Music 265B Week 10 Mixing, Part 1
PDF version available HERE
We may have prepared a rough mix for the band during the recording & editing phase, but we don’t start the actual mixing phase until all of the song’s elements have been recorded and edited into the right place. When we mix a song, we try to accomplish a few goals
- Bring out the ideal timbre in each instrument, with respect to the song’s style.
- Blend & balance the individual instruments into an ensemble performance.
- Prepare the song for the mastering process.
Mixing With A Purpose
Mixing can be extremely subjective, and most mixing decisions come down to a matter of personal preference. There is no one true way to mix any given song, but there are some things to consider. Different instruments and many styles of music have their own “classic” stylistic sounds. A guitarist may want to emulate someone else’s iconic guitar sound, or a band may bring in their three favorite albums from similar artists and say, “Make us sound like those guys.” Your client may not be as good as their musical heroes, but having some points of reference will guide the mix in the right direction.
Use Your Ears, But Don’t Trust Them
Your ears will play tricks on you. Everything from the room’s acoustics to the speakers’ frequency response will “color” the sound that you’re hearing: some frequencies will get boosted, others will get cut, and the way that sound travels through the room will affect how the mix sounds to us. These changes will also affect how the mix “translates” to other systems. In the end, what sounds like an awesome mix in your home studio may sound terrible in your car. To help prevent this, most engineers will listen and A/B the mix through several pairs of speakers, like a relatively “accurate” pair of studio monitors with a subwoofer, a pair of consumer-grade computer speakers, headphones, or an old boom box. In the end, the best mix should sound good on any of these devices.
Another important thing to consider is Ear Fatigue. As you sit in the same room listening to the same tracks on the same pair of speakers for hours on end, your ears will gradually get desensitized. To keep our ears fresh, it is important to take a break every so often to rest our ears. We can even try to “reset” our hearing by taking some time to listen to something that sounds completely different for a few minutes. A change of sound and a change of scenery can help. If you’re mixing for long periods of time, take a break and come back a day later with a pair of fresh ears. What sounded great at the end of the day yesterday might sound awful today, now that our ears have had some time to rest.
Preparing a Session for Mixing
Our computer has a limited amount of processing power to spare. When we had to record musicians as they played along to our tracks in Pro Tools, we needed our system to have minimal latency. A long delay in their headphones would have hindered the musicians’ performance. Since we’re done recording, we can dedicate more of our computer’s power toward processing the plug-ins in our mix, without worrying about a lag in the playback sound. Go to the Setup > Playback Engine menu. These settings will vary from system to system. Find the H/W Buffer Size setting. Raise the hardware buffer size to its highest setting. Under Host Processors, select the 2nd highest option available: for example, if you are using a quad-core laptop, allow Pro Tools to use 3 out of the 4 cores. This will dedicate the majority of your computer’s CPU cores to Pro Tools’ engine, while leaving one available to process the computer’s background tasks. Set the CPU Usage Limit to roughly 80% to start. Set the Delay Compensation Engine to maximum to start. These settings should have a nice tradeoff between raw processing power, and an ability to handle other system tasks. If Pro Tools is frequently suffering from error messages during playback, change these settings around until you find something that works for your system. To see how hard your system is working, select Window > System Usage.
Starting to Mix
Before we recorded our tracks, we stressed the importance of microphone choice and proper placement in order to capture the best possible sound from our instruments. During the setup for our recording session, we used proper isolation to minimize the bleed & leakage from other instruments, and we used special settings like high pass filters on our microphones. Some microphones were better suited for certain sound sources, and we even used multiple microphones on the same instrument in some cases. During the editing phase, we chose the best takes, in terms of performance & sound quality. We also used editing tricks like Strip Silence to clean up the “dead air” in our tracks. Hopefully, after all that work, our tracks sound great. However, the sound of these tracks might not be the exact sound we’re looking for in our mix. Whether we need to gently sweeten our tracks, or completely warp the tone of each instrument, we use some of the same basic tools across all of our tracks: Equalization, Compression, and Reverb. These aren’t the only tools we will use, but they’re often the first plug-ins we turn to.
