Using Filters Part 2 - Shaping Sonics

This article goes deep into the usage of the Filter2 effect for shaping and controlling the sonics of your song mix. It doesn't matter which genre of music you compose in, as these concepts are transferable. Before we go into detail, please look at Beatslaughterer's article on the explanation of filters; I'll be assuming the readers understanding of some basic terminology as we move through the concepts.

Aim For Reductive

Mixing your music in Renoise can easily turn into a quagmire of EQ settings for every track, and often this can be imprecise and messy. EQ used in an unsure way and often in an additive way ends up introducing more noise than needed, especially when using cheap band-EQs. We'll discuss EQ in a later article because it's better seldom used and in place we'll be using Filter2. A better sensitivity to using the filters will eventually lead to better usage of a quality EQ. It will get you in the good habit of "reductive mixing".

In mastering a good principle to hold is: "the quality of the mastering is largely determined by the quality of the mix". So if you have a great mix, then little work will need to be done in mastering. Getting a good mix can be aided by an understanding of the limitations of the digital format, which is the core of your audio-canvass in Renoise. What restrictions you might ask?

Assuming you are working at a digital sample rate of 44,100hz (slices of data per second) then we can already see what the upper limitations are: the highest frequency sine wave (or sound that uses both positive and negative parts of the data) can be 22050hz - because one data slice is needed for the positive vibration and one sample is needed for the negative vibration. I hear you saying, "great" because most humans can't hear notes much above 18000hz. But, and this is a big but, is a sine wave really going to be a smooth curve at 22050hz or near to that at 44100hz? The higher the frequency the more the resolution of that frequency is lost. The sounds no longer have their analogue smoothness. So, the higher end of the frequency spectrum in digital is usually considered as 'noise' or 'harmonic color' - no real musical notes are up that high. Most natural sounds and synthetic sounds have content up in that region but a lot sounds differ in terms of the purity or harshness of that content. Because the quality of that noise is being crushed by the sampling restrictions sometimes those upper frequencies are not needed or entirely not desirable.

What about the low end of the frequency spectrum? Here the problem is not frequency but amplitude. Sub-low frequencies below 40hz, because of the Flectcher-Munson curve, take more amplitude to be expressed. We often do not need those frequencies in music because they are mostly rumble sounds: the kick and bass sit higher than 40-50hz. Why would high amplitude rumble be an issue for the mix? When it comes time to master the track, specifically in gently boosting the levels, large amplitude rumble will quash out higher definition sounds because of the lack of headroom in digital audio. Also, excess rumble will make it hard to discern the clarity of the desirable bass elements - hence the term 'muddy'.

So what is sonically desirable and balanced? What is the sonic model we are aiming for? Here a little listening homework will be necessary! Assuming you have reliable monitors to listen to music on, go get your hands on some recent recordings of classical music. Yes I know, classical might not be your thing, but you'll find no more reliable reference material in any other popular genre. Most classical music has been recorded for optimum fidelity in mind, so that it near enough simulates the experience of actually being there at the gig, hearing all the acoustic instruments in pure form. You'll find that the audio has a clear and deep quality to it that we can use as a sonic model for mixing. Try to find something that has a range of loud and soft in it, like easy stuff such as Beethoven's 9th Symphony or something more special like Arvo Pärt's Tabla Rasa. Listen to the balance of the frequencies: the highs are never too intense, the bass is present but not muddy. Study the sonic expression of the sound, and then use this benchmark to apply a sonic model onto your own mixing. To get anywhere near the clean expression of sound you'll have some filtering to do.

We can use Filter2 to get rid of unnecessary frequencies to get your individual tracks as clear as each instrument in the classical model. It's a very powerful and underused tool that I have on almost every track of every one of my Renoise songs. Every sound, in every song, within every genre, will all have different needs and the best way to find what works is to get your hands dirty with trying different settings.

