Music 265B Lesson 3: Recording from Start to Finish
PDF version is available HERE
For the rest of the semester, we will be recording, producing, editing, and mixing songs in class. Using musicians within the class, we will track drums, bass, guitars, vocals, and others, provided we have the players to do so. Get together and pick out a song of your choosing: a cover or some original work will suffice.
Scratch Tracks: optional, but very helpful
Typically, a band will have a rough idea of their material mapped out ahead of time. It could be in the form of sheet music, a MIDI sequence, or a simple recorded demo from a previous rehearsal. Though this is usually worked out before the band even books the session, having this material serves a few purposes.
- The producer & band can work out any last-minute changes to the song: different tempo, key, feel, arrangement, etc.
- The engineer can use these tracks as a guide during the tracking process.
- Amateur musicians typically aren’t used to playing alone, with a metronome, or without the rest of the band. By recording them within their comfort zone, they can later overdub a better performance over that track.
If they don’t have useable demo tracks on hand, record some if time permits. When recording scratch tracks, we want to record the best (most comfortable) performance we can, while maintaining some level of isolation & control. Typically, this means:
- Recording the entire rhythm section at the same time, and in the same space.
- Recording to a consistent click (metronome) is ideal, but if the band can’t pull it off, counting off at a consistent tempo from take to take will suffice.
Setting Up for Drum Tracking & Scratch Tracks
Once again, we need a good setup sheet. At a minimum, we need four microphones on the drums: kick, snare, and two overhead (stereo) mics. If we recorded scratch tracks separately, we have more room for close drum mics & room mics. If we need to capture scratch guitar DI’s, factor this in. Otherwise, a full drum set can consist of:
- Two kick mics: a sub mic & a dynamic mic inside on the beater.
- Two snare mics: top & bottom.
- Rack & Floor Toms: one each (typically 2 rack toms, 1 floor tom).
- Close cymbal mics: typically one each: hi-hat, crash, ride, etc.
- Stereo overhead mics: typically X/Y pattern, 4 feet above the kit.
- Room mics: typically a stereo pair above & away from the drums.
On a standard 5-piece drum kit, this means we are using up to 14 microphones, with two left over for other things: more drums & cymbals, or scratch guitars. This may seem like overkill, but this serves several purposes:
- We can always mute tracks if we don’t need them.
- Individual isolated drums allow for more options during editing & mixing.
- With isolated hits, Sound Replacer & other plug-ins can be used to trigger a new drum sample: complementing or replacing a terrible sounding drum with a better sample, or adding articulation & definition to your recorded tracks. More on that during the editing phase.
During our setup, we can also utilize some of our outboard gear, namely the compressors. Light compression in the signal chain will help keep the kick & snare levels consistent, while helping to prevent clipping or distortion to an extent.
Beyond this, if we need to record scratch guitars at the same time, Direct Boxes and well-baffled guitar amps will suffice. If you have to prioritize anything, focus on capturing good drum tracks here. At this stage, guitars are just there for reference: everything else will get overdubbed later.
The recording process should be fairly routine by this point. Recording different takes can be done in one of two ways. On a tape machine, recording is linear: the new take has to be recorded after the older one (or on a new track) so we don’t erase over anything: for example, Take 1 could start at 00:01:00, Take 2 is at 00:04:15, etc. While Pro Tools can do this just like a tape machine, there is another way.
Let’s say we have a guide track, or we have mapped out meter & tempo changes inside Pro Tools. Rather than record after another take or make new tracks (which can use up our system’s resources), we can make a new Playlist, also known as a “virtual track” on the existing track. This lets us record new audio on the same track at the same time reference, without erasing our old take or eating up track space & system usage. On top of this, this makes comping tracks much easier. Let’s say the drummer messed up the verse on one take, but nailed it on another. Rather than try to select the good section and drag it across the session, we can switch between playlists, copy the part we want, and paste it at the exact same time reference into another playlist. The only catch to this method is it requires playing to a click, or the same reference track. Beyond this, recording is nothing new.
