- DAW 101
A DAW or Digital Audio Workshop is a computer application that processes sound for replication. DAWs often come with a host of included plugins, otherwise known as effects or FX.
DAWs synchronize synths to a clock to allow them to generate time sensitive musical information. Plugins cover everything from synths to audio FX to in-app cosmetics. Midi is the aforementioned musical information. When put to work in a contained environment, a producer can orchestrate anything they set their mind to.
- Plugins
Plugins go by a few different names. VSTs(Virtual Studio Technology) on Windows, AU(Audio Units) on Mac, AAX for Pro Tools, so on and so forth. Plugins can do a variety of things from synthesizing noise, manipulate phase, automate & script volume, perform midi
Plugins designed for apple products are coded differently than those made for Windows. The same goes for Linux. When given a choice between VST classic and VST3, use VST3. If you do not own Pro Tools, do not waste hard drive space on an AAX version. You will never see it and they do add up over the years.
- FX Chains & Templates
FX Chains emulate physical hardware racks. They are, and always were, user made configurations designed to reduce the time needed to achieve a desired outcome. For some, that's clean loudness. Others want a dynamically flat signal. FX racks can even be used as a one knob solution to the slimiest or most ethereal sounds man could achieve. Take the coveted Ring Mod Sidechain for example. This is a 4 step process. First, audio must be recieved from the Kick & Snare(K&S) channel with 0 input latency. Second, that audio must be inverted using phase cancelation. Third off, the ring mod needs to hear the waveform of the processed K&S to invert the waveform of the pre-sidechain master. 4th, all routed audio needs to be applied to the master channel in order, in time, and with correct routing.
This is a lot of work to do at the beginning of every project file. We save it instead as a Patcher, Audio Effect Rack, or Reason Rack preset. Everything from the generator to the notepad VST can be saved in sequential order, to be dragged and dropped into any project file.
Where an FX chain is meant to recreate a highly randomized or difficult to emulate sound, post-processing is the part that removes noise early, handles the EQ at the specific parts where it will do the most damage, and assures any loudness is achieved in the correct order for the most dynamic range, while simultaneously achieving the greatest possible loudness.
are more of the same, bt bigger picture. A Template is a whole project file layout, ideally optimized to have your go to settings and presets loaded at startup. In the case of my own template, I have left a lot of inconveniences in to force myself to make hard choices when the song picks up steam. This resistance is calculated. My synths are mapped to the macro, forcing me to script my synth, rather than spew out midi triggered garbage on a repeating LFO. The drums are just barely tolerable, but are so unapologetically bad that they will ruin anything I release if they are kept the same. Best option in this scenario is to using something that will offend you in a full mix if left unignored. The same can be said for all of the quality of life things like a broad tuning of E on all EQs and multiband compressors, or a bare-bones "phat rack" on my primary dubstep synth channel. I know I will not change most of this unless I am changing the key of the project. The same goes for anyone who adopts my workflow.
- Recording
This section aims to catch everyone up on decades of well defined and heavily tested standards for processing raw audio.
Please refer to:
Gain staging a microphone or line-level (guitar-cable) audio signal is about balancing silence. For most sources, we set one's own input gain to -10 decibles metered, and to record in mono. -10db is loud enough to combat hardware noise, but quiet enough to prevent baked in distortions. Any inputs quieter than -30db allow for artifacts from hardware and software to bleed into the recording. Those artifacts will be picked up by compression no matter how smart or capable your mixing engineer thinks they are. It cannot be removed once its in the wav. Take the extra day to record these parts again and refuse services from anyone who cannot explain why this is an issue. They are uneducated and they are way out of their element.
Now that we have a basic grasp on what makes a studio grade recording, lets knock out bouncing and rendering. When audio is synthesized, there is more potential for crazy errors. Borrow from the above.
- Highlight and remove all FX(ctrl/cmd+x) before flattening.
- Only keep unison and random phase if it is necessary for the recording's character.
