Taming the Dragon
...and teaching it to sing.
In this final article of a three-part series we discuss the processes of editing, mixing and mastering your music at home so it is ready for release into the wild.
So you've finished recording, congratulations. The tracks are all "in the can" as they say. Miller time? Not so fast Homer! Your piece still needs some polish before it's ready to be shared with your adoring public. Audio post production is a crucial piece to the puzzle in order to make sure your masterpiece sounds to your fans the way it does in your head. We'll start with making sure we have all the pieces in the right place.
If you're recording digitally, you have the luxury of non-linear editing at your disposal. This means you are able to surgically cut, copy, and paste tracks to and from anywhere you like. This can save you a lot of time and trouble and can even be used creatively but it's also easy to get carried away with this capability and end up doing more harm than good for yourself in the long run. I personally have seen more than one band spend so much time on editing their performances that the recording ended up sounding like a completely different group. Granted, the final result gave the impression of an almost perfect performance, but the reality was that there was no way these guys were going to reproduce anything that sounded like that live and what it boiled down to was misrepresentation. Yes, these tools are very powerful, but there's a certain point where you have to ask yourself if you're really being honest with your fans. That said, here are a few nice uses for digital editing.
Sometimes you can take an existing song and create entirely new arrangements or versions that are extended, shortened, etc. This can be done on the fly and reviewed instantly. You can see how the tune sounds with the solo being twice as long or adding a verse to the end. The possibilities are endless.
If you need to re-record a short passage of material such as a solo you can set automatic punch points to make sure only the section you want gets recorded. This takes the guesswork out of punching in.
If you have a few takes of a song and each member's best performance is on a different take you can combine performances to get the best of all worlds. This is called track compositing. Now is when good isolation is priceless because you don't want any bleed from other takes ruining the edit. Also to pull this trick off the drummer is going to have to play to a click track. There is no way around it. Not all drummers can do this but it's crucial if you're going to do any cutting and pasting since tempos fluctuate no matter how tight the band is.
If you're going to be cutting and pasting parts together you're going to have to make your cuts at times when there is silence on the track. Otherwise there will be an audible pop. You will have to find what is called a "zero crossing" Sometimes you can get away with cutting a snare hit or kick in half but that will be hit and miss so good luck. If you have no silence to work with an alternative method is a "cross fade". This means that instead of a clean cut the computer will quickly (we're talking milliseconds) fade out the track volume right before the intended cut point and at the same time fade in the inserted piece. This makes for the smoothest transition possible and it should be very difficult to tell the cut is even there.
By nature, MIDI editing is much more versatile than audio editing. Since you're not really editing any audio and you're only dealing with coded instructions, a lot more is possible with MIDI. You can fix timing problems with quantizing. This corrects every note in a sequence that you specify so that it is exactly on the beat. Some extensions of this feature are the humanize function which randomly shifts notes around so the performance ends up being slightly imperfect. This doesn't ruin the piece but actually helps make it sound less mechanical. A combination of these two features is something called a groove template. This is a preset formula that quantizes a performance but sets key elements like a snare drum to fall slightly "behind" the beat to give the pattern a solid groove.
If you've ever programmed a drum machine you know that if a song primarily consists of one four measure pattern all you have to do is program that one pattern and repeat it as many times as you need. With digital multi-tracks this is called looping and it's a big time saver although you have to be careful that the song doesn't end up sounding too mechanical and lifeless, unless of course that's what you want.
The ability to transpose a song into a different key comes in handy when a singer is having problems with the key in which the song was originally written. Tempo changes are also possible during part of a song or the whole piece.
There are as many ways to approach the task of mixing as there are artists producing music.I will introduce you to a method that will be a good starting point for everyone who is new to music production and that will also leave a lot of room to find what works best for you based on the style, schedule and budget of your project.
Effects / Signal Processing
As you get good, clear, strong, isolated sounds recorded you'll begin applying effects that will start to shape your vision of each instrument on your way to the final mix. This will help you build a solid foundation for the final mix as you record your tracks and it's a process that takes practice. A good set of ears in the studio is not something that comes easily but it is something that comes with patience and practice.
