic range alone.
  • A small amount of expansion has been used to reduce the lower level content (bleed, ambience etc.) which has been brought up as a side effect of the compression.

Now take the kick track out of solo and solo the snare.

Again by switching the dynamics section in circuit a more controlled signal is heard. Similar processing to the kick track has been applied to the snare, with a fast attack to catch the peaks but this time with peak mode disengaged as a smoother dynamic is produced. Again the expander has been utilised to reduce the hi-hats between snare hits.

Let’s look at the hi-hat which provides a slightly different challenge as it contains mainly high frequency content with an attack portion which must be treated with care. The snare bleed present in this track is the over-riding energy, i.e. when the snare hits the signal is at its highest level. This makes it difficult to control the level of the hats as the snare will govern the gain reduction.

  • A medium ratio and fast attack have been used to catch the peaks but as the average level is just as important in this signal the PK button is up.
  • Quite a long release time is used so the hats don't jump around too much and sound dynamically even without the compressor getting too eager.
  • Note that the high-pass filter is engaged to cut everything below 500Hz which serves to reduce the amount of snare in the signal.

Now solo the tom-oh channel and listen to both the uncompressed and compressed signal using the BYPASS ALL button.

This track presents a similar problem to the hi-hat track as it contains not only hats but snare, kick drum and some ambiance. You will notice two Duende channels inserted over this track, and if you look at them you can see why.

The object of this exercise is to even out the level of the tom-oh track, whilst not bringing the level of the ambient portion up too much.

  • The first is set up with a low threshold and low ratio, making it compress nearly the whole signal by a small amount. This serves to generally even out the dynamics.
  • Placed after this is another channel set up in a limiter configuration with fast attack and infinite ratio. Looking at the gain reduction meter shows us that it is catching only the highest peaks and reducing these by about 6db.

Finally all these tracks have been sent to a bus and compressed lightly using the Duende bus compressor. This serves as a final 'polish' to bring the together the drums and add some punch. The use of the Bus Comp is covered in tutorial 5.

By A/Bing the two bounces at the bottom of the screen the difference in dynamic control is quite apparent.

See the tutorial on side-chaining for ways in which the EQ section can interact with the dynamics for even more control.

Tutorial 2: Using the Gate/Expander section

Tutorial 2: Using the Gate/Expander section

A gate can be thought of as a processor which silences (or significantly reduces) the audio passing through it if is below the set threshold. This can be useful for removing unwanted noise present in an audio signal, so the gate will close if the wanted signal is not present. For this to work correctly there must be a reasonable signal to noise ratio. Gates have other uses such as controlling the decay of drums, which we will look at in this tutorial.

An expander is similar, but is essentially a gate with a soft knee. An expander can be thought of as the opposite of a compressor, reducing the part of the signal which falls below the threshold by a set amount. The ratio of the Duende channel expander is fixed at 1:2, and the gate is at 1:20. Where as a gate will open and close almost exactly at the set threshold and reduce the signal to almost silence, an expander will reduce the gain according to the input level; the lower the level, the more gain reduction happens.

Load Tutorial 2 into your DAW; here you will see 5 audio tracks of a drum recording. Because we have the drums recorded separately, we can zoom in on certain aspects of the individual tracks to control their dynamics using the gate/expander section of the Duende Channel.



Solo the kick track and open the Duende Channel placed over it.

Play the track and you will hear that the kick track contains a significant amount of the snare bleeding onto it. This is a consequence of modern mic’ing techniques and is quite normal. However, it may not be desirable to have the snare present on the kick track, as it narrows the options available for treating the kick drum as this would influence the snare as well.

Now press the BYPASS ALL button to turn the gate on. Firstly you will hear that the snare drum has virtually disappeared. You can see that the green gain reduction meter is lighting up indicating the gate is significantly reducing the gain when the kick drum is not being hit.

Now look at the gate controls:

  • The gate/expander button is set to GATE mode as the EXP button is not on.
  • The threshold is the level at which the gate will open and close. The threshold has been set to +4db, meaning that the gate will remain closed if it receives any audio below this level, and open if the input level is above this.
  • The hold time is the period for which the gate will remain open after it has received a signal over the threshold.
  • The release is the ramp time taken for the gate to return to its closed status after the hold time has been completed.
  • The range is the amount by which the gate will reduce the gain in its closed state, from 0-40db.
  • The FAST ATT button selects how long it takes the gate to open once it has received a signal over the threshold. With the button down the gate/expander will open almost instantaneously, and up it will take a small amount of time to react.