After you get a few good takes in the can, it’s time to edit. Whether you’re working on scratch tracks or the real thing, the same methods apply. The first step is called comping: combining the best parts of each recorded performance into one new master performance. With the regions/clips in place, Crossfade the clips together, highlight all of the clips on that track, and select Edit > Consolidate Clip to make one contiguous clip.
Rule of thumb: if it’s not broken, don’t fix it until it is. Depending on the genre & the ability of the band, comping may be the extent of your editing. If the drums need to be tighter or quantized, a function called Beat Detective can analyze and quantize audio. Select the regions you wish to edit, and select Event > Beat Detective. Before you start, copy the regions you want to edit, and paste them in a new playlist in order to save your original tracks (for safety & posterity).
A few notes about Beat Detective:
- Older (LE) versions of Pro Tools can only do Beat Detective on one track at a time. Not a big problem, as we typically want to fine-tune each track anyway.
- Beat Detective works better on tracks recorded to a steady click.
- Because it chops up & moves the audio clips, isolated tracks are easier to work with: if you’re only using the four-drum mic technique, you probably won’t be able to quantize the whole kit properly.
With Beat Detective open, select Bar/Beat Marker Generation under the Operation section. Click Capture Selection & alter the settings as needed (4/4, contains 16th notes, etc), select Sub-Beats under the Detection section, then click Analyze. Purple lines will appear in the region. Slide the Sensitivity slider around until only your isolated hits are marked (exclude the bleed from other drums). Try to get it as close as possible. If there are any mistakes left, fix them manually.
With the hits marked, click on Clip Separation, adjust the Trigger Pad to 10ms (double the value of the Crossfade Length under the Edit Smoothing tab – more on that in a minute) and hit Separate. We’re halfway there.
Select Clip Conform tab, and click on the Strength slider: this allows you to adjust how far Beat Detective will quantize the clip. 100% is snapped to the grid, like a drum machine, whereas 0% is unchanged: the drummer’s original performance. Find a happy medium between the two to your liking if you want to keep some of the drummer’s original feel. Exclude Within slider will ignore notes within a certain distance from the grid. Swing, obviously, swings the clips. For simple quantization, leave these unchecked. Click Conform to move the clips to their new grid. Before we play the newly quantized section, select Edit Smoothing and click Smooth to add Crossfades between clips. If there were any mistakes during analysis, fix them manually, and you’re done. Rinse & repeat for each track. Once again, select the newly quantized clips and consolidate clips.
Another trick is called Strip Silence. This function works like a Noise Gate, stripping away dead air & excess noise by removing unwanted parts of the clips. Select Edit > Strip Silence or press Command U to open the Strip Silence window. There are four sliders: Strip Threshold determines how loud the level needs to be in order to be saved. Min Strip Duration determines how long the gate stays open. The Clip Start/End Pad sliders determine how much of a buffer is left before/after the audio. Leave room so you don’t clip off any of the signal you want to keep. When the adjustments are right, click the Strip button. That’s it.
Our last option is an Audio Suite plug-in called Sound Replacer. Select a clip, copy & paste into a new playlist. Open the plug-in & click Update to show the waveform in the window. Click one of the floppy disc icons below the sliders, and pick a drum sample out of the library. Drag the slider until a yellow line appears on all of the hits (avoid bleed, if any), and click Process to resample the track.
There is one noticeable problem when using samples instead of recorded audio. While kick drums are usually always simple isolated quarter & eighth notes, snare rolls sound different from quarter notes. If the drummer is playing anything slightly complex, like a fill or delicate snare work, the samples alone can sound obviously artificial – a drum machine instead of a drummer. Because of this, it may be a better idea to duplicate the original (edited) tracks, and run their duplicates through Sound Replacer. By blending the two tracks/sounds together, you will typically get a more natural-sounding performance, with added definition from the samples – More on that during the mixing phase.
