Anatomy of a Song
Sat, 28 Jan 2012
One thing I've always wanted while learning about recording and mixing audio was a complete session file, with all the settings, from a professional track. I did a year-long audio engineering course and I never saw one. I did learn a lot of useful information, but I've always wondered exactly how much bass was rolled off on that vocal track, or how much compression was really on that bass guitar.
I think a lot of people have musician friends who can show them their mixes, but I'm also aware that a lot of people don't, because I've seen other people asking for the same thing, and I've seen (and heard) too many home recorders who obviously have little idea what to do. I've still never seen a fully complete session file available for download. I've seen a few sets of all the tracks, but they're not mixed.
With the release of In Our Dreams We're Flying (it's a free album, go get it here or on Bandcamp), I feel like I'm just about at the point where I know how to make a professional-sounding track with the limited tools of a home studio. And therefore I'd like to try to provide what I wanted in the past.
This is the completed track, the song Perspective from the album, but stripped of any 3rd-party plugins and replaced with ones that come with a default install of Pro Tools 8+, apart from the compressor and limiter used in mastering. We'll be looking at a Pro Tools 8 session, which you can download, but I'm aware that a lot of people don't have Pro Tools so this post will also contain extensive notes and screenshots from the whole session.
I've chosen the track because it's fairly simple and I feel it's the sort of music that the average person might be recording in a home studio. It's folky stuff, so if you're making electronic music, you probably won't get much out of this I'm afraid. This isn't my best mix - I did the bulk of this one a while ago and there are a few slightly weird things - but I think it sounds fairly decent. You can of course judge that for yourself.
The Complete Session
Main Edit window shown first with Waveform view on every track, then with volume. Last is the full edit window where you can check out panning etc. Click for big.
First half of the tracks in the Mix window. You can click on the plugins to see their settings.
FXB on the Drums track is FXpansion BFD Lite, a drum software instrument, running off the adjacent MIDI track. This track is inactive as it's been bounced down to the DrmsB track.
The Piano track is inactive as it's been run through my TL Audio 5021 compressor and recorded back into the PnCm track.
The rest of the tracks in the Mix window. Again, click on the plugins to see their settings.
VOXM, the main vocals track, has a send to VOXFX for the reverb. The two female vocal tracks are both routed into an auxillary track that applies EQ and reverb.
Pr-53 on the Blip track is a Native Instruments Pro-53 software synthesizer. This track is inactive and has been bounced down to Blip B since the Pro-53 is broken in Pro Tools 8.
Mastr is the stereo buss that everything runs through. SLC is a compressor and L3M is a limiter. These don't come with Pro Tools but I'll talk about what you can do with what you might have. MstrCl is just an alternate clean stereo buss track without the plugins on it (just so I can see the levels).
This section contains specific comments to help explain why I did whatever I'm doing on each track. It doesn't cover the real basics like how a compressor works, but you can find that information from plenty of other sources. I also don't claim to be any sort of expert!
A lot of what you do in mixing depends on the recording, so remember that one solution doesn't always work.
DO invest in some good microphones, and DO take some time to find the best microphone positions for recording. Only once you have a good recording can you really do good mixing. This isn't a tutorial on recording, but there are many books etc. available on the subject.
Compressor: Average gain reduction on transients: 3dB/Max: 6dB. I don't have a real drum kit, and BFD Lite puts all the drum sounds through one stereo track. I want to add some compression to the kick and snare (mainly the snare) to get a "bigger" sound. I'm using a fairly long attack time here, because I want the hit of the snare to still be pretty clear and loud. The release is also fairly fast because the snare sound drops off quickly. I think these settings are pretty much the same as the default Drum Comp preset, which is a bit lazy of me.
There's a reasonable amount of compression going on here. If this was a heavy rock track you might use a bunch more, and possibly add a reverb as well. I don't want the drums to be really heavy here, but I also have to consider that all the drum sounds are on this one track. If I compress too much then the cymbals will start obviously being "sucked down" when the kick and snare hit. Separate drum tracks for the overhead mics etc. would alleviate that.
EQ: The EQ will depend entirely on your drum sounds. I've added a bit of bass because the original sounds were pretty bass-light. Normally you'd add the EQ before the compressor so it's not compressing stuff that's going away anyway (since most good EQ is cutting), but in this case I didn't want it to compress the extra bass away again.