- Flatten audio above -30db. Do not flatten anything that reads -60db to -72db at operational volumes. A software synth can still create artifacts at lower volumes. Turn the master output up.
- Re-apply any post processing (ctrl/cmd+v) after flattening is complete.
- Do not be afraid to edit the raw audio by selecting louder or more satisfying chunks of the render.
- Performing a piece correctly is more about presenting a product that is worthy of a listener's attention, than it is about pride. Respect your fans and they will respect you.
- Once a song is published, there are no do-overs. Tune your drums. Change your strings. Make your vocal delivery personal.
- Practice your own parts to a metronome to avoid frustration in the booth.
- Expensive cables and microphones do not clean the dead skin and oils from your instrument. Buy fresh strings for the clearest audio quality.
- Is your signal red? It is clipping.
Everything else in the chain will carry your clipping noise up to the master output until it is present for our ears to bask in until the very end of time...
Turn it down.
- Mixing
The recording is now in the DAW. All performance flaws and imperfections birth the soul of a recording.
This really does need an explanation. I thought it didn't, but I am only allowed to be wrong once.
Mixing is the act of balancing every sound, normalizing the volume of every performance, adding volume to
things that are buried, removing volume from things that cannot be heard because - hear me out - if you can't
hear something, it doesn't exist and if it doesn't exist, its distorting your master.
Mixing is what allows a song to be understood as music. A good mix is intentional. The listener will interpret everything correctly without needing to be told what to listen for, because it is your job to show, not tell. A bad mix will leave the listener deeply underwhelmed or confused about who or what is delivering the story in a song.
The challenge of mixing a song down is that there is no creation left. If a guitar part sucks, you cannot mix a good performance into the song. It needs to be recorded again. If there are no adlibs or harmonies in the vocals, it cannot be mixed in. These parts need to be recorded. Same goes for a bad drum loop. Producing a good song is the first line of defense for every bad mix engineer.
Recording is not mixing. Adding a riser is not mixing. Adding a phaser to the bass guitar is not mixing.
Watching the mix engineer and then saying "i think the drop could use more bass" is not mixing. Picking a
new snare is not mixing; but the right snare will make the mixdown significantly easier. As mentioned before,
writing a good song is the end all be all.
You are not ready to mix a song down until the song is finished.
Assuming every recording sounds good without heavy editing and there's no unintended vague emptiness in chunks
of the full song, there's a few tricks professionals use. We will tackle the common ones for each genre.
I have put this off for a year for a good fucking reason. bare with me because this might take a few weeks.
- Old-school: No sub bass in the entire instrumental. Sub bass only comes from the kick.
- New-school: Get your volume from stereo and mix everything a little lower to allow headroom and prevernt sub-harmonic distortion on the master.
- Sidechain the vocals into the instrumental.
- Turn vox down -6db and run em 50/50 to the sidechain send and to the master.
- Hard Pan vocal harmonies 100% left and 100% right with the main in mono.
- Splice all breaths out of the vocals
- EQ(cut) like a menace. That 808 will only pay your bills when you let it work.
- Throw an obscene amount of resonance on your snares to get that glassy clink.
- Lead your sidechain in a little before the snare plays. This will create silence in the instrumental. That silence allows the snare to CRACK when it plays.
- Here is a comprehensive run down on how industry leaders record their guitars. Enjoy.
- Record a minimum of 2 perfect parts and tighten them up with splicing.
- Record additional palm mute stabs for "chugs".
- EQ like an idiot. Seriously. No bass allowed. Shelf your mids out.
- Allow your kick to choke the sub out of your bass guitar when sidechaining.
- Mix the bass loud.
- Play everything legato. Do percussive notes on purpose.
- Use the shittiest recording devices you have accessible. Use twitter to share How offended when anyone notices you're actually just poor as shit.
- Record in a bare closet. Or a bathroom. Just disregard the listener completely.
- Do everything in one take. Bonus points: do the ukulele and the vocals on the same recording.