If you're using software effects like those found on Garage Band try and use them as non-destructively as possible. By this I mean that you should patch them through the auxiliary sends of your mixer and keep the dry signal intact. Your ability to do this will depend on how much you end up taxing your CPU by having many tracks and effects going at once during playback. If it turns out that your system can't handle playing back all the effects you need in real time you may have to make virtual copies of each track, mute the originals and then use destructive effects on the track copies to get the sounds you want. This is more time consuming but much easier on your computer because it doesn't have to process all the effect signals "on the fly".
There are many types of effects out there. You may have them in the form of rack mount analog or digital units, stomp boxes, or software. You can get good results from all of these so don't be afraid to mix and match to come up with a sound of your own. Most effects can be put into a few major categories. We'll briefly cover the ones that are most frequently used and then discuss them in a little more detail.
Equalization (E.Q.) is the process of boosting or cutting certain frequencies or groups of frequencies in the spectrum of sound. Most if not all musicians have at least a basic grasp of what E.Q. is and does. It's how you shape your sound. E.Q. controls can be as simple as a bass or treble knob on a stereo or as complex as a fully functional parametric E.Q. with which you can specify exactly which frequency you're adjusting.
Delay and it's close cousin reverb are in the business of simulating acoustics, meaning they trick your ears into thinking that sound is located in a place of a certain size, shape, and material.
Modulation effects such as chorus, phasers, and flangers mix sound with a delayed copy of itself in different ways to achieve a wider, sometimes sweeping cyclic wash of sound that is controlled by tweaking different parameters.
The first thing to do with each track is determine if you even need to use EQ. If so, it's much easier (and better) to effectively cut (attenuate) frequencies than it is to boost them. That is why having a nice, full realistic sound during the recording stage is the way to go. This is achieved by recording with as little E.Q. alteration as possible (a flat response through the mixer). Think of the sound you have recorded as a block of stone or wood. With E.Q. we "sculpt" a sound out by cutting away everything that does not need to be there. Boosting frequencies isn't forbidden but you don't want to create a situation where things start to compete for space. A lot of times I'll use E.Q. as a combination of boosting some frequencies and proportionately cutting others to obtain the desired effect. This helps keep a consistent signal/noise ratio and helps prevent clipping (distortion). During every stage of the mixing process revisit each track and see if you need to adjust the E.Q. in order to make better use of the frequency spectrum. A good final mix does not sound like it's thin or lacking in any area BUT there is only so much room in the spectrum that humans can hear and each track should only have what absolutely needs to be there to satisfy the mix.
Good engineers are able to make you think you hear something that really isn't even there (or barely there) by only including the essentials. Listen to a song you love. Chances are the drums don't sound anything like they do when you stand next to them. The attack of the kick drum may only come across as a slight thump or you may mostly just hear the "click" of the beater hitting the drum head. But it's EQ'd correctly and there is enough other things (like bass guitar) going on to fill up the space and do the job of convincing you that you can even feel the kick, just like you would at a live show. "Efficient use of sonic space" is a mantra you should remember.
Common stages of an E.Q. unit include...
Input: This is where you can adjust the level of the signal coming in. If you did a good job of recording it usually shouldn't have to be messed with and can be set at unity gain (nothing added or subtracted).
Type: High Pass (lets high frequencies pass and blocks the lows), Low Pass (vice-versa), High Shelf (controls a specific frequency and all those above it), Low Shelf (vice versa), Parametric (lets you zero in on very specific frequencies)
Gain: Choose whether you are boosting (+) or cutting (-) frequencies.
Q: Widens or narrows the range of frequencies you are adjusting.
After E.Q., this is the second most important skill to master since reverb in all it's different forms is so prevalent in modern music mixes. It's an extremely powerful tool when used correctly. This is where we start to get into the idea that effects are like make-up on a woman. Too much and she looks like a clown, so unless you're going for that effect be very careful. A little goes a long way. As a matter of fact, when all the tracks are up and the song is rolling you really shouldn't be able to pick out much reverb going on at all. It's more like a gentle transparent subliminal glue (yes you can quote me) that helps everything gel together and coexist nicely. Usually the determining factor on where to use reverb in a basic mix can be broken down to three things.
If a sound can be relatively short and cuts off quickly.
If a sound is in the higher register.
If a sound is a focal point, meaning it's something people pay attention to specifically.