With the drums edited for timing, we can begin overdubbing the rest of the instruments. Guitars, specifically bass guitar, typically come second in the process. As part of their complete rig, guitarists will have a selection of guitars/basses for different tuning/tone, effects pedals (distortion, wah, etc), and a specific kind of amp (solid state, tube, etc). Since they invested untold hundreds (or thousands) of dollars in their signature sound, we want to capture as much of this as possible, even though we may disregard them and process a clean DI signal anyway. Just don’t tell them about it. Guitar rigs may not require as many microphones as the drums do, but we can still go overboard:
- A Direct Box before the effects pedals
- A Direct Box after the effects pedals, before the amp
- An assortment of close-mics on the amp cone: dynamic, ribbon, or condenser
- And a room mic or two, some distance away from the amp.
If you have to prioritize something, a clean direct signal is the most important thing to get: we can always process the signal through our own effects, or reamp the signal. After getting levels and recording a bit, we can use this time to compare microphones and decide what we want to keep. Once again, we can always mute tracks that we don’t need. Keep an ear open for a low buzzing sound from the guitar/amp. If you hear it, press the Ground Lift switch on the direct box & amp.
On stage, guitarists are used to playing a part, stomping on a pedal & chugging on. For more control in the mix, try to adjust & record these different parts as if they were separate instruments: clean guitar, distorted guitar, solo, verse, chorus, etc.
With all of that in mind, record bass & guitar part-by-part, overdubbing on top of the guide track or newly edited drums.
Comping is more of the same: grab the best parts from each take, and combine them into a new master take. If we have a clean DI signal recorded, we can run the signal through any assortment of amp-modeling plug-ins: Amplitube, Ampeg, Amp Farm, Sans Amp, etc. Each one provides an assortment of emulators for various guitar amps: tube, solid state, different cone/cab sizes, effects, and microphone placement.
Though you will most likely do more of your guitar tweaking during the mix, amp modeling lets you either replace a terrible sounding amp, or change the tone of the guitar tracks entirely. If the guitarist came in with a p.o.s. combo-amp, you could emulate the tone of a Marshall stack, and so on.
Keys: Acoustic Piano & Synthesizers
Recording synthesizers are almost identical to guitars. Direct boxes & amps (if any) follow the same principles. With the exception of vintage analog synthesizers, all keyboards have MIDI ports. Through these, we can capture the player’s performance in the form of MIDI data onto an Instrument Track, as well as audio on an audio track. Run your MIDI cables to your interface, and route the signal inside Pro Tools through the MIDI I/O – If you can’t see it, click on View > (Mix or Edit) Window Views > Instruments. Typically, you can get away with just having the input set to All. If you have multiple keyboards, or want to control multiple Virtual Instruments, you can route the MIDI input/output accordingly. With that MIDI information, we can later edit (quantize, etc) the data and reroute it back through the keyboard, or any other virtual synthesizer within Pro Tools.
Acoustic (grand & upright) pianos are a different matter. First of all, since they are a stringed instrument, they can get out of tune. In fact, finding out after the fact that your piano was out of tune can effectively ruin your recording, your harmony, and your entire project. Though the facility will usually have their pianos tuned regularly, check before the session. If it’s out of tune, find one that isn’t.
A piano’s construction provides a certain level of isolation. The lid can be opened slightly to allow for close-micing, while the wooden beams below/behind the soundboard can provide an overall isolated pocket. Both of these positions are great when you have to mic a piano along with the rest of the band, but if we can overdub it by itself, there is a better way.
Open the lid all the way, and place some condenser mics high above the strings. This gives the natural overtones time to develop. The middle of the strings provides a sweeter, mellow sound, while the ends of the strings have a more brittle sound. You can also place microphones underneath the center of the soundboard, above by the hammers, and/or by the end of the soundboard. Experiment to your liking.