Edit & Mix Window Signal Flow
Just so we have a clear understanding of how our sound gets processed through Pro Tools, let’s review our signal flow.
Signals get into our tracks through the Track Inputs (the Interface, or an internal Bus from another track). They are recorded as Audio or MIDI Clips, depending on the kind of data we are working with. These raw clips are edited, which usually involves trimming, moving, quantizing, and warping with functions like Strip Silence, Beat Detective, and Elastic Audio. They may be processed and rendered into new clips with AudioSuite plug-ins. The clip’s playback volume may be altered with Clip Gain, as well as Fades at the start & end of each clip.
After all of those functions in the Edit Window, the sound of our clips travel through the tracks’ 10 Inserts in order from the first insert at the top of the channel, to the last. The inserts are where we stick the majority of our plug-ins, including software instruments (for MIDI clips), Equalization (EQ), Compression, and other effects, which we will discuss throughout the mixing phase. A portion of the track’s signal can be split off and routed somewhere else through any one of the track’s 10 independent Sends, located below the inserts. Sends can be used to send this secondary signal through one of Pro Tools’ internal Busses, or through another path. We may have used these in the past to create a separate headphone mix while we were recording, but during the mix, we typically use these for special effects like Reverb & Delay on a separate Aux track, among other things, which we will discuss later.
The track, and each individual send has a Pan knob, which can be used to shift the signal around between the left & right sides of the stereo image. Tracks & sends also have a Fader, which controls the track’s output level. The Mute button will silence the track or send’s output. Sends can be set to Pre-Fader mode, meaning the send’s output level will be independent from the track’s fader/output volume. Lastly, the entire signal exits the track through the Output section of the I/O. We can use the track’s output to route the entire signal to another track, like a submix on an Aux track, or out of the system to our ears through the Master Fader.
Standard Plug-Ins are signal processors that affect the sound of an entire track. When we activate a plug-in on a track’s Insert, the sound of the clip flows through the plug-in, in order from the first insert to the last. Processing a signal with regular plug-ins is non-destructive: we can add, remove, or make changes to a plug-in’s settings at any time without changing the original sound of the clips in our edit window. In other words, inserted plug-ins only affect what we hear during playback. Since this applies the same setting across the entire track, we use these inserted plug-ins for the majority of our mixing work. On the other hand, AudioSuite plug-ins (available through the AudioSuite dropdown menu) are used to make changes to individual Clips in the edit window. We often use AudioSuite plug-ins to create special effects or “fix” problematic audio clips in specific sections, rather than processing the entire track. When we select a clip and process it with an AudioSuite plug-in, our changes get rendered into a new clip, which replaces the old one in the edit window.
In other words, to process the entire track’s signal, activate and use plug-ins on the track’s Inserts. To make changes to individual sections and clips, use AudioSuite plug-ins.
Imagine two runners on a track. One lane is clear, but the other is full of hurdles. If the two runners race at the same speed, the runner who has to jump over the hurdles will eventually lag behind the runner in the empty lane. To stay in sync, the runner in the empty lane may need to slow down. When we use powerful plug-ins or add more to certain tracks, the system has to work harder in order to keep up with all of the processing. Normally, the sound of tracks with lots of plug-ins may eventually lag behind our other, like the runner jumping through hurdles. Pro Tools tries to mitigate this with Delay Compensation. When a track’s playback is lagging behind the others (delay), Pro Tools will delay all of the other tracks by that same amount (compensation) in order to make sure the tracks are heard in sync with one another.
The sound spectrum is arranged from low to high-pitched frequencies measured, in Hertz (abbreviated Hz: 1000Hz is abbreviated to 1kHz, or kilo-Hertz). In digital audio, the audible range is typically 20Hz to 20kHz. We can divide this spectrum into several distinct areas.
The Low frequencies (~20-60Hz) are the chest pounding bottom end of the frequency spectrum. They include the fundamental pitches of most kick drums, and the lowest notes on the bass guitar. These frequencies are usually played back through a subwoofer: a standard pair of studio monitors may not be able to accurately play these frequencies. Too much of these extreme lows can sound muddy.
