In previous articles, we've addressed basic hardware you need for recording, including recording devices and mics. We've also discussed mic placement and the recording process itself. This article is devoted to mixing and mastering and finishing the process up.
Put simply, mixing is analogous to editing in film, it's when you take everything you've recorded, pick the best takes, and put them together to get a nearly finished result. This is actually the part of recording that in my experience takes the longest, is most frustrating, and requires the most real work. Where do you start?
First go through the takes you have from each instrument and pick out the ones that gel the best together. Did the drummer not have a single take where they nailed the one beat right after that break? Then pick only instrument takes that line up with this mistake. The goal is not to have the least mistakes in a recording, it is to have a recording that overall sounds very consistent. Some truly great albums have had huge mistakes and improvisations on them.
In the process of whittling down to the best takes, you may find that you need part of one take and part of another to have an overall more consistent take. This is easily accomplished in most pieces of sound editing software. Usually, it just entails splitting a take and dragging it over the part of another take you wish to replace. This can often result in a clipping sound when the resulting sound wave doesn't match perfectly. The thing to do here is to zoom in on the take and move it over till you can get the wave looking reasonably consistent. It's helpful to cut directly before the attack of a note to minimize the chance a distracting clipping will occur. This process is hardest with bass because the wavelengths are the longest and I've frequently found that moving a bass part such that it won't clip will make it sound slightly out of time.
Some sound editing software packages will have a built-in plugin for repairing a clip such as this. Trial and error is the best way to work through these things. And remember, if a vocal part is going over a mistake, it takes a very discerning ear to pick up on it. I have literally spent entire days trying to clean up tracks in this manner to a great degree of frustration, only to have the result come out sounding alright to everyone but me. A lot will depend on how much of a perfectionist you are.
When mixing down, you want to wind up with a single take of how ever many drum tracks you laid down simultaneously (for me, it's usually kick, snare, and a stereo overhead track), one take of bass (if you want to overdub bass, it can sound cool in many circumstances but the frequencies it works in are some of the least articulated in the human hearing spectrum), and usually two takes of guitar. The reason for this is that a mix will sound much fuller if you have two takes of guitar, one panned left and one panned left. Frequently a guitarist will come up with more than one part for a given section of a song and it's very easy to pan one left and one right. Listen to Pearl Jam's Ten for many great examples of this. It fills up a mix and adds far greater depth to it. If you can't come up with two good takes of a guitar part to pan, there are a couple ways to get around this and still make use of the stereo spread. The first is to copy your one take and pan a copy that has been phase reversed to the opposite side of the spread. This can be done quite easily in Cubase and most other programs. It will yield a somewhat bizarre sound that doesn't actually sound like it's coming from anywhere at all. Bear in mind that you are limited on volume by it because, by nature, out of phase sounds cancel each other out and thus you'll find that by lowering the volume on one, the overall volume of the guitar track will increase. This option is often used for live albums when a band has only one guitarist.
The second option is a little more complicated: copy your one good guitar take and place it in its own track. Now move it so it is slightly later than the original. This doesn't have to be a huge difference, usually half a second or less will do the trick. This is technically creating a delay, which is the principle chorus pedals are based upon. To add to the effect, you can pitch correct the new sound sample by plus or minus 5% to make it seem like it was a later take. Overall, this will do a lot to make the take sound different in the mix and you now have two takes to pan left and right. If you used the Billy Gibbons trick for recording the guitars with one mic out of phase with another, just pan one mic right and one left. Pure and easy.
When it comes to vocals, sometimes you can benefit from a technique called doubling, which is similar to what we just did with the guitars. Doubling with vocals occurs when you have multiple takes running simultaneously. It results in the vocals having a smoother quality and overall sounding like the notes in them are more consistent. You hardly hear any pop recordings these days that do not utilize a good deal of doubling with vocals. Usually there is one main take and one or two others that sit quieter in the mix to bolster the sound. This can make for a great dynamic shift in a song, for instance you can use it to set apart the vocals in a chorus from those in a verse by adding overdubs in the chorus. This effect can also be simulated with a short delay (200 ms or so).
If you've recording additional instruments, take note of where they fall in the frequency spectrum compared to the other instruments. I've found that if you have guitar and piano or keys in the same song, they'll be taking up a lot of the same space. The solution with two instrumental parts that are in the same part of the frequency spectrum is either to pan them away from each other or to take out one of the instruments. If there are too many instruments taking up the same part of the spectrum, you will see a pronounced hump when it comes time to master. You can fix this to a limited extent with EQ or a multiband compressor, but the less you have to use these tools the better. The moral is to try and find instrument parts that will stand out from each other in the mix. And whenever you have two parts that seem to just sound like mud or you need more definition out of them, try panning them away from each other in the mix. This will also make the mix sound much more interesting in headphones. Ever listened to Bohemian Rhapsody and tried to catch where it sounds like everything's coming from? It's a treat. For most mixes, you'll wind up with drums, bass, and vocals panned center, with guitars panned to the sides and sometimes vocal overdubs/harmonies panned slightly to the side, too.
Once you have your final takes, it's time to start processing them with effects. You can add the effects during the processes above, but I've found that if you can make a mix work without any effects, it will sound a hundred times better when you add them in. I've had experiences where I get tempted to EQ before I've got panning and relative volume levels set and a lot of times I've completely had to throw out all my effects and start over again with panning.