- Lead your snares in with reversed white noise
- Swell the sub bass into the kick
- If the hihats sound weak, use a white noise loop to add sustain.
- G-CLIP or MISHBY on the master. One is a soft clipper. One is an eldritch abomination with a great distortion algorithm.
- EQ, then compress, then distort for sheer loudness.
- Do not use a limiter anywhere, unless it is on the master.
- Less is more. Write with your desired dynamic range in mind.
- Match every volume to your kick and snare. These two set the average loudness of every sound in your song.
If you're ready to begin mixing down your song, take a moment to skim this page. I will teach you all the vocabulary and science behind mixdowns that are often twisted and mangled to better fit someone's $300 course. The secret? You don't know what to google to get all the info for free! I'm here to put a damper in that grift.
A couple quick things to look out for in your friends' projects:
- Panning the bass - Its not cute. You will give someone a migraine.
- Leaving the kick and snare any more than 1db louder than the mix. Bad.
- Cutting the sub out of the master - Why? Let the listener do that on their own.
- Mixing to [-0db averaged]. This is an experimental practice with no practical function. It can only be achieved through aggressive distortion and hurts to listen to.
- Only listening on one set of speakers - Your Holy Chalice may get you free meth, but everyone listening on car speakers can hear things that you don't. Give it a spin, Kyle. You might hear something new.
If you see any of this dumbass behavior play out, you have my endorsement to smack them right upside the back of their head. They obviously don't take to instruction.
- Vocals
A rule of thumb when processing vocals is to concentrate, rather than cut. The human voice has a built in bandpass filter called "lips". Lips are a device our ancestors have used since the dawn of time to captivate and enthrall. Too much EQ will filter their lips off and that sounds boring.
Now that we understand our instrument, first steps should be a De-esser. Your De-esser reads frequencies where human voices generally produce the most unwanted noise. In other words, we keep the pretty air noise, and remove all the ugly kuh-siss-puh-tohh sounds. It is one of those things that you dont know is on until you take it off. Invest in a good one.
Back to how the mouth filters like an EQ. We will use an EQ to prevent the 3rd effect from smooshing the nuance of human enunciation. Our Pre-EQ will take any unnecessary bass out of the recording. Bass can enter a microphone from a number of sources and many of them will not be audible until compressed.
In most modern music stylings, a reliable multiband compressor will do so much work. Your vocalist's microphone provides its own EQ. Multiband compression will bring out the $40,005 audio quality at no added cost. Try to keep it under 40% and add no more compression than you need. Even 9% can do it if the vox fit the mic.
After compression, you have the question of Stereo Processing. The answer is no.
Your lead vocals are now in mono and ready to receive some polish. Add your first EQ before the multiband compressor. Feel it out. When using compressors, our EQ objective is to cut around the second octave to remove any street and electrical noise and use as few bell curves to nudge any heady bass tones or flat highs in our favor.
You may now look into low mixed reverb or my personal preference; delay. Reverb can get cakey like bad concealer or unmixed paint. If you have lots of staccato notes, pick delay. It is much easier to control and provides the same richness as reverb. Read my delay topic to learn why.
Once you have chosen your prefered room sound via delay or reverb, the last step is gainstaging. I use saturators becuase they provide more volume with less clipping. If you're uncomfortable with putting a distortion on your vocals, you may alternatively compress the vox behind your time fx and throw a soft clipper at the end. Always close this step off with a limiter any time agressive boosting is required. Saturation will crack up transients. Compressors will eat dynamics like pop rocks in a blender. Limiters will simply listen, adjust, and then leave that shit alone.
I sidechain my vocals into my mixdowns. Whether it is a guitar, a pluck, vocals, percs, kick, snare... we use our vocals as an input signal to make room.
In addition to creating room for the vocals, you want your stereo mix to come from comping and harmony. Take Billie Eilish for instance. Her harmonies will be hard panned like guitars while her lead vocal track will be hard center. The lead vocals...[to be continued]
- Guitar
Please refer to:
- > Distortion for achieving all guitar tones.