The Low-Mid frequencies (~60-250Hz) include the low fundamental pitches of most instruments. This is the warm, fat, and punchy low end of most instruments. Too much low-mids can sound boomy, and too little can sound thin.
The Mid frequencies (~200Hz-2kHz) include the high fundamental pitches and lower overtones of most instruments. It is one of the widest ranges in the spectrum, covering 3 octaves on the piano starting at middle C. These ranges cover the attack sound of most string instruments. The midrange has a lot of troublesome areas. ~200-500Hz can provide a nice fullness, but too much will make our drums sound like cardboard boxes. ~500Hz-1kHz has a honking hornlike quality. ~1-2kHz can have a tinny, nasal quality.
The High-Mid frequencies (~2-6kHz) include the upper overtones of most instruments. This range covers the attack sound of drums & cymbals, and part of the Sibilance range in vocals (hard consonant sounds around 4-8kHz). Boosting the high-mids can help instruments pop out & cut through the mix. Too much can be fatiguing.
The High frequencies (~6-20kHz) include the highest overtones, and the upper sibilance range. Boosting this range can add brightness, brilliance & sparkle. Too much can sound brittle, too little can sound dull. As we get older, and lose our hearing, these higher frequencies are typically the first ones lost. Many people can experience a significant loss in frequencies above 15kHz.
Equalization (EQ) plug-ins let us boost or cut back specific frequencies within the audible sound spectrum (20Hz-20kHz). Pro Tools comes standard with a powerful Equalizer plug-in called EQ3 7-Band. As the name implies, we can use this EQ plug-in to boost or cut back up to 7 different bands (frequency ranges) within the sound spectrum. The plug-in displays our current settings on a color-coded graph: each color representing the different bands on the EQ. The vertical numbers (measured in Decibels, dB), represent a change in the EQ curve; positive numbers represent a boost, 0 is no change, and negative numbers represent a cut in the curve.
In the upper left corner of the plug-in window, the Input knob can boost or cut down the incoming volume (Pre-EQ). The Phase Reversal button (a zero with a slash) will flip the incoming signal 108 Degrees out of phase, when pressed. The Output knob controls the outgoing volume from the plug-in (Post-EQ). These knobs are defaulted to 0 dB (no change).
Beneath the Input & Output controls are the filters. The High-Pass Filter (HPF), or low-cut, can be used to roll-off the Low frequencies, just like the roll-off switch on our microphones. Activate it by pressing the IN button next to the HPF section. We can choose the shape of the HPF curve: either a Low-Shelf (default), which will affect all frequencies below the Frequency knob’s setting, or a Bell Curve which will affect a narrow bandwidth. Use the Q knob to alter the bandwidth and intensity of the filter.
The Low-Pass Filter (LPF) works just like the high-pass filter, except it rolls off the high frequencies, allowing the lower ones to pass through.
The bottom section of the EQ3 plug-in contains 5 individual EQ bands, corresponding to the Low Frequencies (L), Low-Mid Frequencies (LMF), Midrange Frequencies (MF), High-Mid Frequencies (HMF), and the High Frequencies (HF). Each section contains 3 knobs: a Frequency knob to select the center of the curve, a Q knob to adjust the width, and a Gain knob to boost or cut back this frequency range. These bands don’t need to be used within those specific ranges. When combined together, the bands may overlap and create different EQ curve shapes.
EQing a Track
Back in our discussion of microphone choice & placement, we described the frequency ranges of many different instruments, and some of their sonic characteristics. Every note has a fundamental pitch (its lowest frequency) followed by a series of overtones. If we know what an instrument’s lowest note is, or how it is tuned, then we can start by eliminating all of the frequencies below that lowest fundamental pitch. Let’s imagine we are trying to EQ a violin in standard tuning. When the strings are in tune, the lowest string is tuned to the note G3 (195Hz). In this case, anything below 195Hz is not coming from the violin: we can consider it to be noise, since it probably contains leakage from other instruments. We can activate our High-Pass Filter (HPF), set it to 194Hz, and chop off every frequency below that. From there, we can listen to the sound of the violin, and adjust frequencies to our preference. The lows may sound thin, so we boost them. The mids might be ok, so we leave them alone. The high-mids may sound scratchy, so we cut them. The high overtones may sound dull, so we boost them to add some sparkle.