Now try moving the threshold control up and note that if this is set too high, nothing will get through the gate and the audio is almost silenced. If the threshold is set too low then everything will get through the gate and it won't function.

Turn the range control to 0db and note that no gating takes place. This is because in its closed state the gate will reduce the gain by 0db.

If the hold is set too long it will let more than one beat through, and as you slowly increase this from 0-4 seconds you can hear the snare being let through.

The release time has a similar effect but is a ramp up to full rather than a held level.

If the FAST ATT button is switched out, the first part of the kick drum is missed as the gate is not opening fast enough to catch the initial attack. This is generally best left on FAST ATT for transient signals such as drums. The hi-hat track has had a similar gate treatment to make the notes slightly shorter therefore giving a tighter sound. As you can see and hear a gate is a very powerful tool for separating drum signals, but also gives you some control over the length of the individual hits and their attack and decay characteristics.



Now lets look at the expander mode and its use for reducing (not silencing) the low level aspects of a snare drum track, whilst leaving the high level parts untouched.

Solo the snare track and open the Duende Channel plug-in.

You can hear that the snare track contains a fair amount of hi-hat. It may desirable to reduce the level of the hi-hat as the snare can then be treated more independently, which in this case means compressing it and adding some reverb.

Now press the BYPASS ALL button to place the expander and compressor in circuit. Looking at the green gain reduction meter it can be seen that similar to the gate on the kick drum the expander is reducing the gain in between snare hits, but in a more subtle manner. If you now move the range control on the expander to 0db so it is no longer working, the hi-hats and ambiance present on the snare track is brought up in level, partly due to the use of the compressor, which is used to give the snare some extra punch. If the range control is now turned back to about 20db, a cleaner, tighter sound is heard.

Take everything out of solo and hear that the kick and snare have more definition within the drum mix, and the kit generally sounds tighter and cleaner. Further down the mixing process this may bode well as there is now more 'space' in the drum mix.
Often if you wish to use large amounts of compression on the signal, it can be a good idea to use a gate or expander before hand to compensate for the fact that the compressor brings up the low level information. Using a gate/expander and compressor in this series configuration gives the user a great degree of control over both the low level and high level information present with in an audio signal.

  • Reducing unwanted noise, i.e. guitar amp hum
  • Reducing ambient noise from bad recording environments
  • Separation of drum recordings
  • Control of attack and decay on drum recordings
  • Reduction of headphone bleed on vocals, acoustic guitar and other instruments
  • Increasing the dynamic range of a signal
Tutorial 3: Basic EQ

Tutorial 3: Basic EQ

EQ in its most basic form is a tone control, used to change the frequency balance of a signal. We have all encountered this in the home hi-fi featuring either hi, mid and low tone controls or as a graphic equalizer. As you move the respective controls the sound is changed in such a way as to give more treble, less bass etc. By doing this you can get the music sounding to your taste, and with a well featured EQ can solve problems caused by a difficult listening environment or badly recorded material.


The Theory

The Duende Channel EQ is predictably rather more complex and powerful than those found on a hi-fi, yet works on the same principal. It is a four band EQ, meaning four sets of different frequencies can be attenuated or boosted at the same time and the Q (or bandwidth) is variable on the middle two bands. The high and low bands are switchable from bell to shelf independently, and to top it off hi and low pass filters are also simultaneously available. In addition to the four band EQ, two filters are simultaneously available.

The diagrams below attempt to clarify the difference between shelf, bell and filter shapes.


As can be seen a filter cuts out the frequencies above or below its cut-off point. A lowpass filter will let through only the low frequencies therefore cutting out the highs, where a high-pass filter will cut the low frequencies. These are useful when unwanted frequencies are present at either the high or low end of the spectrum. An example of this would be cutting out the stage rumble picked up on microphones in a live situation.