After comping, you can edit the MIDI information, rearrange it if need be, rerun it through the keyboard, or control a Virtual Instrument. Like amp modeling, this lets you effectively “re-amp” your keyboards through whatever instrument you wish, from a virtual piano to a Kontakt library. If this is the case, put your desired virtual instrument on one of the Inserts (while on an instrument track) and pick your tone.
Though the keys will typically play melodic/harmonic parts, keep in mind Synth Pads can cover a wide frequency range, which can muddy up your mix: frequencies that you would want to set aside for kick, guitars or vocals can get buried.
Singers need two things to record: a rough mix and reverb in their headphones. They may want everything else from drugs & mood lighting, to candles & incense, or worse to get in the mood. Oblige what you can, have the producer deal with (or deny) everything else. While you want to do whatever is reasonably possible to nurture the best possible takes out of a singer, you (at least the producer) still need to manage the singer well. If left to their own insecurities & indecisiveness, a singer could have you trapped in the studio for hours on end, recording take after take, even if you already have plenty of good material in the can. Beyond this, singers & horn players typically can’t sing or play for as long as the rhythm section can: they will wear down or go hoarse or “blow out their chops” much quicker, so you will typically have to space out your tracking over several days. Keep these in mind before you start tracking.
Some singers already have their own microphone preference: Shure SM7B, Neumann U87, etc. For everybody else, we typically start the vocal process by setting up a number of microphones: condensers, ribbons, dynamic, and so on. After recording a performance and comparing the sound, we pick out our favorite mic and record the rest with that. Any number of factors can dictate what mic we pick in the end. A light jazz singer has a completely different style & dynamic from a rock, metal or punk screamer. Where an expensive condenser might be great on most vocals, a cheap dynamic mic is better suited for others. Don’t forget to use a Pop-Filter and/or a Windscreen. These serve a few useful purposes: they prevent the singer from “eating the mic” by keeping the singer’s mouth at a minimum safe distance, and they keep gusts of breath from overloading the microphone –pops & plosives.
We can also use some outboard gear while recording, – all aimed at reducing noise, while maintaining a strong, clean signal – these can also be done after the fact with plug-ins, but it is typically best to do this before the recorder, if at all:
- A De-Esser when adjusted properly, will help tame sharp S & P sounds, called Sibilants and Plosives. It temporarily gates high frequencies when they get too loud, while remaining open for natural overtones.
- EQ, more specifically a Hi-Pass Filter or the Roll-Off switch on a mic/console, can be used to remove/tame low frequency noise, while allowing the high frequencies to pass through. For example, if you know a singer’s lowest note in a song is A = 440hz, everything below 440hz is noise.
- Compression: We want to maintain a relatively strong but clean signal, without clipping or distorting. Light/minimal compression is used for safety to prevent severe clipping. While we can use it to make our vocal track louder overall, we want to avoid saturation while recording. We can always add more compression in the mix later. We want a clean vocal recording first.
When tracking vocals, we are typically more concerned with capturing the best feeling & performance from the singer, not necessarily rhythmic & pitch-perfect accuracy. While we can (and will) go through the pitch-correction process, we can’t really add emotion or expression after the fact to a vocal track.
Editing Vocals: Tuning
Once again, we comp our takes into a new master take. If we have to make a choice, err on the side of a clearer, more emotional performance, rather than a pitch-perfect melody. If needed, we can tune the vocals. These same methods also apply to any monophonic/melodic instrument (like our bass tracks). With the exception of the new Melodyne D.N.A. (Direct Note Access), computer software cannot properly analyze or identify harmonic material. While it can pitch audio up or down, it can’t turn a C Major chord into C minor. It can, however transpose a C Major chord into B Major, and so on.