In some cases, we don’t know where the sounds we hear are coming from. In that case, we have to use our ears & EQ bands to figure out where the problem areas are. Most drums don’t tune to a specific pitch. Instead, the head is tuned to the shell’s lowest possible pitch. To figure out where this is, we can solo the snare drum, activate the EQ plug-in, and enable the High-Pass Filter. While the snare drum plays, we can gradually raise the frequency knob until we hear the lows of the snare drum disappear. Then, we can gradually turn the frequency knob back down until we find the exact spot where the snare drum’s lows come in clearly, while removing the low kick drum leakage from the track. The frequencies immediately above this area should be the low frequencies of the snare drum. We can center our Low Frequency band over this area, and either boost or cut back this section to our liking. We may have to figure out where the attack sound of the snare drum is coming from. Since this is buried somewhere in the middle of the track, we’ll need to use a hidden feature in the EQ3 plug-in: Band Pass Mode.
Band Pass Mode lets us effectively solo an individual EQ band inside the plug-in. To use it, hold down the Control & Shift keys, and click on one of the EQ bands. This will bypass the Gain knob, but it will let us hear just the frequencies inside of our selected EQ band. While the track plays, we can still click and drag on the Frequency & Q knobs, and scan back & forth through the spectrum until we find the sound we’re looking for. Once we’ve found it, we can let go and adjust the Gain on this band to boost or cut back these frequencies. This trick is great for finding trouble spots, like the unwanted ring in a snare drum, the nasal sound of a vocalist, and so on.
Boost or Cut?
Equalization will raise or lower the volume of specific frequencies within a track. In the end, this will usually make the overall track louder or softer in the mix. There are a few ways to approach this. With Additive EQing, we boost the frequencies we want to hear (the good parts), and then turn down the overall volume to match the track’s level. In Subtractive EQing, we cut the parts we don’t want (the bad parts), and raise the overall volume to match the track’s level. In the end, if we’re boosting, cutting, and then adjusting the output gain to match, then there is no difference. However, since most people don’t bother to match the overall volume when they EQ, their tracks often wind up getting louder and louder until they’re in danger of clipping. Because of this, as a general rule, it’s usually better to cut when we EQ. If you’re not sure where to start, the EQ3 7-Band plug-in comes with a wide variety of Presets available in the Preset dropdown menu. Look for the words “<factory default>” near the top of the plug-in window. These Presets are good starting areas for finding an ideal sound for any given instrument. They may not be exactly what we want, but with some fine-tuning, they can be a timesaver. However, you should always trust your ears, not the preset: their idea of how an instrument should be equalized may not work for your recording.
One Instrument, Multiple Microphones
Having multiple microphones on the same instrument can be tricky. We may have used two microphones each on the kick & snare drum. In this situation, we might have each track focus on what the others lack. If we used a sub microphone and a regular dynamic on the kick drum, the sub might have great low end, and terrible highs. The dynamic might have decent lows, and strong mids & highs. In this case, we can focus our sub track on the fat bottom end of the kick: use a low-pass filter to remove the highs (and bleed from the rest of the kit) on the sub, and blend these lows into the mix. The dynamic microphone’s track may be EQed to focus on the kick drum’s attack and punch in the mix. Individually, these tracks may not sound good at all. When blended together, these two signals should produce a massive, full-range kick sound. For the rest of the drum set, our overhead microphones usually provide most of the overall drum sound in our recordings. With this in mind, we can use our individual snare, tom, and cymbal microphones to focus on each drum’s attack sound, which may be lacking in the overheads. We can then gently blend these other microphones under the main overhead signal. One signal always supports what the other one lacks.