If a sound satisfies at least two of these requirements it should be considered for reverb treatment. Snare hits, vocal lines, and guitar solos all fit relatively well into all these categories. Kick drums and bass don't because even though they can be short sounds, they're not high or a focal point. Of course this is all subjective but it's just meant as a starting point. Common sense is king here. Also, if you live in magic digital land, you should create a virtual "wet" track for every instrument with effects on it, side by side with the dry ones. This can be done for each track with all other tracks muted so your CPU doesn't have a coronary. Then either the wet (printed) track or the dry track (with aux effects) can be played back depending on if you're tweaking or just listening. This also saves compu-sweat. Printing your effects like this helps give each track it's own seperate space in the effect process and it doesn't sound as mushed up as feeding everything with reverb into the auxiliary using whatever comes out for the final mix. It's cleaner. With analog you will have to do your best with this in mind and be very mindful and judicious about your effects use. Most reverbs (hardware and software) come with presets that can be used as starting points for experimentation.
Common stages of a reverb unit may include...
Input: How much signal comes in. Again we want as strong a signal as possible without distortion.
Mix: Lets you determine how much of the effect you hear at the output in relation to the dry signal. If you're patching signals into the unit via the effects send you will almost always want to keep this setting at 100% wet and use the aux send and effect return to control how much reverb is heard in the mix.
Algorithm: is a formulated model of different kinds of environments that provide natural reverb.
Size: sets how big the room or environment you've chosen sounds.
Diffusion: determines how reflective the surface of the space you are simulating is.
Decay: How slowly/quickly the sound fades away. Usually the faster the tempo of the song is, the more quickly you want the reverb to fade so it doesn't muddy-up the mix.
This effect uses a slightly delayed and de-tuned copy of the original signal to produce a wide, shimmering sound. It's commonly used by guitar players, keyboard players and singers. When it's used on vocals it gives the illusion of a wider, fuller sound and can help mask subtle pitch problems.
Common stages of a chorus include...
Delay: controls how long it takes for the copied signal to appear. This is measured in milliseconds.
Depth: controls the amount of change in modulation or pitch of the sound. If you want a more extreme effect, set this higher.
Rate: controls how fast the modulation will cycle.
This effect is essentially an adjustable echo that can be placed on any signal you choose to achieve a desired effect.
Common stages of a delay include:
Delay: the amount of time the effect waits before playing the duplicated signal.
Feedback (Regeneration): Adds an adjustable portion of the delayed signal back into the input.
Modulation: alters the playback speed of the delayed signal to achieve various alternative sounds.
Mix: Again, lets you determine how much of the effect you hear at the output in relation to the dry signal. If you're patching signals into the unit via the effects send you will almost always want to keep this setting at 100% wet and use the aux send and effect return to control how much reverb is heard in the mix.
There are two main methods for getting effects of any kind onto your signal. Effects can be "inserted" into any individual channel in the mixer. The inserted effect will only process the signal on that specific channel. As mentioned before, physical mixing boards often have a quarter inch insert jack. Special cables are required to use these insert points that have a TRS (tip ring sleeve) stereo jack at one end (the mixer end) and they split into two mono jacks at the other end that can each be plugged into the input and output of the effect. Virtual mixers are able to automatically patch effects in without cables but they still only effect the one channel.
The other method requires use of the AUX (auxiliary) loops on the mixer. Any effects plugged into this system can be applied to any channel on the mixing board through the use of the AUX 1, AUX 2, etc. knobs on each channel. This method gives you a little more bang for your buck not only because it saves on gear but also on CPU sweat if you're working in virtual reality. It doesn't tax the system nearly as much. You're also free to add as many auxiliary channels you want to a virtual mixer as long as your system can handle it.
The "Big Picture" of Mixing
Hopefully even before you started the recording process you had some sort of idea in your head of what you want your final mix to sound like. A good way to do this is to use a professionally mixed and mastered song as a guide. This gives you a real example of what you want things to sound like that you can compare and contrast to your own mix along the way. However, don't put yourself through the agony of trying to sound exactly like another recording. It's not gonna happen and you'll just be chasing your tail. Use it as a guideline on the way to finding your own sound.
