High-Pass Filters - Retaining Fundamentals
The Ambiguous Approach
Using a high-pass filter is a no brainer right? Just slap one across every channel and start cranking them up! Your mix will be clean in no time! Right!?
"I always high-pass my vocals at about 120Hz." Always?
"Just turn it up until you can hear it change and then back it off a little." During what part of the song?
As a young engineer it is easy to fall into false assumptions before your ear is trained well enough to hear when certain tools stop being useful and actually become harmful. When I first began dabbling in audio mixing, I was very fortunate to have lots of mentors and peers around who knew much more than me about a ton of things. But in hindsight, lots of the information given to me was either ambiguous or purely based on habitual workflows. Even turning to online forums (a whole other can of worms) and articles did nothing to clear up certain ambiguities.
A necessary disclaimer: I believe mixing is an art and therefore forego the constraints of right and wrong. I do however, when necessary, respect certain approaches and techniques that have been proven to yield a desired result. No need to constantly reinvent the wheel right?
So here is my approach to the well loved high-pass filter.
In my mind, the qualifiers of a well recorded sound are very simple. A sound that is representative of the source and doesn't have any or many defects. Defects are easy to identify so let's define those. Defects are the eccentricities that add no desirable character to the sound and are oftentimes unintentional. Defects typically stem from the acoustic environment and/or microphone placement at the time of the recording. So assuming these variables are under control, we are now presented with a sound that has some amount of frequency content. Now on to the source.
A musical pitch is created by some energy vibrating some matter. In an acoustic guitar, it is the vibration of a taut steel string. In a clarinet. the player's embouchure acting on the reed. In a human voice, air vibrating the larynx. Each pitch exists as a fundamental and a series of overtones unique to the specific source. No acoustic sound exists in a vacuum. The acoustic environment plays into the emphasis on certain frequencies which is again why great care must be taken when choosing the environment, and selection and placement of the microphone or microphones is yet another active decision in what frequencies are highlighted.
So we can now begin to see that it is the way in which a source voices a fundamental and set of overtones that makes it unique. Also why I find it so incredibly impossible to subscribe to any "one size fits all" technique in any aspect of recording or mixing.
Fundamentals and Pitch
Beyond dealing with how fundamentals and overtones create a source's timbre, there is a direct relationship between pitch and frequency. Obvious to some of you, maybe not so much to others. But the simplest example I can give is the one we all know: A440. We tune our instruments to it every day (unless you're into the A432 thing). If you play a sine wave at 440Hz, we have agreed to call that the note A. If you then play the A above middle C on a piano, you have a note with a fundamental of 440Hz and lots of nice overtones. Play the A one octave down and you have a fundamental of 220Hz. Pluck the open A string on a guitar and you have an A with a fundamental of 110Hz. The open A string on a bass yields a fundamental of 55Hz.
Feel free to check out this webpage for a full list of frequency and pitch relationships: http://pages.mtu.edu/~suits/notefreqs.html
Briefly reducing this thought the theoretical, let's use sine waves to view how musical tones lay across the linear frequency spectrum.
C major triad
Any equalization decision we make is changing the amplitude of which these musical notes are voiced whether they are fundamentals or harmonics.
In The Mix
Now we can venture back to our track and see how this is applicable. Imagine you have an acoustic guitar that has a pronounced low end. Instinct says, "crank up that high pass filter." But now we know better. In fact, if we set a high-pass filter above around 82Hz on a guitar tuned to standard E, we have already begun to impose on the instrument's full range. To cut off any fundamental frequency would leave us only overtones and render the sound not only unnatural, but inconsistent because notes voiced above the filter will have fundamentals unaffected.
My recommendation is to find the lowest voiced fundamental in the track and set your high-pass there. Meaning if the lowest note the guitar plays in the track is G on the 6th string, then you can cut below 98Hz and not be loosing any fundamentals. Then to clean up whatever low end that is causing trouble, use a low shelf across an octave or more to pull out what you don't need. Again, be sensitive to how much you reduce and where the Q comes back. Remember that you may be dealing with fundamentals in some chord voicings and overtones in others. This is where an engineer has to use their ears and make his or her own artistic decisions.
In practice, to constantly be staring at meters & analyzers and referring to exact numbers on a chart would be overly tedious. This article is only meant as a suggestion to think about filters in a more musical rather than arbitrary way.
This could even mean a simple rethinking of workflow. For example, if I need to set levels quickly on audio I haven't worked with, I'll find the key and then bandpass across the entire mix bus a couple of the lowest octaves and most of the upper ones so I am left with mainly fundamentals in the midrange to work with (in my experience, this is the genius of NS10s but less specific to source audio). Then once the fundamental frequencies of the individual tracks are set, I can open the spectrum back up to balance the lower elements and start making broad strokes across the higher frequencies to shape the character of the mix.
Remember that there is no rulebook when it comes to working with audio. And even though there are many really great tools and resources available to us all, don't forget to use your ears!
My FabFilter Pro-Q 2 filter presets are organized by key and have frequency centers at all useable octaves ready to become a high-pass, low-pass, shelf, etc. Free; your generosity is always appreciated.
Install: Click to download .zip file. Unzip and drag "Octave_Filters" folder to ~/Library/Audio/Presets/FabFilter/FabFilter Pro-Q 2 for macOS, and My Documents\FabFilter\Pro-Q 2 for Windows.
Frequencies for Equal Tempered Scale - http://pages.mtu.edu/~suits/notefreqs.html
Overtones and Harmonics - http://hyperphysics.phy-astr.gsu.edu/hbase/Music/otone.html