With the BELL button up on the LF band of the Duende Channel a shelf EQ will be selected which will boost frequencies from the cut-off point downwards, or upwards in the case of the HF band. The shelf shape gives a gentle roll off, as opposed to the more heavy handed cut action of a filter.


If the BELL button is down then instead of the shelf shape, a curve is selected which will only boost frequencies around the cut-off point with a gradually reducing gradient towards the zero gain point. The HMF and LMF bands have one more control alongside the gain and frequency. The Q, also referred to as bandwidth, is the width of the filter and governs how much of the frequencies either side of the center frequency are boosted or attenuated by the EQ. This type of EQ is referred to as fully parametric. A smaller Q value will give a wider curve, therefore not only affecting the cut-off frequency but many others either side. This is often said to be a more musical way to EQ, and can be good for general sound shaping when the tonal balance of a signal needs to be changed.
With a large Q value, the curve is narrowed allowing you to zoom in on frequencies more precisely. This can be useful for attenuating problem frequencies within a signal, but large gain changes can sound unmusical.


The Practice

Now we are going to use the channel EQ over various drum tracks to clean them up and increase the separation between them.

Load the appropriate DAW project and let’s have a look first at the individual channel EQs.

Solo the kick track and open the corresponding Duende channel. The kick drum is not sounding so good with the channel bypassed. There are some frequencies in the middle which tend to mask the lower and higher end of the kick drum. Instead of adding low and high EQ, it can be preferable to pinpoint the nasty middle frequencies and tone them down.

Now press the BYPASS ALL button to bring the processing into place.

  • Using the LMF band an 8db cut has been made and by experimentation the offending frequencies have been found to be around 600Hz. The Q has been decreased so a more musical result is achieved.
  • A small boost at 3.6kHz has been used to give the kick slightly more click, emphasizing the range of frequencies caused by the beater hitting the skin.
  • Also a small amount of bottom end at 64Hz is added to give the kick slightly more weight.
  • The filter section has not been used as we need all the low frequencies of the kick drum to be present and the high end is important to retain the attack information.

Now solo and open the Duende channel over the snare drum. Listen to the unprocessed signal and then press the BYPASS ALL button to enable the channel processing.

  • The high pass filter has been placed at about 90Hz to remove the unwanted frequencies, as there is little of interest happening down in this area in the snare drum.
  • To compensate for this low end cut the LF band is used in shelf mode to boost some of the frequencies above this cut so the snare still has some weight.
  • Again there is some nasty middle present, this time by sweeping the LMF with an 8db cut and high Q value it has been found to about 285Hz. The high Q has been used so that the frequencies either side of 285Hz are not attenuated too much as there is a significant amount of wanted information around this area.
  • To give the snare a bit more snap the LMF band boots about 2kHz with a low Q value.

Solo the hi-hat and again listen to the processed and unprocessed signals.

  • Here you can see the high-pass filter has been wound up to 500Hz to cut out the kick and some snare drum in the hi-hat track as this is not wanted. Because the hihat is nearly all high frequency information this can be done.
  • Pulling 6db out of the signal using the LMF band at a frequency of 540Hz significantly reduces the presence of the snare in the hi-hat track.
  • Some of the harsher frequencies have been attenuated using the HMF band with a slight cut at 7kHz.
  • Then to compensate for this a slight amount of air has been added using the HF band in shelf mode at 17kHz.

By toggling the BYPASS ALL button in and out the difference in clarity is quite noticeable in the EQed signal.

Finally solo the toms-oh track and open the appropriate Duende channel. Again the middle frequencies are quite muddy, and some work should be done to clear these up.

  • The high-pass filter has been used at about 60Hz to clean up the bottom end and a slight cut at 180Hz using the LF band in BELL mode reduces some of the boomyness in this area.
  • Using the LMF band a small cut has been made at 430Hz to reduce the boxyness of the snare drum.

The Proof

Now solo the processed track and A/B with the unprocessed track. By carefully using filters, EQ and a small high end boost in places the overall clarity and openness of the drums has been improved.

Please bear in mind that all this EQing has been done with the channels in solo. This is often not the recommended way to mix, as to help a signal to sit in the mix you need to hear everything else that is going on around it. For these purposes however, the solo method has been used to help clarify what is happening with the EQ as otherwise it can be difficult to hear.