Tuning typically comes in two forms: a simple plug-in, which you can “set and forget,” and more elaborate plug-ins for more precise editing. On the simple side, we have Auto-Tune, and several others. These work well for (fretted) bass guitar, and horns, when needed. On fretless instruments (upright bass), and especially vocals, they tend to produce a very obvious pitched digital effect (sometimes called the “T-Pain Sound”). Unless you actually want that effect, you should do your fine-tuning in software like Melodyne. Before editing in Melodyne we need to finish editing the isolated track we want to retune: comp the best takes, delete any dead air or glitches, etc. Save a copy in a new playlist. Now we can tune.
Fire up Melodyne and import the tracks. The Melodyne editor analyzes the track and displays it as blobs of audio in relation to meter, pitch and (sometimes enharmonic) key. Click on View and select everything. On the left side, we see a grid of notes: one line per pitch. Right click on this to change the scale reference. Up top, we see the audio in notation form, as well as our transport & toolbar. Change the tempo/meter settings as needed. The bulk of the editor is the actual audio itself. With all the view options selected, we see the orange audio blobs in relation to pitch & beats. The grey boxes & outlines show where Melodyne thinks the notes should be: on pitch.
The quickest (least accurate) way to edit in Melodyne is to select everything with the Main Tool (looks like an arrow), and click Correct Pitch and Quantize Time. There are a few potential problems with this. Melodyne corrects/quantizes where it thinks the audio should go. This is also where your music theory/harmony work comes in handy. Depending on the performance, Melodyne may misread the audio: for example, an eighth note slur from E to F# (two separate notes) can be misread as one F quarter note – out of key & out of rhythm. This is why we edit manually.
Use your ears, and with the other tools, separate the notes (Note Separation Tool), quantize them (Timing Tool), and edit the pitches into place (Pitch Tool). Just keep in mind that there may be non-diatonic tones, and the rhythmic “pocket” may be ruined if you quantize the timing: a major factor in jazz music.
With the track edited to your liking, save it as an audio file, and import into Pro Tools. Spot it back in place of the original, and you’re done. Remember, this is the same process for editing any melodic track. Editing a fretless bass in a jazz track will make a huge difference, tightening up the overall harmony in the track. Only the newer versions of Melodyne (with D.N.A.) can edit harmonic content.
Editing the Overall Project
With everything tracked & edited individually, it’s time to edit the entire project in context. Even though we comped & quantized our tracks as we recorded, there is still the possibility that instruments will be a little bit off from one another. Though we still want to maintain the feel/pocket of the musicians, it’s a good idea to line up any unison hits/stops that the band plays together. The same applies to sectional pieces, unison lines, or anything meant to be played together. Zoom in and look at the waveform in the edit window. You can actually see the wave peaks out of alignment. Nudge the region forward or back with the + and – keys on the number pad until the beat is aligned. If we recorded/edited to a quantized grid, it may be better to do this editing in Slip Mode, rather than Grid Mode for fine detail work.
One major problem with multiple audio tracks is phasing: phase cancellation, beat frequencies, etc. This is destructive interference caused by audio being slightly (or perfectly) out of phase with other audio tracks. Typically, this occurs when two separate mics are on the same instrument: kick, snare top/bottom, overheads, guitar cabs, and so on. Fixing this is fairly easy. Zoom in until you can see the wave peaks, then Nudge the clip forward or back with the + and – keys on the number pad until the waveform is aligned. If the clip is perfectly out of phase when they’re aligned, select the clip, click Audio Suite > Other > Invert, and then click the Process button. Some EQ plug-ins have a Phase Reversal button which performs the same function, but it’s better to do the first method.
Next, cut off any lingering bits of “dead air,” or unwanted noise from your audio regions. This is like doing “strip-silence” manually. Use your Multi-Tool & Trim the dead space in front of/behind each clip. Leave the decaying audio “ring-out” at the end of the song – we will later Fade Out the ending to an extent, but we still want the cutoff to sound natural, not abrupt. With everything trimmed, add Fades to all of the clips: Fade-In at the front, Fade-Out at the end, and Crossfades where two separate clips meet. Aside from making everything smoother, this also prevents unwanted pops & clicks in some plug-ins (especially reverb units), which can ruin your recording. While you can preform basic fades directly with the multi-tool, you can create more advanced crossfades by highlighting the clip ends you want to fade & selecting Edit > Fades > Create or Command F.