It goes without saying that the job of mixing takes into consideration what the song sounds like as a whole. To break this down into manageable bites you can start to think of the different "dimensions" of a mix individually. Where does each signal reside, in the foreground or background? This is primarily controlled by the overall volume of each track and is the basic starting point of the mix. Even though everything must ultimately sound good together it's too much of a task to work with the entire mix at once. If you've been building a rough mix as you've been recording please keep these settings intact, solo the drum tracks and start checking the mix systematically, adding tracks as you go. Your best friends on the mixing board as you do this will be the mute and solo buttons. These guys let you hear tracks individually, in groups or as a whole quickly and easily as you move through the mixing process. Also, most importantly, no matter how anxious you are to finish a mix do not attempt to start and finish an entire mix all in one sitting. Spread it out over a few sessions. Trust me on this one. Your hearing, judgement and nerves all start to get warped after more than a few hours of obsessing over the same parts of the same song. Take breaks and stay as fresh as possible when you're making these decisions or you risk having to go back to fix your work later. Take your time.
Start with the drums. Depending on how big of a kit your drummer has and how many tracks you devoted to drums this can turn into a mini-mix job of it's own. I recommend setting a level for the kick drum and then basing everything else on that. There's no set method for this but it does give you a starting point to work from. Once you have a mix you can live with for the drum kit by itself you can send all the drum tracks to a stereo bus that can serve as a submix fader so you can adjust the drum volume without going crazy trying to keep the levels the same with respect to each other. Some software programs let you lock (group) faders together and move them as one unit. This achieves the same result.
After the drums add bass, rhythm instruments (keys, guitars), horns, backing vocals and then lastly lead guitar and lead vocals. As you mix, be sure and listen to the song at different volumes and on different speakers if possible so you can get an idea of how the mix will sound on different systems. Different speakers for example will bring out different frequencies in the mix. It's doubtful that you will dial in a mix that sounds great on every system, but you should try to achieve a good balance of quality between them. Every time you add a track you're going to be faced with new problems to solve. At each step pay attention to this checklist of dimensions and be fairly satisfied with the mix before adding the next track:
Foreground / background: primarily controlled by each tracks individual volume as stated above.
Depth: Effects (especially reverb, chorus and delay) play an important role in how deep or far away a signal sounds from the listener. Remember that less is usually more when dealing with effects. If things start to get too crowded, muddled up, freaky, or just plain F.U.B.A.R., consider pulling back on some effects first to try and solve the problem. We should be able to hear each instrument clearly. That's why they're there right? Yes, sometimes things that aren't completely clear or that are "subliminal" can have some positive effect but those things should only be considered when all the main components of a mix are up and running well together.
High / Low: This has to do with frequency (E.Q.). There is only so much room in the frequency spectrum of human hearing. Just because you may have unlimited tracks to work with doesn't mean there's unlimited room in your mix. Each sound should occupy it's own space. Sounds should not compete with one another. Remember the sculpture analogy? We concentrate on trimming the fat and cutting frequencies instead of boosting whenever possible. Boosting frequencies and levels too much has a tendency to get you into trouble during mixing because it starts to make sounds compete for space.
Left / Right: Your pan controls determine where a sound lives in the stereo picture and is the same as a balance control on a stereo, only now you have control over each track by itself. Generally it's a good idea to give everything it's own space or "slice of pie" in this dimension as well. You can think of how the band is set up on-stage. Kick center, hi-hat a little to the right, bass a little to the left, etc. Sounds have an easier time co-existing in the same space of one dimension the farther they are apart in another dimension. Lead vocals and kick drum are more forgiving of each other when they are both dead center in the stereo spectrum because they are a good distance apart in frequency spectrum, etc.
The Final Mix
Let me state a hard but simple fact of life. Chances are slim to none that you will arrive at a mix which completely pleases every member of the band. That's just the way it is, especially at this level of D.I.Y. production. There are a few common complaints, most of which usually stem from the fact that deep down, everyone wants to hear more of themselves, but compromises have to be made plain and simple. You have to concentrate on what serves the song best and what brings out your band's sound the best. How you come to an agreement about that is up to you, but just know that it's something that will have to be addressed at some point I guarantee you.
Once that bridge is crossed the final mix can occur. The term "final mix" is somewhat deceiving since there are a couple of more steps we will discuss in this process. Whether you have been working with analog or digital up to this point I highly recommend mixing down to a digital medium since it is much easier to work with when it comes to mastering and duplicating your finished product. The final resting place for the stereo mix can either be a DAT tape, CD or hard drive. Please consult your user guides for specific instructions on bouncing down to stereo with either software or workstation units. If you're working with a hardware (real) mixer take your main stereo output and plumb it up to your stereo recorder of choice, press record on the two track, play on the multi-track and make some history.