Tutorial 4: Using the Duende Channel side-chain

Tutorial 4: Using the Duende Channel side-chain to alter the response of the compressor

Sometimes it is desirable to hear the sound of a quality compressor making a drum kit pump. However, often a bigger, more even sound can be achieved by processing the side-chain of a compressor with an EQ. By doing this the compressor becomes much more versatile and the way it reduces the gain of the signal can be controlled. This is known as frequency conscious compression and many other effects are possible with this technique.

Every compressor has a side-chain circuit even though the user is often not aware of it. The side-chain generates a control signal that dictates, in conjunction with the compressor controls, how the gain reduction circuit behaves. Often the side-chain signal is identical to the input signal. The information used to control the gain reduction, is identical to the audio input to the compressor.

Low frequency sounds such as a kick drum have much more energy than high frequency sounds such as cymbals. This results in the low frequency elements of a signal dictating the behavior of the compressor and can result in the level of the signal being dragged down when a low frequencies are present.

How to

When the Duende Channel is first loaded, it's compressor is of the traditional type where the side-chain signal is identical to the input signal. Due to the flexibility of the Duende Channel, the compressor behavior can be altered quite considerably by assigning either the filter, EQ or both to the dynamics side-chain.

This can be done easily on the Channel by depressing the DYN S/C buttons found in both the Filter section and the EQ section. Depressing this will move the respective section into the side-chain circuit and remove it from the audio circuit. This means that if you have an EQ setting dialed in, and the DYN S/C button is subsequently depressed the EQ will no longer process the audio signal, but will become a control signal that is not present at the audio output.

To listen to the side-chain signal, press the S/C LISTEN button found in the OUTPUT section. This enables you to audition exactly what signal is controlling the dynamics. When any section is assigned to the side-chain, the Process Order graphic will reflect this to let you know how the routing of the channel is configured.

Process order graphic showing EQ and Filter assigned to Dynamics side-chain


No side-chain processing

Solo track 1 of tutorial 4, and you can hear the kick drum causing the hi-hat to reduce in volume when large amounts of compression are used. The snare drum has an influence too but to a lesser extent because it contains less low frequency energy. The hi-hat can be heard jumping up in between kick and snare hits, because as there is very little energy present in the side-chain at these moments the gain reduction is minimal.

High pass filtered side-chain

Solo track 2

Here the filter section is routed to the side-chain, and the hi-pass filter is moved to 500Hz to reduce the influence of the kick drum on the gain reduction circuit. Pressing the S/C LISTEN button enables you to hear this.

The sound of the drum kit is now more open, and the kick is let through more as the compressor acts less on these hits.


EQ and Filters to side-chain

Track 3 takes this a step further.

Firstly the EQ is turned on and assigned to the side-chain. Then a large EQ cut at 700Hz is dialed in which reduces the snare frequencies from the side-chain signal. Again, depressing the S/C LISTEN button will let you hear that the presence of the snare drum in the side-chain is considerably less. The kick drum is also reduced more as it has some energy in the 700Hz area.

The drum kit now has a more open, even sound where the kick and snare drum don't kill the rest of the kit and it's ambiance. Note that the output level is roughly the same as it was when nothing was assigned to the side-chain, yet the perceived loudness of the kit is considerably more.

This technique can be used on any signal where you wish to change how the compressor reacts to various frequencies that are presented to it. By assigning the EQ or filters to the dynamics side-chain, they are of course assigned to side-chain of the expander/gate, allowing frequency conscious gating and expansion to be carried out.

Tutorial 5: Duende Bus Compressor

Tutorial 5: Duende Bus Compressor

The Duende Bus Compressor can be used on sub-groups, master faders and even channels and is generally employed to compress groups of instruments or a whole mix.

When the Bus Comp is inserted over the master fader in Pro Tools it appears post-fader in the signal path. What this means in use is that if you move the master fader level (i.e. to perform a fade out at the end of a song) then the level into the Bus Comp is changed. So if the master fader level is lowered the input to the Bus Comp is also lowered, so by the end if the song there will be no compression taking place as the input would be well below the threshold. In this scenario it may be preferable to use the Pro Tools trim plugin placed after the Bus Comp to perform the fade out.