When it comes to EQ and other effects, I could give you a greatest hits list: vocals absolutely need compression and reverb and delays sound beautiful on a snare, but the vast majority of what I know about these things comes from these two pages, so I'll just give you their links instead:
In my experience the latter page has more useful information condensed into less space, but every once in a while I hit a snag in a mix that the former page has the solution to, in addition the compressor settings on it are invaluable.
When setting a compressor, always make up gain to make sure the peak of the track is at the same place it was before so that your relative volume levels don't have to change very much. I've never found a good guide for reverb settings, so you'll really have to use your ears. I will tell you, though, that the longer a reverb tail, the lower it should usually sit in a mix. Backing vocal tracks can usually afford to have more reverb on them as it will make them sound more distant in a mix and so they won't battle with the lead vocals as much. Also bear in mind that when you listen to a solo vocal track running with reverb on it, it sounds worlds different than it will with all the other instruments running as well.
This is another spot where things can become frustrating and you can experiment. Some of the most off-the-wall mixes have become hit records. Listen to Purple Haze for a great exampleputting the vocals only through the right side? Similarly, the Rolling Stones' Ruby Tuesday puts the drums completely through the left channel. Blind Melon's Change completely switches the channel the guitar is coming through as the vocals come inin mono it becomes very disorienting! And yet all these records work, so don't be afraid to play around a little bit if you've got an interesting idea.
When it comes down to it, mixing is getting all these separate tracks whittled down to a single stereo spread track for mastering.
Mastering is the final process by which a recording is prepared for reproduction. The techniques behind this are often very so secretive that mastering artists can seem like occultists. Basically in mastering you are trying to sweeten the mix a little. Part of the process of mastering involves final stages of compression, to ensure the volume is as constant in a mix as possible. Most tracks you hear on the radio are the same volume for almost the entire duration of the song. Try opening some of your favorite MP3's up in your audio editing program to see how consistent they are. Of course, jazz is a different story, as are genres like bluegrass, classical, and to some extent country. All of these genres will show a lot of dynamic changes in the course of the song. Anything in the rock format or pop will be consistently as loud as it can get.
Beyond compression, however, mastering is also used to fix certain parts of the mix. For this part of the process, a person needs a spectral analyzer and either an EQ or multiband compressor. The latter will be better for this part since it will dynamically respond to a mix and thus will sound more natural. Either way, check your song to see how it looks on the frequency spectrum. If you're like 90% of first time recorders, there's a huge notch between 800Hz and 1500Hz. This is because guitars sound powerful with their mids dropped out and everyone loves that powerful sound. The problem is that between those frequencies is the meat of guitar sound. No other instruments fill that part of the spectrum as well. Open up one of your favorite songs and put it through the spectral analyzer and you'll see that the way a song should look is an almost completely flat line between about 150Hz and 15kHz, with volume tapering off at either end of this. This is what you should strive for in your mixes.
You can fix the mid frequency notch by either going back to your original mix and EQing accordingly, or by compressing the frequencies around these mids. You can use a multiband compressor to make those mids louder, but overall it makes a mix sound a hair unnatural. Why bother? Because most people have their stereos EQed to drop out mids, so they're already doing this work for you, so when your mix gets on their stereos it will sound very thin and harsh and the low end will be complete mud.
Another important point to emphasize: listen to your mix through every stereo and/or set of speakers you can find! You can never guarantee what system a person is going to be listening to your CD through, so make sure it sounds as good as possible for all sets of speakers.
An additional trick for mastering is if the mix sounds too cold or harsh to transfer it to audio tape. This is really hard to do outside of a mastering studio, but there are plugins to replicate the warming and compressing effect of tape. They aren't perfect, of course, but they can be good for what you need them for.
Mastering is your last chance for fixing anything you don't like, so don't be afraid to try! It's easily the most arcane part of the process and the most difficult to measure. Every piece of software I discussed before can be used for mastering, but when it comes down to it, if you don't trust your skills in this area (and I'm making it sound a lot simpler than it actually is), there's no shame in taking your mix to a local mastering studio. Rates are usually pretty good for individual songs and the people there can make your mix sparkle like none other. I'm still trying to get this part of the recording process down.
In The End
What comes next is artwork and duplication and distribution. You can find CD duplication in any town. Check the classifieds of your favorite local music magazine to find a list or search on Google for your hometown and CD duplication. You're bound to find something. You can also distribute online. Each of the programs discussed for this series of articles allow you to mix down in MP3 format. I recommend mixing to 128 kbps and a sample rate of 44,100. Below this and strange things start happening to the sound. Above and you can keep the quality pretty consistent with your original mix.
Some good online distribution venues for music include Myspace.com, Garageband.com, and Rock.com. I'm sure there are a whole host of others out that that will be recommended in comments appended to this article.
For artwork for your CD, cruise your local art school. You're bound to find someone who'd like another portfolio piece and will do artwork for you on the cheap or for free and having good artwork will make your CD look that much more professional.
So, that's it for me. I hope you've all enjoyed these articles and learned a lot from them. There are probably hundreds of things I forgot to mention that you'll discover on your own or things I'm going to realize in the next few years were misleading or I could have done better. The process of recording doesn't end until you decide you no longer want to work on it. And by all means the learning process never ever ends. I was asked by one person in a comment in a previous article how I learned all this and the simple answer is I experimented and tried new things out to see what worked and what didn't. Keep working at it and you never know where you can end up. Thanks for reading.
Oh, and here's the shameless plug: visit my band online here ;)