With everything trimmed & in place, there are a few more steps to go before we finally start the serious mixing. First off, let’s do some basic routing & preparation. Create several Stereo Aux Tracks, and name them Drums, Bass, Guitars, Vocals, Reverb, etc. Route the Output of your various tracks to their new respective busses drum tracks to Drums, etc: these now function as a sub-master for each section of instruments. For the Reverb aux, set its Input to any unused Bus. Hold down the Option button, and on the first track, assign any unused Send to that bus you just selected. Holding down Option will assign the same action to every track in Pro Tools. You have now assigned an adjustable send (like an Aux Send on a console) to the Reverb channel. If you’re only going to use one reverb/Send for this project, select View > Sends A-E > Send A only to display a visible meter on that track.
If we have any tracks that we don’t need anymore (scratch/guide tracks, excess mics, etc), rather than delete them entirely, we can select them, right-click on the track name and select Hide and Make Inactive. This way, you save any information you may need to salvage later. If you need to make the track reappear, right click on the track’s name in the Tracks list, right-click and select Show and Make Active.
We can also do a bit of automation in this phase. After hearing everything in context, we may decide we want to cut out certain parts, troublesome notes, etc. Rather than delete these parts outright (we may find out that we really needed them later), we can mute these regions. Highlight the undesirable parts, and select Edit > Mute or press Command M instead.
Starting to Mix
Panning moves sound around within the Stereo Image (left & right). Most of the time, the extent of this just means moving the knob once & forgetting it. Typically, we recreate the drums’ position, either from the drummer’s perspective, or ours as the viewer: Kick & snare in the center, toms & cymbals spread out to the left & right, overheads panned extreme left & right, etc. Everything else is usually panned to where the performer would typically be on stage: imagine the band on stage, and recreate that positioning. If we need to change this around during the mix (special effects, etc), we can do some more automation.
In the edit window, select a track, and click on the dropdown list (default is waveform). As you can see, there are many automatable options, but select pan for now (the same procedure applies to the other options). The line down the middle of the track is the current pan position: middle is center, up is left, down is right. Use the Hand Tool and click twice on the line to make two separate points. Leave the first dot in place to keep all of the automation before it unaffected, and then move the second dot to your desired position. A vertical line between points makes an immediate change; a diagonal line between the two points makes a gradual change.
Specific EQ varies widely depending on musical genre, instruments, style, etc. However, there are a few tricks that apply to any mix. Individually, an instrument can have a great, full-range tone. In context, it could be cluttering up the mix by covering up other instruments. For example, our kick drum has a low-end thud, midrange punch, and a high-end click. It sounds great by itself (and we would want it to cut through the mix), but there’s a synthesizer pad or a bass guitar muddying the sound. There are two tricks we can use here.
We can use EQ to cut the problematic frequency ranges out of other instruments. A high-pass & low-pass filter combined can notch out a very narrow frequency band (like the bottom end of our kick). Roll-off will gradually cut the low-end bleed off of our tracks. This is good for cymbals & overheads, since it cuts out the unwanted low-end from the other drums. With Subtractive EQ, we lower & cut unwanted frequencies instead of boosting everything else, which essentially makes the overall track louder. Typically, it’s better to cut frequencies than it is to boost.
Side-Chain Compression uses a separate signal to trigger a compressor on a different track. With this, we can use the sound of our kick drum to compress a synth pad (or something else). Assign another unused bus/send on the kick channel and raise the send volume. Throw a compressor on the synth pad, and assign the compressor’s key input to read from that bus. Adjust your settings, and the compressor will activate whenever the kick drum hits. This may create a bit of a pulsating effect on the track, but it will allow the trigger track to cut through more.