But a few explanations of the different settings in Filter2 first. I'll be using some shorthand terms: "Cut" (cutoff) refers to the frequency cut-off of Filter2, and "res" (reso/Q) refers to the resonance or amount of feedback Filter2 is using. Next, choosing an filter with the right 'slope' is important. The slope refers to how sharp the cut-off is, e.g. -12dB refers to the volume of the sound dropping away from the cut at a slope of 12dB per frequency octave. So, a slope of -48dB is steeper than -12db. The Moog filter has a strong immediate cut and thus has a slight resonance to it suitable for automated filter sweeps. Also, khz is a fancy way of saying 1000's of hz.

What follows is some of my common usages of filter2 for making a slick mix:

LP Filtering

LP filters are useful in knocking out unwanted high frequencies of each track. Here are some common examples:

• Temper distortion: a Moog or -12dB cut of anywhere between 4khz to to 12khz. This makes the distortion sound smoother and warm. Usually the larger the amount of distortion the lower the cut. Sometimes just a small amount of res to brighten the cut point.
• Reducing the hiss aspect of old samples - when you simply have no alternative. A Moog cut around 12khz-16khz usually works.
• Harsh high-hat and cymbal sounds can be fixed with a Moog cut around 14khz to 16hkz, sometimes with 2-3% res to re-compensate for brightness. Alternatively you can use a smoother -12dB roll off depending on the characteristic 'sizzle' of the sound.
• Some pad/string sounds really benefit if you filter out the highs, especially saw-tooth Juno-like pads. Here you can go as low as you like, but also use a small amount of resonance relative to how 'distant' you want the pad to be.
• A deep cut on a kick drum, sometimes all the way down to 100hz can have a nice effect for a pulsating sub-kick heartbeat, suitable for all sorts of genres.
• Any other sound that doesn't need the upper frequencies expressed - it's almost always better to cut the tops than to have a really busy noisy upper range.

HP Filtering

HP filters are useful in knocking out unwanted low frequencies of each track. It is ideal to have a tuned sub-woofer with your speakers so that you can get accurate monitoring all the way down. Here are some common examples:

• Many kick drums benefit from having a -12dB HP filter with a cut around 90-110hz. This keeps the kick punchy but rolls most of the subsonic amplitude.
• Taming the bass - A -12dB HP filter with a cut around 120-200hz depending on the character and key of the bass sound. A small amount of res (less than 6%) will boost the low end punch without over doing the subs.
• Keeping vocals behaved - A Moog HP at 80hz to take out any dangerous plosive amplitude bursts. Great for any acoustic recordings.
• Handy on counter-percussion/loops - use any of the HP filters above 300hz to take out the lows so they don't interfere with your main kick drum.
• Any other high melodic instruments will benefit from clearer expression without their low component - experiment and find points that are suitable for your sounds.

Using Two Filters

Quiet often it becomes necessary to use both of the above techniques in conjunction to really focus down onto the necessary sound. Yes, you could use a Band Pass filter, but I prefer using two Filter2 plugs so that on each I can have the freedom of using different slopes and resolutions. Here are some common examples:

• Reverb and Delay decay - Cut the lows below 600hz or higher and cut the highs with a Moog HP 2k or higher. This will help to get a more authentic echo and one more reminiscent of the cool analogue dub delays of days gone by. Because of the narrow frequency range the sound will become more distant (therefore magical).
• Kicks and Basses - as discussed before in HP filtering, but used in conjunction with LP filtering at any desired setting. Taking out all the high sounds in the low instruments will free up the canvass in the mids.
• Narrow the pads and ambience - just to the mid-range with resonance on the LP filter. Again this will improve the illusion of distance.

Ready, Set: Filter!

Of course, there are probably millions of more examples that could be stated on how to use filters to shape the sonics. Because of the wide variety of genres most of these examples will be fairly academic, and are best found through experimentation. The general principle holds: reduce bass for control of sub-rumble and mix-amplitude, reduce highs to lower digitalized noise. Once you have tidied up your mix in this fashion you'll reap the rewards in the mastering process - and - your precious composition will be all the more enjoyable for an audience.