EQ: Rolled off the very low end to make room for the kick and make it less "boomy". Cut the highs slightly because I want to hear it but it was masking the rest of the mix a little too much. Remember that you want to try and give each instrument its own space in the frequency range. Cut a bunch of low-mids because my bass is kind of boxy/nasally sounding there. You can get away with a lot of EQ on bass.
Compressor: Av. 4dB/Max 6dB. You also often want a lot of compression on bass guitar to keep the notes at roughly equal volume, and increase the note sustain. I think some people pretty much just slam bass guitar with tons of compression the whole time. What I tend to do is watch the gain reduction, and set the release and threshold so that there's quite a lot of compression when the note hits, and just about zero gain reduction at the end of a long note. As the note decreases in volume as it's held in reality, I'm pretty much keeping it at equal volume with a minimum of compression applied.
Sometime bass guitar just works for me, and sometimes I still spend ages trying to get it to sit right. Bass is hard. Make sure you've got a set of speakers that you can play your mix on which can play low bass!
Recorded with a matched pair of Rode NT5s at either side of the piano at the front, under the keys, with the front panel off.
EQ: I get the best sound from my current piano by recording from the bottom (under the keys) with the cover off the bottom half of the body.
Compression: As stated earlier, the piano was compressed using a hardware unit and recorded back in - a TL Audio 5021. I normally use a medium-to-fast attack and fairly slow release, but I use a low ratio and don't compress very much. If you search forums, everyone's saying use light or no compression on piano, but many pop and rock artists are blatantly using tons of compression. e.g. listen to the starts of Porcupine Tree's Piano Lessons or Anna Nalick's In My Head.
Reverb: OK, so there are basically two reverbs that come with Pro Tools 8, D-Verb and the AIR Reverb. I don't really like either that much and in reality I tend to use Wave's RVerb or TrueVerb. There are lots of reverbs better than these. D-Verb in PT 8 LE onward has quite a blatant chorus effect on the Hall and Church sounds (Digidesign says it's "subtle") which can kind of mess up the sound of things. D-Verb is a really basic reverb, but what I do like about it is that it never overpowers things. You can add quite a lot and it's always pretty smooth and stays in the background.
I used D-Verb here because when I tied the AIR Reverb, it came through quite harsh, even when I messed with it turned down the treble a bit. Note that the wet/dry mix is automated - the reverb ramps up at the end. In fact, if you listen to the wavering of the reverb tail sound from the final piano note, that's D-Verb's "subtle" chorus effect.
Pipe Organ Synth
EQ: A minor highpass filter to stop any low rumbling. Note that Pro Tools' frequency display is logarithmic: the low cut is way less range than the high cut. The lowpass frequency is automated, so at the start of the track the organ has all its treble, but as the other instruments start it rolls back to how you see it now, so it fades into the background a bit behind everything else.
Reverb: Again, automated. Lots of reverb as the song starts, then it fades to dry.
Recorded with a matched pair of Rode NT5s, one pointing at the fretboard and one at the bottom of the body. Don't record too close, or aimed at the sound hole, or you'll get tons of bass.
EQ: Massive low cut! Some tracks cut the acoustic guitar so much that it just makes that clicka-clicka sort of sound in the background. I don't normally cut quite as much as I have here, but I think I was recording a little too close so it was a little boomy originally. When an acoustic guitar is playing along with other instruments, you can usually afford to cut a lot of bass from it, and it will tend to actually make it clearer in the mix. The extra dip at 3K is making a little space for the vocals.
Reverb: Just another hall kind of thing.
Comp: Av. 1.5dB/Max 2.5dB. Why is the compression before the EQ here? It probably shouldn't be - we'll be compressing the bass which we later remove. Anyway, the compression here is very minor so it doesn't really matter.
EQ: Another huge cut to the bass. Same reasons as the acoustic guitar above.
Delay: Since this instrument has lots of abrupt, on-beat notes, I've used a delay instead of a reverb here. The track is played to a click so it's matched to the tempo in Pro Tools, which means I can also easily time the delay to the beat (here the delay is a crotchet, or quarter-note). I always use the basic Digidesign delay plugins. It's a delay - what features do you need?