In Cubase, Nuendo and Logic the insert is pre-fader so the level of the fader does not affect the input level into the Bus Comp.



A bus is referred to in different ways in the various DAWs. Traditionally a bus is a path over which a signal travels in order for to reach a certain destination. This destination is a channel which can accept various inputs from various buses and is referred to as a sub-group. Pro Tools and Digital Performer use this definition and the signal will be sent to an Aux Input or Aux Track via a specified bus. In this case the Aux Input or Track is equivalent to what we refer to as a sub-group. Logic refers to a bus as both the path over which the signals travel and the destination sub-group. Cubase and Nuendo handle this in a similar way but refers to them as groups.
Refer to the diagrams below which attempt to make this clearer!

So you’ve got that – a bus comp is not placed over a bus, but in fact a sub-group!

The Bus Comp found in SSL consoles has been most famously used over the mix output of the desk, and it has been said that strapping the Bus Comp over your mix is the final ingredient to make it sound like a record. Try placing a Bus Comp over main output quite early on in the mixing process, and using about 1-3db of gain reduction.

Bus compressors are generally used over signals which exhibit a variation in frequencies and transients, i.e. a drum kit, backing vocals or any group of signals.

It is often desirable when mixing a drum kit to send all the tracks to a bus before they reach the mix output. A compressor can be inserted over this bus, which in the case of the Duende Bus Comp provides a way of bringing all the separate drum tracks together so they all start to sound like a whole kit again. It also gives an overall control over the dynamics of the drum kit and lets you govern the interaction between the tracks.


Open the appropriate tutorial 5 DAW file;

You can see that the 3 mono and 2 stereo drum tracks have been assigned to a subgroup. Inserted over this sub-group is a Duende Bus Comp.

Play the track and open the Bus Comp.

Listen to the drums with the COMP IN button switched out to get used to the dynamic in them. Now hit the COMP IN button and notice how both the density and punch of the kit is increased.

  • The THRESHOLD is set at such a point so as to control the higher level elements of the drum track. By decreasing this threshold, more compression takes place which can be clearly heard and is reflected in the VU meter.
  • The MAKE UP GAIN is set at the point where the apparent loudness of the kit is about the same whether or not the COMP IN button is down. As mentioned in the compression and expansion tutorial this is a good way to A/B as louder signals generally sound better to us humans.
  • The RATIO is at the lowest setting of 2:1. At this ratio the Duende Bus Comp features an extremely soft knee characteristic, meaning even the very low level portions of the signal are affected by the gain reduction circuitry, but less so than the high level parts. At 4:1 and 20:1 the knee is harder and behavior is similar to a regular compressor.
  • A long ATTACK time of 10ms is used. This allows the initial attack of the drums to come through before the compressor kicks in. Gradually lower the attack time and notice how the kit starts to sound squashed as the compressor reacts more quickly. The main elements of the kit (kick and snare) become increasingly compressed but the rest of it (hats, crash and ambiance) stay about the same. The attack time for percussive signals is often set to be quite long to allow these transients to ‘breathe’.
  • The RELEASE is set to 600ms. Use fast release times to make the compressor ‘pump’, or longer times to provide a smoother response.

The release time is a control which can be set according to the tempo of the track. A track at 140bpm will require the release to be set faster than one at 90bpm. At 140bpm the compressor must recover more quickly, and the release should be set to it’s fastest setting without causing the pumping that is generally avoided.

See the diagrams below where the thick black lines show the gain reduction:


With the release set too slow, you can see that the compressor does not have chance to recover before it reduces the gain on the next peak. Because the compressor does not have time to bring the audio back up to full level, the level of the next transient is compromised. This will results in the first beat being at a high level and the following ones being lower. It also means the average signal level is lowered because there is always some gain reduction happening.

The second diagram shows a shorter release time which gives the compressor time to get back to its no gain reduction state. This release time will probably sound most natural (which may not be what you are after!), and will yield the highest signal level without a pumping effect being heard.