Vocals can be problematic. In general, we want to preserve a singer’s bright overtones, but the Sibilance can be too harsh. Any time the singer uses a sharp, hard consonant sound like “S” as in the word Sibilance, “T” as in Time, “ K” or “C” as in Car, and so on, these frequencies might be too loud. To counteract this, we can use a De-Esser plug-in. Whenever one of our troublesome sibilant Frequency gets past a certain loudness (or Threshold), the de-esser will reduce the signal’s level by an adjustable amount. Pro Tools comes standard with a plug-in called Dyn3 De-Esser, available under the Dynamics plug-in category. Near the bottom of the plug-in, the Frequency knob controls the frequency we want to target. The Range knob determines how much the signal will be reduced. Under the Options section, HF Only will suppress just the high frequencies (anything above the Frequency knob) instead of the overall signal. The Listen option lets us hear just the frequencies that are currently being affected. Use this to fine-tune the de-esser’s settings.
De-essers are great on vocals. Since we use them to correct a problematic signal, they usually work best when placed in front of an EQ plug-in (one of the inserts above the EQ) in the signal chain. This lets us focus on enhancing the sound of a smoother vocal use our EQ. Later on, we may even decide to place a de-esser in front of our reverb track in order to control some unwanted sibilance.
Dynamics: Compressors & Limiters
A Compressor plug-in is used to reduce (compress) the dynamic range (the contrast between soft and loud volumes) of a signal. Overall, it can make the quietest parts of a track louder, and the loudest parts quieter. Pro Tools comes with the Dyn3 Compressor/Limiter plug-in, among others. Compressors have a Gain knob that controls the level of the incoming signal. This part of the compressor is used to raise the overall volume of a track. In other words, this gain function makes the quiet parts louder. The rest of the compressor is designed to suppress the louder parts of the track. Whenever a track gets louder than the compressor’s Threshold setting (measured in Decibels, dB), the compressor kicks in, pushing back against the signal. How hard the compressor pushes back depends on the Ratio we use. For example, with a 2:1 ratio, when the signal peaks at 2 dB louder than the threshold, the compressor will only allow the signal to go 1 dB louder. If the signal goes 4 dB over, the compressor only allows it to get 2 dB louder. A Gain Reduction (GR) meter shows how much signal is being reduced during this process. A Limiter does the exact same thing as a compressor, but limiters typically operate on a much larger ratio: 10:1, 20:1, or ∞:1. While a compressor will still allow a signal to peak above the threshold, a limiter with a high compression ratio will usually prevent the signal from going too far overboard. We typically use this during the mastering phase to prevent any potential clipping.
The Dyn3 Compressor/Limiter has a graphic display. The orange vertical line represents the Threshold. The white line to the left represents the level of the incoming (uncompressed) signal, and the line to the right represents the compressed signal. The Ratio determines the slope of the compressed line. The point where the threshold and ratio intersect (the orange & white lines) is called the Knee, because sit looks like a bent knee. This Knee curve lets us choose between having a hard transition from the uncompressed to the compressed signal (a “Hard Knee”), a gradual transition (“Soft Knee”), or somewhere in between.
We can adjust how quickly our compressor reacts to a loud signal. A compressor’s Attack time affects how quickly the compressor will activate whenever a signal peaks above the threshold. The Release time determines when the compressor stops pushing back. If these settings aren’t adjusted properly, we may hear the compressor “pump” when it turns on and off. Normally, we don’t want that to happen. A typical compressor should make the track sound smooth and relatively even. The compressor may make the track sound thicker and more full when the signal is compressed. However, too much compression will lead to Saturation and distortion. Extreme compression can be extremely fatiguing to our ears, and it will ruin the dynamic contrast in a song. The loud parts won’t sound loud and exciting when there aren’t any softer parts to create a contrast.
Special Compression Techniques
Side-Chain Compression allows us to use a signal from one track to compress the signal on another track. For example, if the bass guitar is drowning out the kick drum, we can use the kick drum to trigger a compressor on our bass track. The compressor will cause the bass guitar to get softer for a moment whenever the kick drum is struck. To do this, activate a Send on the kick track (or whatever track we want to trigger the compressor), assign it to any unused Bus (e.g. Bus 1), and raise the send fader to 0. Activate the compressor on the bass track, and look for the skeleton Key icon with the “no key input” dropdown menu: this is the key input selector. In this menu, choose the same Bus we selected earlier for our kick drum signal. In the compressor’s Side-Chain controls, select the small Key button. The compressor will now use the incoming kick drum signal to compress the bass guitar. This technique is commonly used in dance music: a kick drum will compress a synthesizer track, but the compressor’s exaggerated settings will cause the synth track to “pump” in time with the beat.