- > Phase for amp recording etiquette.
Index
When recording, there are a few standard practices that are not up to interpretation. Leave the tone knob rolled up all the way up. Your producer is not your guitar tech. Mixing cannot create frequencies that a tone knob un-alives. Double tracking should be 2-4 channels of dedicated re-recording each part, per left, per right channel.
We should also talk about the impact of a good set up before continuing.
The "Neck" of a guitar cannot be mixed out of a recording. Buzzy frets? the neck is backbowed. This forces strings to canoe upward and touch strings with open frets. Are notes sharp or flat for no reason? Check that the neck isnt flexing up over the body. This will disproportianately increase the travel distance of a string, altering the pitch accuracy per note, per fret.
Fix either of these issues at home by unscrewing the cover on the top of most "Headstock"s. For backbow, loosen the truss rod with a hex key via counter-clockwise motion. For arching, rotate the truss clockwise. A truss is meant to provide stability to the wodden neck. Do not attempt to loosen or tighten based on rigidity. When in doubt, make the motion of the clock in sequential order in an orientation you recognize and before adjusting the truss. You only get one neck.
Setting the neck requires the fretboard to be flat. This may be tested by rocking a perfectly flat object along 3 or more frets. Something as simple are a speed square or other machined level tools may work in a pinch. Raised frets contribute to buzz, but do not affect intonation. These must be hammered back in to remove this type of fret buzz.
Intonation is a little more complicated. Setting the pitch of the 12th fret to the whole tone is my prefered method. No oscilliscope required. Take the pitch of the ghost harmonic at the 12th fret and reference it to the pressed 12th fret. If the 12th fret plays sharper than the open note and the ghost note, more distance must be created between the saddle and the nut. Do this by tightening the saddle post. If the 12th fret is flat when pressed, shorten the length of the string by loosening the saddle post.
Adjusting saddle distance is not the same as adjesting saddle height. Saddle height should not be changed unless a heavier string gauge is causing unwarranted fret buzz and no other issues. Same for if a string bends multiple semitones when pressing down lightly. decreasing saddle height will improve pitch accuracy.
Lastly, grab a tuner and play note by note to confirm that there are no dead notes or sharp spots. Bowing in the neck creates sharp notes. The strings must be loosened and the neck must be tightened again. Check one more time with that fret rocker for gaps or wobbling. If a string is floppy, the saddle was raised
Pickup height is the last factor to consider when adjusting a guitar. A player will know if their pickups are too high. What they might not know is how high pickups have the same magnets as low pickups. The magnets may pull strings out of tune, causing them to go sharp. This is an issue that should only happen to people who do their own setups without guidance. Factory set guitars will not do this.
The only time pickups should be adjusted is to increase dynamic range (lowering) or increasing the "bite" under distortion (raising). Do with that information and a size 0 Phillips screwdriver what you will.
An optimal room and speaker cabinet set up has a few approaches with specific requirements.
- Live Room:
Acoustic Treatment in the form of bass traps, broadband absorbers, surface diffusion, and hard reflections + recording medium. - Direct Input + Amplifier Out:
real time signal splitting and 2 channel recording (DI), a high quality dynamic microphone, an acoustically dead chamber - can be a stuffed closet or anechoic box, and an interface+DAW or mixer with 2 recording mediums. - Dynamic Microphone:
Acoustic treatment for a [1]Live Room can be as inventive or as simple as one wants. I personally require my traps to handle frequencies below 50 hertz. That means the room needs to be massive and the absorber density needs to be super low. One distincition that needs to be made before we continue - a critical listening room needs as little room coloration as possible. Stripping the room of its characteristics is the opposite of what we want to achieve for a recording space. Any acoustic treatment will be in service of the instruments performed. In a monitoring space, the opposite will be accomplished through fine tuned reflection canceling and frequency build-up elimination.