The Bus Comp can easily be made to produce a pumping by setting a fast release time, 0.1 or 0.3s. This can be heard quite clearly over the drum kit by pushing the ratio up to 20 and bringing the threshold down. A denser, more exciting sound can be achieved like this and can be a very desirable effect. Notice how it sounds like the drummer is hitting the kit twice as hard and he’s put a few pounds on!

Tutorial 6: Advanced EQ

Tutorial 6: Advanced EQ

Note to Pro Tools LE and M-Powered users:

Please set the H/W Buffer Size to 512 Samples. This setting can be found in setup>playback engine.

We all know the situation when you work for an hour on a particular sound, only to find when you take your mixer out of solo you can hardly hear the fruits of your labour against the rest of the mix.

This is generally down to the fact that there is not 'space' for that musical element within the mix. Other elements are present which occupy similar frequencies, therefore clash with each other. This is one factor which can often differentiate a professional mix from a demo.

The most obvious way to create space is to put all the elements at a sensible volume and pan settings so they stay out of each others way. Reverbs and careful use of dynamics are tools which can also contribute to placing an element in it’s own space, but EQ is often the first port of call when trying to achieve this. The more complex the mix, the more parts are fighting for room. This means that giving each part it’s own space is more important than ever and may take a while to achieve. If a song has just an acoustic guitar and vocal then finding space for them should not be an issue, where as a song using 100 tracks of audio would generally require a fair bit of ‘space management’.

Take for instance a recording of a hi-hat captured whilst the whole kit is being played. We are generally only interested in the higher frequencies present in the hi-hat track as the rest of the spectrum is other drums and unmusical content.

In this case it is worth employing the high-pass filter on the Duende channel to remove frequencies below 500Hz as we are not interested in these. Also when drums are close mic’ed there is usually more low frequency content present than would normally be heard if you were to listen to the kit in the room. This is due to the proximity effect which occurs when a microphone is placed close to the source, and can be compensated for by either filtering or rolling off the low end.

It is also worth bearing in mind that bass frequencies take up far more energy than high frequencies for the same perceived volume. So cutting low frequency energy out of signals which are only interesting in the highs can be a good practice, as it essentially means your mix can be clearer and louder because there is less low frequency energy present that takes up valuable headroom. But be careful not to just do this as a habit, and make an informed judgment about where the filter should be set based on what energy is present and what is wanted in the low end.

The same can be said for using the low-pass filter on sounds like bass instruments which are predominantly low in frequency. This is not so important from an energy point of view as this high frequency information contains comparatively little energy, but can help to clear up the top end of your mix.

Using the high and low pass filters in this manner is a quick an easy way to create some space in your mix. This can also be done by using the channel Eq high and low bands for a less heavy handed approach. Using the high and low band EQs in shelf mode means that the frequencies are getting attenuated rather that cut out all together. This can provide a more subtle result yet still prove effective.

Using the Eq section of the channel to control the middle elements of a mix (where the music is!) can be much more involved yet once it is done well can produce great results.


Using EQ to create space in a mix

Load the Tutorial 6 session.

Play the track and the notice how some of the parts are getting on top of each other. Try soloing different tracks together and notice the interaction between them. For instance, the rhy gtr 1, rhy gtr 2 and bass tracks have very little definition when played together. The drums sound slightly muffled as there are some middle and bottom end issues that need clearing up. Lets start with the drums and bass as they are the foundation for a track like this.


Solo the kick channel and open the corresponding Duende channel. Now bring the bass track into the soloed mix so we can get these two parts working together so they complement each other.

Press the BYPASS ALL button on the kick track to bring the processing into place with the bass still soloed.

  • The high-pass filter has been set to about 50Hz to let the bass breathe, and a boost of 5db has been added at 130Hz to compensate. This has been done to enable the kick to sit slightly above the bass.
  • A boost of 5db is used at 4kHz to emphasise the click of the kick drum. This is where most of the attack information is and will allow the kick to come through the mix.

Now open the channel on the bass track and look at the EQ settings.

The low-pass filter at 3kHz clears up the top end slightly as this is mainly noise.

  • A boost at 72Hz is used to give the bass some more weight.
  • A cut at about 740Hz is used to move the bass slightly out of the way of the snare which can be demonstrated by bringing the snare track into solo.
  • Finally, a boost at 950Hz is used on the bass to give it a little more body.