Parallel Compression blends a heavily compressed signal with a lightly compressed (or uncompressed) signal. This can have the effect of making a track sound thicker without adding excessive volume, or ruining the dynamic range of the track. To do this, we can duplicate the track, or split our track’s signal onto a separate Aux track. We can then heavily compress the second track, and gradually blend it alongside our original.
Dynamics: Expanders & Noise Gates
Compressors restrict the dynamic range of a track; Expanders do the opposite. An Expander has controls that look and function juts like a compressor. However, instead of limiting any signal that peaks above the threshold, the expander will raise them. We can use an expander to exaggerate the dynamic contrast in a track. A more extreme form of an expander is called a Noise Gate. A noise gate will suppress a signal until is gets louder than the threshold. When the signal is loud enough, the gate swings open for a set amount of time before it closes again. We can use noise gates to suppress the background noise in a track, similar to how we used Strip Silence in the editing phase. Imagine a drum set. The snare microphone may capture a lot of leakage from the kick drum, or the rest of the kit. If the snare was recorded properly, the snare’s signal will be strong, and the leakage from the kick drum will be softer. To remove the kick drum’s leakage, we could try to EQ the low frequencies away, but we could also use a noise gate, with the threshold set higher than the kick drum’s leakage, but softer than the snare drum’s level. Whenever the snare is struck, the gate will open, and we will hear the snare sound with minimal bleed from the kick drum. We can adjust the Attack, Release, and Hold settings to keep the gate open until the snare is done ringing. We could use it to cut off the snare’s ring too.
The Channel Strip plug-in combines an Equalizer, Compressor/Limiter, and Expander/Gate into one plug-in. Each module within the Channel Strip functions just like their stand-alone versions.
When we hear someone speak or play an instrument in a room, the sound radiates out from the source. This means the sound travels straight to our ears (or microphone) as the original “dry” sound of the instrument. That same soundwave also hits the walls, and bounces back around the room: Reverberation. Eventually, this “wet” reverberation reaches our ears as well. The size, shape, and surfaces in the room will affect the sound of this reverberation. When we record an instrument with closely positioned microphones, we capture this “dry” signal, while minimizing the “wet” reverberation. A Reverb plug-in attempts to add the sound of a room back into the mix. Because we record with closely positioned microphones in isolation, we can blend in as much or as little reverb as we want. We can even change the type of room reflections we use, from a small room, to a long hall, to a massive church.
Mixing With Reverb
We usually use individual EQ & compressor plug-ins on each track, since every track requires a unique setting. We may be tempted to add a reverb plug-in onto each track, but this can be inefficient. Whether we’re mixing a dozen tracks or several hundred, adding extra plug-ins may take up more of our computer’s processing power, and some fancier reverb plug-ins can require a lot of power. Ultimately, we want the band to sound like they’re playing in the same room, and we can do this with one reverb plug-in.
First, Create one Stereo Aux Input track, and call it “Reverb.” Next, activate one of the track’s Inserts, and select a Reverb plug-in, like D-Verb. Assign the reverb track’s Input to any unused Stereo Bus (e.g. Bus 1 & 2: any busses that aren’t currently being used). The reverb track can send its output to the main mix. Next, activate a Send on every track in the session (except for the reverb track), and route this send to the same stereo Bus we used earlier. DO NOT Send the reverb plug-in back to itself: this can create a Feedback Loop. To blend some reverb into the mix, simply raise the Send Faders on the tracks that require some reverb.
Get To Work
These aren’t the only plug-ins we will use, but they’re enough to get us started. We will cover more specialized plug-ins, and some advanced techniques in the next lesson. At this point, our focus is on shaping the tone of our tracks. For now, Equalize, compress, expand, or gate the tracks that need it. Mixing can be mostly subjective. What sounds like a good tone for one style of music may not be suited for different one. Individual instruments may sound great when they’re soloed, but they could sound terrible in the mix. Try to find a nice balance that complements the other tracks. Remember to take a break, rest your ears, and check your work the next day. Just like editing, mixing can and will take some time. Get to work.