Noise Gates work well on tracks with a lot of bleed, especially drums. Louder hits pass through the gate, which then closes to shut out the quieter bleed. If the drummer is doing delicate snare work, process the track section by section with the Audio Suite noise gates. Do this if you didn’t use strip silence & sound replacement.
Compression can raise overall volume while reducing loud peaks. Musically, this is like turning piano (soft) and forte (loud) dynamics into mezzo piano and mezzo forte. We already used some outboard compression during the recording process, but we can add more through plug-ins as needed. Extreme compression can create a noticeable saturated sound, or even a pulsating sound in some cases. If a track needs a lot of compression, it is usually better to use two compressors with light compression: it’s a bit like lifting a heavy object. One person will strain noticeably, while two people could lift it with less trouble.
Parallel-Compression blends an uncompressed or lightly compressed signal with a heavily compressed version of the same signal. Gradually mix the compressed signal back in until the track sounds fuller/punchier without sounding louder. This way, you can preserve the track’s dynamics while making them sound more present. This technique works on just about everything, even in the mastering process.
Brick-Wall Compression is extreme compression that obliterates a track’s dynamics while boosting its volume to the limit. We typically see this more on the mastering end, but it can be used in the mix as well. It is used a lot on tracks with no real dynamic contrast (like a soft & subdued verse against a loud & energetic chorus). It uses loudness instead of contrast to try to hold your attention – like “smashing your head into a brick wall.” It can get very boring very quickly, and it’s often fatiguing to listen to.
The other major problem caused by brick-wall compression is ear fatigue. After a while, like hearing a jet engine or another extremely loud noise, it hurts to listen to it for an extended period of time. Typically, you need to leave the room & rest your ears before you can hear properly and mix objectively.
Another technique that we employ throughout the mixing (and mastering) phase is called Reference Monitoring. This means actively listening to and comparing several different kinds of speakers as we mix – a system with a subwoofer for full range, large far-field monitors, small near-field monitors, even headphones or an old boom box. What may sound great on one sound system could sound lacking or terrible on another, so we find a compromise that works well on any system.
Specific volume, EQ, and the like vary widely depending on instrumentation, tone, genre, personal preference and more. Despite all of this, a few principles & techniques still apply. The lead instrument or vocal typically sits right on top of the mix. To check this, adjust your mix, turn your control room volume all the way down, and then gradually raise it back up. The lead track(s) should be the first thing you hear.
Earlier we touched on Automation – programming our mix (volume, pan, etc). While we can program all of the options manually by making points with the hand tool, there is another way. With a Control Surface (like the FaderPort or C24 console), or by clicking & riding the fader, we can program automation in real-time. Beneath the track’s I/O section is the AUTO (automation) mode. Off ignores automation: faders and the like will stay where they are until you move them. Read follows and performs the programmed automation. Write lets us program our automation in real-time: wherever the faders are, moving or not, Pro Tools will record that position in time. Touch works like read-mode, but only writes while the fader is touched/moving. Latch does the same, but will keep writing until you stop playback/automation. Switch the desired track into the proper mode, press play, and move your fader. Keep (re)doing this as needed until your mix is finished.
In the end, make sure you have a good overall volume level without clipping on the master fader – no distortion. Give yourself a few decibels of headroom to stay on the safe side. With everything mixed, we can move on to mastering.
Typically, we mix a song (multiple tracks) in context, and master an album (multiple songs) in context. Though there are many vague definitions for the mastering process, it always results in turning a mixed song into a finished Master copy, from which all other copies will be made. When mastering an album, we balance the levels of each song for consistency, so listeners don’t need to adjust their stereos between songs. There is much more to mastering than that.