Now the bottom end is sounding tighter, clear all solos and lets concentrate on the hihat. With the Duende channel bypassed you can hear low end information which is not really needed on a track like this.

By bringing the Duende channel into play, you can hear how this has been cleaned up. The low-pass filter is wound up to 300Hz to eliminate this low end, and a boost at 4.5kHz gives the hats more bite and presence.

Lets have a quick look at the snare before bringing in the rest of the kit mics. Bear in mind that a lot of the snare in the drum mix will come from the tom-oh and room mic tracks.

  • A resonance has been found at about 730Hz which has been reduced with a LMF cut in this area (more information on this below).
  • The bottom end has been removed with the high-pass filter and a slight amount of punch has been added at about 200Hz.
  • Finally a small amount of top end has been added to bring a little life back into the sound.

Solo the whole kit using the 'kit bus' sub-group. Notice there is a bus comp placed over this group, and by bringing this into circuit using the 'COMP IN' button you can hear the drums working together more sympathetically to each other . Bring the bass track into solo and you can hear the rhythm section working well together.

Guitars etc.

With the drums and bass in solo, bring the rhy gtr 1 track into the mix. Open the Duende channel and hit the BYPASS ALL button.

  • The low end has been trimmed off using the high-pass filter and the lower-mid boosted.
  • The HMF band has been used to move the guitar out of the way of some of the drums and a top end boost to open the sound up without it sounding harsh.

Now bring the rhy gtr 2 track into the mix. With the Duende channel bypassed this track tends to bleed over the bass and rhy gtr 1. Hit the BYPASS ALL button to turn on the channel and look at what has been done to sit this in the mix better.

  • Again the low-end has been removed with the high-pass filter and the frequencies around the snare have been toned down.
  • A high-mid boost pokes the guitar through the mix and some high end opens it up slightly.
  • Some compression is used to even out the dynamics slightly.

Finally the pad has been sat at a low level and probably doesn't need any EQ.

Now A/B the processed and un-processed mixes using the bounces at the bottom of the screen. With the Duende channels switched on the mix sounds more open and energetic with more definition between instruments. This is because each one has been moved into their own place within the mix.

Notice that the peak level on the master fader is in fact lower with the processed mix, but the mix sounds louder and bigger.

As in tutorial 3 these EQs were arrived at by processing ‘in place’, that is not in solo. For these purposes however we are soloing to highlight the different processing taking place.


Let’s go Hunting

Sometimes you will be presented with signals which exhibit resonances. These resonances are high levels of energy at a particular point in the frequency spectrum which are often not related to the music.

They tend to mask other frequencies and are generally not pleasing to the ear, and can become very annoying after a while. Vocals tend to exhibit these, as do instruments recorded with poor microphone placement.

A well trained ear will be able to hear, pin-point and pull out these frequencies without too much of a problem. However if you don’t know you’re frequencies off by heart there’s a tried and tested method for dealing with these gremlins. If you suspect a signal has a resonance which you don’t want to hear in your mix, here’s what you can do to control it.

Look at the vocal track in the tutorial file. Even though the vocal sounds fine on first listen, there are in fact resonances present which when controlled will enable the vocal to sit better in the mix and generally sound 'flatter'. By flatter we mean that the frequency spectrum is leveled out, so there are no obvious peaks across the range. This can easily be seen on a spectrum analyser.

By using a high Q value and applying a large amount of gain, it has been found that resonances are present at 247Hz and 2235Hz. If you increase the Q values on the LMF and HMF bands to 2.5 and wind the gain up to about +15db, you can hear an exaggeration in the signal at these points. Now reduce the gain down to -8 db or so and decrease the Q to about 2. You will notice that the vocal sounds more open and intelligible.

Obviously how much cut you use to pull out the frequency is subjective, and beware that using large amounts of cut with high Q values can sound quite strange. This is why it is advisable to back off the gain and Q value once the resonance is found.

Tutorial 7: Parallel Compression

Tutorial 7: Parallel Compression

Note to Pro Tools LE and M-Powered users:

Please set the H/W Buffer Size to 512 Samples. This setting can be found in setup>playback engine.

