I've been slamming my tracks into the ableton limiter (well not slamming) but it honestly sound terrible for anymore more than a few db of gain reduction. I know nothing of the technical aspects of producing music and it has been driving me nuts.
Like how does volume actually work? I know this sounds sort of like "magnets how do they work" but how do I know how loud something actually is. Why do two sounds that both peak at -6db individually peak higher when played togheter. I need a book or something that just gets me going on understanding the basics.
http://www.sengpielaudio.com/calculator-spl.htm
You absolutely do not need to know any of that to mix well, but there it is.
I struggle with panning too - and basic mixing. I also don't understand the whole production -> mixing -> mastering pipeline when working as a lonely solitary producer. Especially since 95% of my music happens in the box and I rarely use any of my live instrument skills - mostly because I have terrible recording equipment and I find the in the box stuff sounds more powerful and as a result I find it hard to mix the live stuff without resorting to dumb shit like duplicating tracks a bunch of times or ridiculous compression+gain staging.
Mixing is all about creating space. You create space a lot of different ways; leveling, EQing, compressing, panning. The trick is knowing when to do each of these things, and that is partially dictated by the style of music.
EQing is HUGE. The audible frequency spectrum runs from 20Hz to 22,000Hz (22kHz) - this is the full range of what the human ear can hear. We divide that spectrum into three areas - lows, mids, highs. How these areas are defined is a little nebulous, but there is a rough consensus.
What are the lows? I consider the lows to be roughly from 20-250Hz. Someone else might say as high as 300 or 500 Hz. This can be broken down further into two areas - bass, and sub bass. Sub bass would be 20-60Hz. We call it sub bass because you can't really hear it, but you can feel it. The more audible bass spectrum would be from 60-250Hz.
I consider the midrange to be from 250-4,000Hz. Some might go as high as 6,000Hz for defining the upper end of the midrange. This is where the most action takes place - your melodies, vocals, snare drums all sit in this range. This range is also broken down into two areas - low mids and high mids. I'd say low mids range from 250-2,000Hz, then the high would be from 2k-4kHz. Low mids are where the meat of these sounds are, while the high mids are where the clarity is. A snare drum really knocks around 2,500Hz, vocals sound nice and crisp around 3k-4kHz
The highs would be everything above 4,000-5,000Hz. I consider this the area where shit starts to get annoying. This is where brightness comes in. And the super high frequencies like around 16,000Hz and up is kinda like sub bass in that you can't really hear it too much, but you can sense it, like an airiness. I'll give a tiny boost to a cymbal around 10,000Hz to give it a nice little shimmer.
Another helpful thing to know: Pitch is intrinsically tied to frequency - ie: a pitch of middle C rests at 261.6Hz. You don't need to know the frequency of every pitch to understand that when you're EQing a given instrument, you generally want to accentuate it's natural frequency range while hiding some of the frequencies you don't need.
So, for a bass guitar, we know where the body of that sound is - do we cut it off above, say, 500 or 1,000Hz? Not necessarily; the pluck of a bass guitar string sits way up in the 2,000Hz area - you probably wanna keep that pluck audible in many cases, but in between 1,000Hz and 2,000Hz, you can remove some stuff, and you can remove most of the stuff above 3000Hz. But those frequencies above and below the meat of a given instrument are resonant frequencies - frequencies that, while not carrying the full weight of an instrument, still hold some of it's presence, and thus are still important. If you remove too many resonant frequencies, you lose a lot of warmth (another nebulous term); basically, what I mean to say is you don't want to build walls around your sounds when EQing and make your mix too cold, you want to cut off just enough so that they they have their own space in the mix but still blend naturally into the other sounds without intruding too much - those overlapping resonant frequencies help with the blending.
By the way: the above two paragraphs are total bullshit - we hear professionally made EDM routinely cutting off melodies and vocals down to like 500Hz all the time for the sake of an arrangement transition. Alternately, something that might seem counter-intuitive is boosting a resonant frequency to mask the harshness of the meat of a sound - if a guitar is annoying you for some reason around 4000Hz, boosting a little around 500 might actually hide that harshness. I say all that to say this: don't let EQ box you in creatively. Move it around however your need to. They're versatile enough that you can really sculpt the shit out of a sound the way you would whittle a boat out of wood, only with EQ you could periodically change that boat into a car on a whim if you really wanted to.
More about placement
As a general rule, you want things in the low end to be as narrow as possible while things in the upper frequencies get wider and wider. What do I mean by narrow and wide? You're sitting in front of two speakers, one on the left, one on the right, but the sound coming out of them sounds like it's directly in front of you - you're listening in stereo, this not exactly mind blowing information. But you can change the perceived positional presence of a sound so that it either sound's more directly in front of you (narrow) or more spread out to your left and right (wide).
