I make a lot of food analogies when I’m talking about mixing, mastering and whatnot; I make no apologies for that. Food analogies are great for a number of reasons;
– they are universally understandable
– they are often actually more relevant and accurately descriptive than they first appear
– they give you ideas for dinner
Far and away the most useful – or at least, most used – food analogy I draw on is The Cake. I must refer to cakes in some way or another every other day; it’s just so simple and effective that everyone can understand it, can learn or reinforce important, fundamental concepts from it, and can return to it when they are in danger of trying out ‘fancy stuff’ that is likely to ultimately lead them down a cul de sac. So without further ado, I present The Cake.
Building a mix is exactly like making a cake. You have certain ingredients, mixed together in certain proportions, which are then baked together before being decorated/iced/smothered in cream and jam. The ingredients are chosen with the flavour of the end result in mind – lemon, chocolate, guitars, screechy wobble bass, whatever. This determines the flavour (but not the quality; this is important) of the final cake. You can’t make a thumping house track with only acoustic guitars any more than you can make a chocolate cake with lemons. You get the idea.
Once you have the ingredients ready, it’s a question of these three main factors;
– using the right amounts of the main ingredients
– using certain methods to prepare each ingredient optimally
– using the right amount of generic, general ingredients to bind the main ingredients together
So, you get your ingredients (chocolate, synths), you do what is required to make sure they will work in context (melt the chocolate, choose the right synth sound), and use generic ingredients to get it all to work (eggs, flour, sugar, reverb, delay). It is of crucial importance to stress the individual, ingredient-specific nature of this; melting chocolate in a pan works really well; you can’t throw a lemon into a pan and expect it to melt. You can crush nuts by bashing them with a rolling pin; this tends not to work so well with marscapone cheese. Each ingredient requires specific individual attention in order to contribute fully to the result.
When the ingredients are ready, you mix them together. It’s time to bake the cake.
Baking the cake – buss processing, dynamics, and toasted nuts
This is where The Cake (or any other cooking analogy) really comes into its own.
In these days of – apparently – having to buss compress absolutely every single everything, this aspect of The Cake is the one I refer to most. Applying processing – particularly dynamic processing – on busses binds the ingredients together permanently. It effectively bakes the cake. When your cake mix is still a liquid/dough/whatever, you can still make adjustments to it without too much concern about side effects. Once the cake is baked, it’s a different story.
To take an example, you’ve got all your ingredients together, so you taste the cake mix and realise it needs more sugar. No problem. You add more sugar and mix it in, taste again – perfect. Easy, no hassle, no side effects.
Now say you didn’t add that bit of sugar. ‘It will be fine,’ you think, ‘I followed the recipe to the letter’. So you bake the cake, take it out of the oven and taste a piece. ‘Bollocks,’ you think, ‘it needs more sugar’. At this point you are left with no option but to try and ‘cheat’ by adding sugar all over the top, or putting extra icing on it, or some other ad-hoc remedy. Maybe all that sweet icing will distract people from noticing that in truth, the actual cake itself isn’t very good.
It probably won’t.
Better to get the ingredients right, to get the mixture right and thus get the cake right, than to bake something soggy and disgusting, then try to make it nice by throwing icing all over it.
Other comparisons highlight different aspects of mixing. Say you’re making a cake with crushed, toasted nuts in. You could;
a) toast the nuts, bash them up, add them to the mixture
b) bash them up, put them in the mixture and then put the mixture under the grill to toast the nuts
c) put them in the mixture, put the mixture under the grill, then bash it all to break up the nuts
It’s fairly clear that b) and c) are completely arse-about-face. I can imagine people thinking ‘only a total moron would do b or c’. And yet, effectively, I hear people doing just this in their mixes all the time. They haven’t prepared their ingredients correctly, and instead resort to throwing processing at the result, rather than taking the time to address the root causes, the ingredients.
All this is directly analogous to buss processing. If you mix into a master buss processor by default for whatever reason, you are – to some extent – doing one or all of the above. People very commonly mix into heavy/drastic bus processing from the very start, then wonder why their mix sounds wildly different – and usually awful – when they turn it off to ‘help’ the mastering engineer. They’re covering the whole thing in icing before they have taken good care of the cake itself.
That’s not to say that buss processing is bad, or necessarily to be avoided; just that it is secondary to correct individual treatment of the ingredients. It’s lazy, and it is suboptimal. A crap cake covered with good icing is still a crap cake. The cake should taste good even without the icing.
When the ingredients have been taken care of correctly, you might well find that the 8-10dB of buss compression you needed to keep control of everything becomes only 1dB, or even completely unnecessary (I’ll be writing something about tapering of processing at some point).
So, to really emphasise the point; when you’re making a cake, you don’t start with the icing. You don’t even think about icing until the cake itself is finished. The worse the cake is, the more stuff you need to put all over it – but it never tastes as good as when the cake is right in the first place.
The Cake In Real Life
Enough analogies now; let’s look at a classic real world example. I see this at least three times a week, and is, in part, the origin of the concept of The Cake.
Bass – let’s assume it’s a classic sine sub bass – can hard to get right in the mix. A combination of small monitors, acoustic considerations, inexperience and just liking things to sound fat invariably leads people to have their sub rather too loud. We have a track with ‘skinny drums and big sub’. Let’s assume this sub isn’t continuous, it has a rhythmic stop-start pattern. The sub bass in question is also 10dB too loud.
Now once the mixdown is done and it goes to mastering, that’s not a huge problem – sub typically hangs ‘off the bottom’ of a mix and has no harmonic content to interfere with the rest of the mix above. Too much bass? Cut it back with eq. Easy. No significant side effects, no nasty surprises.
What happens if the same mixdown was sent to mastering with a lot of heavy master buss compression/limiting? Well, a number of things, which I’ll describe below.
From a dynamic point of view, the compression will keep the overall level more consistent. Now sub bass is big enough at the best of times, eating up headroom like nothing else. When you compress a mix with heavy sub bass, the bass will inevitably cause everything else to duck; when the bass plays, everything else gets quieter.
[N.B. I’m only discussing standard/full band compression and limiting here – the article is long enough as it is, and I’ll address more specific considerations in the future. Sidechaining is a very particular case but it does bake the cake]
So what happens now when the bass is cut back to the correct level? Your track is now jumping in and out to bass that isn’t there any more. Say you’ve limited the track aggressively, because you’re worried it won’t sound loud enough. Instead of (just) the pumping/ducking from the bass, the bass also now causes lots of lovely intermodulation distortion (fuzzy distortocrap) every time the bass plays. This time, when your long-suffering mastering engineer cuts the bass back to the correct level, you hear the distortion even more. There’s no getting rid of that.
This is like burning the cake and then scraping the burned bits off. You can still taste that stuff, no matter how hard you try to get rid of it and hide it. Nobody does that to a cake; it’s absolutely plain common sense. And yet people do exactly the same thing to their audio all the time.
Key point for real-world mixing: You had better be VERY sure your mix (and that includes a group/submix) is good before you bake that cake. You’d better be willing to bet your house on it.
So there you have it. I could write lots more, stretch the metaphor yet further, and doubtless cause more confusion in doing so. But I’d rather leave it there, and simply ask that you remember this;
When mixing your track, ask ‘would I do it like this if I were making a cake?’