Also where this tutorial references the 'drum' track, please take this to mean the 'dry bus'. Do not adjust the level of the 'drum' track level as it will result in the audio being out of sync.

Have you ever felt yourself wanting that compressed, pumping drum sound that still retains the dynamics of the playing? If you dial in the required amount of compression the kit just sounds flat and lifeless, yet if you back the compression off the intensity and energy are not there.

Parallel compression is technique that involves mixing the original uncompressed sound with a suitably compressed version of it.

The Duende Bus Compressor can be very useful in this situation. Drive it hard and you get that classic SSL compression sound, but mix this with an uncompressed version and the life and dynamics are retained.


Load the tutorial 7 DAW session;

Here are two tracks; drums and parallel bus. Play the track and get used to the dynamics of the unprocessed drum track.

This track is send to a sub-group (bus) via a pre-fade auxiliary send. This means that the level of the fader on the drums track does not affect what is sent to the sub-group. The send to the sub-group is set at 0db.

Now pull the drums fader all the way down and push the parallel bus fader up to 0db. Now you hear the same track being smashed by the Bus Comp. Under normal circumstances this would probably be far too much compression to use as the main drum sound in a song.

  • A low threshold and 4:1 ratio are used to obtain a large amount of gain reduction, about 20db on the peaks.
  • The attack is at 3ms which allows some of the initial punch of the drums to come through, but doesn't hesitate too long before clamping down on the signal.
  • A very fast release of 0.1ms is used to cause the pumping effect.

Now you have the parallel bus 'over' compressed, push the level of the drums fader up until the overall sound regains some dynamic and clarity.

Mix the level of these two faders according to taste, depending on what you're after.

This works because the parallel bus and the drums track are exactly in phase and time thanks to Automatic Delay Compensation available in most DAWs.

This technique can be used on any material, and can be considered an alternative way to work where the desired compressed sound can be mixed in to taste with the original uncompressed sound.

Download project files

Click here to download project files

Select the appropriate Project Files for your preferred DAW using the links below:

Please note that when the project file is opened for the first time it may ask you where the audio files are. In the case of Cubase/Nuendo and Pro Tools they are found in the 'audio' folder within the main project folder. With Logic the audio files are found in the project folder. Once you have pointed the application to the audio, it is advisable to save the project so these locations are stored and this doesn't need to be done again. Thanks to Jon Jannaway at SessionLoops for providing us with the majority of the audio.

Cubase & Nuendo

Tutorial 1: Compression and limiting - 10.3mb
Tutorial 2: Using the Gate/Expander section - 2.26mb
Tutorial 3: Basic EQ - 5.14mb
Tutorial 4: Using the Duende Channel side-chain - 980kb
Tutorial 5: Duende Bus Compressor - 2.14mb
Tutorial 6: Advanced EQ - 19.5mb
Tutorial 7: Parallel Compression - 1.34mb


Tutorial 1: Compression and limiting - 8.5mb
Tutorial 2: Using the Gate/Expander section - 2.25mb
Tutorial 3: Basic EQ - 5.12mb
Tutorial 4: Using the Duende Channel side-chain - 985kb
Tutorial 5: Duende Bus Compressor - 2.13mb
Tutorial 6: Advanced EQ - 17.7mb
Tutorial 7: Parallel Compression - 1.34mb

Pro Tools Mac

Tutorial 1: Compression and limiting - 9.6mb
Tutorial 2: Using the Gate/Expander section - 2.49mb
Tutorial 3: Basic EQ - 5.40mb
Tutorial 4: Using the Duende Channel side-chain - 991kb
Tutorial 5: Duende Bus Compressor - 2.16mb
Tutorial 6: Advanced EQ - 19.8mb
Tutorial 7: Parallel Compression - 1.47mb

Pro Tools PC

Tutorial 1: Compression and limiting - 9.91mb
Tutorial 2: Using the Gate/Expander section - 2.34mb
Tutorial 3: Basic EQ - 5.22mb
Tutorial 4: Using the Duende Channel side-chain - 1.06mb
Tutorial 5: Duende Bus Compressor - 2.23mb
Tutorial 6: Advanced EQ - 19.5mb
Tutorial 7: Parallel Compression - 1.44mb