Automation / "Playing the Board"
While the song is playing you may (and probably will) need to make adjustments on the fly because a track setting may sound great at the beginning but terrible during the guitar solo, etc. These adjustments usually have to do with volume but can be most anything depending upon the characteristics of the music. Here's where automation is worth it's weight in virtual gold. Lucky stiffs with this capability get to record fader movements, pan movements, E.Q. and effect adjustments separately and have them play back exactly the same way every time, making the necessary adjustments perfectly every time. If you're working with a software program you more than likely have this trick in your bag. Some digital workstations have it too. Four-tracks don't and most hardware mixers usually don't unless they are higher end. If you don't have these magic powers you will have to go the old fashioned route of getting as many hands as necessary on the board during mix-down to stereo and making the proper tweaks on the fly and if someone screws up you try it again until you get it right. And no matter which way you slice it, that's just good clean fun! Hilarity/frustration will ensue at first but with practice you'll be able to do it in your sleep. Digital users also usually have a function on their gear that allows them to evenly fade out a song at the end either through fader automation or a linear fade process that can be applied at the end of the song. Analog kids will still have to man the master fader(s) at the end to get it just right.
Most musicians know less about the mastering process than they do about mixing. Mastering is the process of maximizing a songs overall volume, making final overall E.Q. adjustments and taking all the songs you have recorded and making them sound good together as a unit or "album" of songs that belong together. This can be as simple as making minor volume adjustments or can get more involved in levels of E.Q. and dynamic processing (compressor / limiter) treatment. Any final stereo mix can be imported back into a software multi-track program for the purpose of mastering. There are also software packages on the market that have been developed especially for the purpose of mastering music. Whether you decide to use these or not, here is a simplified procedure for mastering that anyone can follow. Since mastering is a process of "fine tuning" different mixes in relation to each other it must be stressed that all adjustments are to be made slowly and sparingly while paying great attention to the details of the results. Again, TAKE YOUR TIME.
While using dynamic processors (compressors / limiters) at this stage, become best friends with the bypass switch on the processor. This will give you a constant and clear picture of exactly what you are doing to the sound at all times. Remember, you didn't come this far to screw it all up now. Use compression first to slightly reduce the dynamic range of the audio. Remember we said that "dynamic range" is the difference between a songs loudest and softest parts. In recorded music a tighter dynamic range is perceived by our ears as louder overall. This means there is less of a difference in volume between the loudest and softest sounds and we can set a higher listening volume to be able to hear it all clearly. We want to be able to hear quieter sections without blowing our speakers when the chorus kicks in. Compressors help us do this by limiting loud sounds "compressing" the overall dynamic range. We try to even out the sound and be able to hear the entire collection comfortably without riding the volume knob on the stereo. Again, computers afford you the convenience of saving copies of your mixes to be treated in different ways without destroying the initial quality. You should always be able to start again at square one when things get out of control.
Next, apply some limiting in order to boost the final signal as loud as possible while still avoiding the red clipping warnings on the meter. This helps to reduce noise on the track and gives you a cleaner sound as well. You'll find that after applying these last two processes the songs will start to gel with themselves and with each other.
The final stage is tone and sequence as you attempt to put the songs in the most cohesive order and making slight overall E.Q. adjustments until they flow as evenly as possible.
The Acid Test
Once you have your first final mix you should continue the task of playing it on as many different sound systems as possible to see if it's sounds right. Try it in the car, on your home system, your computer, your buddy's system, etc. This will probably reveal some inconsistencies which are always fun to deal with especially since you've spent so much time and energy getting to this point. Return to the mixing board, rinse and repeat. All I can say is just like with anything in life, practice makes perfect and you will have to put the time in to really know what you are doing and avoid having to remix. It happens to everyone. Don't let it discourage you. Chances are if this is your first attempt at mixing you will go through a few final mixes before you arrive at something you can live with.
Wrap It Up
When all is said and done and you have burned the final mix to a CD that is ready for duplication or uploading, if you have spent enough time and attention to detail on your project, you will have a product you can be proud of. The whole process is probably a lot like giving birth. If out mothers really could remember the pain and anguish they went through bringing us into the world they'd probably never do it again. But, unlike passing a sack of potatoes out of your behind this creative process does get easier and better with time.
And it's a lot less messy.