There are two ways to approach this. The first way is by using plug-ins directly on the Master track in Pro Tools. Depending on how big the session is, this may drain too much of your computer’s processing power, as mastering suites like T-rackS or Ozone require a lot of resources. The other way requires bouncing down the song as a stereo track, and opening it in another session. Bounce the song down to a 48kHz, 24-Bit (mastering quality) WAV file, and open it in a new session. If you are mastering an entire album, having all of these tracks in one session can make mastering in context easier.
Though we spent a lot of time listening to a song in context with itself (balancing levels, changing tone, and so on), we do more of this in the mastering process in a different way: balancing frequencies, sections, and so on. We are already familiar with EQing instruments. EQ while mastering is more of the same: balancing out lows, mids, and highs against itself. If you are having trouble locating specific frequencies, try using a scope or a frequency analyzer on the master track.
Multiband Compression lets us compress specific frequency ranges. Typically, the low (bass) end may have an overpowering pulsing effect. With a multiband compressor, we can focus specifically on the problematic frequencies, and compress it. We can use Parallel-Compression here as well, in order to give the track a bit more punch. Duplicate the track, heavily compress the copy, and mix it back in. Parallel-compression aside, try to avoid severely compressing the master track as a whole. We want to preserve a sense of musical dynamic contrast, and we especially want to avoid a mix that will cause ear fatigue.
Typically, our loudest peaks will be right at the meter’s limit, without peaking. Just to give ourselves a little bit of safety, we can create a noise ceiling with plug-ins like Maxim. Click an insert on the master fader, and select Multichannel Plug-Ins > Dynamics > Maxim. Slide the Ceiling slider down to -0.1 dB. This should prevent any audio from crossing over the threshold and peaking/clipping. While you’re at it, click Dither, Noise Shaping, and 16-Bit resolution. These settings will help us in our next step.
One of the problems with digital audio comes in the form of errors & digital artifacts. Though they can be heard with highly trained ears, (if you know what you’re listening for) most listeners don’t notice them. We bypass some of these by recording at a higher quality, and bouncing-down to a lower quality. Higher qualities have more data: a longer stream of numbers (mastering quality has 24-Bits of data). Since the errors are usually at the end of the number chain, bouncing down to a lower quality (16 bit, 44.1 kHz – CD quality) ignores those number errors. We help this process along with a Dithering plug-in. Dither converts 24-bit audio into 16-Bit, among others. If that didn’t make any sense, don’t worry.
Mixdown & Burning
With the track mastered to our liking, it’s time to bounce it down one final time. This time, we are bouncing down to CD quality: 44.1 kHz, 16 Bit. We can go lower with varying forms of mp3’s and the like, but CD quality is as low as we want to go while preserving audio quality. For that matter, any stereo or media player can play an audio CD – we can’t say the same for other formats.
While we’re on the subject, just about any media player can read & play WAV & AIFF formatted files. Other formats do exist, but they may not work as well. Other more compact lossless formats like FLAC have been showing up in recent years, but they do have their drawbacks. In order to play the file, your media player needs to unpack the file, if it can even read it to begin with. It’s a bit like having to open an envelope (FLAC) instead of instantly reading your mail (WAV/AIFF). Given the abundance of storage space & the common usage of WAV & AIFF formats, you’re better off sticking with one of those two. WAV & AIFF are more or less interchangeable, and a standard CD-quality WAV/AIFF file will sound better than a lower-resolution MP3.
With the tracks bounced, we need to put them onto a CD. While media players like iTunes can do this, they often add their own compression onto the track. Software like Toast is designed specifically for creating CD’s and DVD’s. Launch Toast, select Audio tab, Audio CD from the drop-down menu, and drag your tracks into the window. From here, you can give your CD a title (album title), name your tracks, fill in artist info, and so on. You can even pick the track order, and how much of a gap is inserted between tracks. When you’re ready, burn your disk. As our final step, play the finished CD in a CD player to see if everything is working properly. If you need to change anything, make your corrections and burn another. You’re done.
You have just recorded, edited, mixed, and mastered a song from start to finish. Find a new project and repeat this process on your own for your final projects.