EQ: A distorted guitar recording is naturally compressed (or at least clipped) and I didn't add any compression here. Like the acoustic guitar, a lot of bass is rolled off to prevent low rumbling that I don't care about (and to prevent that rumbling from masking the bass instruments I do care about). The second dip at 1K is just to get the sound I wanted - every electric guitar recording sounds pretty different.
Recorded with an AKG C4000 B with a pop filter.
Comp: Av. 3dB/Max 4dB. As you can see from the volume automation of the main mix window, I do a lot of manual volume riding on vocals so that I don't have to do so much actual compression. Vocals can have a lot of dynamic range, and you want to keep some of it, but probably not as much as there is naturally. I'm pretty much using the Vocal Comp preset here, with a pretty low ratio and not too much total compression.
EQ: Lows are cut a bit, but not so much that it makes the vocals sound thin. If vocals are recorded very close to the microphone, unless it's set to Omni, the proximity effect will give it more bass. Sometimes that means you need to roll off more. Your microphone itself may also have a 100Hz rolloff switch that you can turn on, but the best thing to do is just not record super close to the mic.
Mids and highs are boosted a bit to add clarity and presence in the mix. That 3K cut on the acoustic guitar is complemented here with a 3K boost, so the clarity of the vocal can easily come through without a big boost.
Reverb: Reverb on the vocals is bussed out to a separate track, so there's more control over the volume of each and increasing the reverb volume doesn't decrease the vocal volume like it does with a wet/dry slider. Some people will always put effects on separate tracks. Naturally the reverb is 100% wet here. I tried the AIR Reverb here but it always really brought out the sibilance on the vocals - even when I rolled the treble way off - and it didn't sound any good. So D-Verb it is.
Comp: Av. 3dB/Max4dB. Both compressors are the same except for slightly different thresholds. They basically match what the main vocal compressor's settings. There's one for each track (since the backing vocals won't be singing exactly the same at the same time), but then the tracks are bussed out to an auxillary track that handles both. This makes it easy to change the volume and settings for both the backing vocals at once.
EQ: Basically the same as for the main vocals, but the bass is rolled off a lot more. This is partly because they're female vocals so they don't start as low, but mostly because I just want the backing vocals to be a bit "thinner" than the main ones.
Reverb: Pretty much the same as on the main vocals. There's probably a little more reverb on the backing vocals overall.
Blip EQ: This instrument is meant to be just a thin, trebly sound in the background, hence the huge cut to the bass and low-mids. This is the quiet Electric Piano-sounding synth that has a delay on it (the delay comes from the synth's own software instrument).
Nothing EQ: This is sort of like a combination of the sound of Blip and a pad. This is meant to be very much in the background, hence the shelf cut to the high-mids and highs. Cutting treble will always let an instrument fade into the background more, and boosting treble will bring it out. Sometimes it's better to cut the bass and bring the volume up rather than boost the treble, since a boost is always more obvious sonically than a cut.
Stereo Buss (Mastering)
EQ: A steep cut right down at <30Hz to get rid of any super-low bass, which is inaudible but nevertheless can be there. A minor shelf boost to the highs on the whole track. I hardly ever do more EQ in mastering than this. Since I'm mixing and mastering in the same place, it's normally better to go and fix individual tracks. However, a lot of people do mastering in separate programs (or send their tracks away for mastering) where a lot more EQ may be done. You should always, of course, try to make the mix sound great before mastering it.
Other: The way I master, is almost always EQ->Compressor->Limiter. I'm using Waves stuff here with something from the maximizer series for the limiter. So I EQ as above, then apply a bit of careful compression to the whole track, then limit it at some point to bring the overall volume up a bit.
I spent a lot of time on mastering for In Our Dreams We're Flying, getting the tracks sounding good and at roughly the same volume as each other, as well as all having the same sort of treble/mid/bass balance. A lot of this was editing instruments individually in the mix, rather than true mastering.
Please don't be one of those people who squashes their mix to be ultra-loud with no dynamic range (see: The Loudness War), especially if you don't have a fancy look-ahead limiter. You can use the limiter to bring up the overall volume a little bit by pushing down the big peaks and bringing the whole track back up to 100%.
After mixing, make sure to listen on different sets of speakers (and headphones!). What sounds great on studio speakers can sound terrible on a Sony mini-system, and the key to mastering to to make a track sound as good as possible on all systems. It's especially important to listen on something which can reproduce low bass.