A solution I favor is the [2]Still Air Box. I would frame a 3x3' cube with medium density fiberboard (MDF) and pack it with rigid insulation. This creates an anechoic chamber, designed to isolate the sound of the speaker going into the mic. One could also spend the same amount of money adding bass trap columns, tuned & broadband absorbers, and cutting their own diffusion panels. Neither option is "correct" and both have a similar amount of context sensitive strengths and weaknesses.
Monitoring through a real amp and speaker while recording a clean DI(direct input) signal is a must. The trick is to mirror one's guitar input to a dedicated amp-out box. Since the dry signal is recording at the same time as the amp, mic position and distortion can be edited as mixing problems arise. IK Multimedia offers a one box solution and I advise you do not pay for this POS. It requires a dedicated 1-account login and that you ONLY use it on mac or windows. Stick with the annoying solutions and they'll pay for themselves in time, peace, and flexibility.
A dynamic mic - you know which one, is prefered over condenser and ribbon mics because guitar recording is about driving the speaker and amplifier until they get a little toasty. Dynamic mics can handle that ripping hot voice coil, often hitting 100-110db. I may include some of my favorite non-standard mics when they're ready to buy their timeshare. Sadly this is the type of thing you cannot achieve in an appartment, or even most neighborhoods. The speakers will reach +80db at minimum. This is then fed into the recording interface at -18db to -10db signal level. Why? Analog recordings can be turned up, but they cannot be turned down. Ever. This limitation is why amateur mix engineers try so hard to mix EDM at -20db and then crank a limiter in post without questioning why everything has to be so quiet.
| EDM created in a box doesn't use that headroom. Your input signal does.
Same principles here. Roll up the tone, roll down the volume knob(s). Double tracking each left and right channel is a non negotiable for good sounding guitar. The trick here is a mix of convolution and harsh EQ.
The tone knob does not get a say in the studio mix because it is a destructive edit. A lowpass pass placed at the beginning of the signal will do the same thing our live preset will do. The volume knob rolled all 10s can also be simulated with shelf EQ and a soft clipper. Blasting it to shit before the DAW can save it is the #1 way to sound like Death Grips.
- Bass
Bass guitar sound has become standardized by a direct input and a Fender Power Bass, or P-Bass. Direct in is optimal =as it removes the potential comb filtering from a recording source with bad acoustic treatment. Any parts that have sloppy tail noise should be chopped out of the wave. Frequency splitting is the best way to achieve clarity and texture and this is done by professionals by duplicating the entire recording, compressing => running a lowpass on the sub layer and using a highpass on the top, saving all distortion, chorus, digital wash, etc for this top channel. The sub layer will stay raw or minimally distorted to retain low end.
TL;DR - Bass is always recorded in mono. Standing bass does not require much beside a room with acoustic treatment and hard surfaces with optional wall and cieling diffusion. Bass guitar is more complicated and will be the only bass mentioned from here forward.
Clone a recording of the bass track without any modifications. One is your dedicated mono sub. The other is your bass top. This one may be micd up to a live speaker cabinet, processed digitally in your guitar group, or treated strictly as a clean digital track for ease of EQing.
Filter the high end out of the Sub Layer at octave 3 at a slope of -6db. Optionally raise your cutoff frequency 7 semitones from the root to allow wiggle room on melody forward songs. This layer provides both sub and bass presence. The only assistance the bass top will provide is in the mid range. highpass filters will be applied on the master. Any cutting of the low end at this stage will negatively affect the bass response on premium listening setups such as cars or home entertainment centers. Only use a shelf to reduce bass distortion at the master level.
For the bass top, filter the low end out beginning at the 4th octave at a slope of -12db. I reccomend that lowpass filters are only added to the pre-EQ to reduce mechanical noise. This top can now be treated like a guitar or like a synth. Place white noise generation behind the EQ to create a humming fuzz. Use distortion to add mid-range. Do stereo processing for textural enhancements and improved dynamic presence in stereo settings. Add flangers to make the bass melt. The sky is the limit.
- Wind & Brass
For the time being, please refer to vocals. Many of the recording techniques are shared among soloists.
- Drums