Let's try something: get a bass instrument going in Ableton, and put Ableton's Utility plugin on it. Slowly increase the Width parameter all the way up to 200% to spread it out. What happens? The bass disappears into the stereo field - we never ever ever want this to happen. Bass, as it turns out, only sounds good when you keep it narrow; turn that Width all the way down to 0% to bring it back, and you can feel it in your chest like you're supposed to, it'll be directly in front of you. Bass is not something you wanna spread out. Similarly, a kick drum is not something you wanna spread out - that thump needs to be dead center or else it's useless.
Try the same thing with a vocal or synth; the wider it gets, the more ear-candy-like it becomes - this is good. Just don't don't overdo it. You wanna spread out sounds in the upper frequencies, like guitars, piano, snares (to a small degree OCCASIONALLY), cymbals, vocals, a synth. How you do this is entirely up to you, but I tend to just go wider the higher - for me a violin would be wider than a cello.
A free stereo tool for Ableton and non-Ableton users:
http://www.fluxhome.com/products/Freewares/stereotool Same principles apply, plus it's nice to be able to see it visually.
Panning kinda goes hand in hand with widening and narrowing in that you need to do it selectively and don't overdo it - I never hard pan anything all the way left or right. I like balance; if I got maracas panned slightly to the left, I need a hi hat or a wood block or something going slightly to the right - I absolutely hate when a song has too much going on on one side and it sounds lopsided.
You can definitely combine panning and widening to great effect - I like a wide ride cymbal panned slightly to the side, for example.
Another cool trick for widening things: Delays.
Throw Ableton's Simple Delay on a synth or something in the mid range. Unlink the left and right, click and change the sync to time, set the delay on both L & R to 1ms, turn the feedback to 0% and the wet/dry to 100%. With these settings your synth or whatever will sound pretty plain, directly in front of you, nothing special. Change the delay time on the right side from 1ms to 10ms. Even such a small difference in the delay in the output of the right side will trick your ears into thinking the synth just got wider - it didn't, one side is just lagging behind slightly. I do this on vocals all the time. It's a very neat trick.
Common problem: Getting a kick drum and a bassline to coexist. Sometimes you'll hear one cancelling out the other and weird phasing issues. Solution: EQing and/or sidechain compression. If you look at a kick drum through Ableton's spectrum analyzer, you'll see most of the thump is peak centered around 40-50Hz, or 60-75Hz, or maybe around 110Hz sometimes.
So, method one: set your bassline EQ to remove only that frequency. I like Ableton's EQ8 with a pole set to notch with the narrowest Q. When the bass is solo you'll hear a slight loss of low end in that spot, but when the kick drum is playing alongside it, the bass will actually wrap around the kick and the bass will sound as if you hadn't removed anything from it at all.
Method two: throw a compressor on the bass, activate the sidechain and set it to accept input from the kick track, set the fastest attack possible, release around 5ms, ratio at least 2:1, threshold as needed, but not too drastic. What's gonna happen is the bassline is actually going to duck out of the way whenever the compressor detects that kick drum playing. Increase the release time if you want the bass to stand up again more slowly (you've heard this effect in EDM alot, that pumping synth sound). I also like sidechaining vocals against melodies, like making a guitar duck down a little when a vocal is playing. You just gotta be subtle with it.
Don't even get me started on using compression though, if I use compression for an individual item it makes mixing that item exponentially more difficult - I recently noticed the first kickdrum of my song was always significantly louder than any other kickdrum in the song turns out it's because my attack on the compressor I was using was too slow or the release too long? I honestly don't know why I had a compressor on that kickdrum though other than a general feeling of I should because it's what everyone does when I watch youtube videos of the producing. If I use compression on a mix and the settings sound good for one very busy section it might make another quiet section sound completely blown out and super loud and vica versa.
That's a common mistake - doing things that don't need to be done because you saw someone else do it. And I'm guessing you're overdoing it. As a general rule, I will only use drastic compression ratios on individual instruments. I never compress an entire mix more than 2:1. And I like to use a soft knee on a whole mix - a hard knee will make the pumping effect of the release phase more apparent. I'll use a hard knee on individual instruments I need bring out the attack on, like drums or some kind of plucky instrument; I like when those pump sometimes. If the sound doesn't have a tail, keep the release short probably. When it's time to go for loudness - LIMITER.
So you have a lot of choices to make, but you get to make them creatively, not just technically. If you've got two instruments occupying the same space, you gotta decide: maybe one goes left, the other goes right; one dips at 3kHz, the other at 500; one is wider, the other narrower; one is more compressed, the other more dynamic. You'll probably be combining some or all of these on any given instrument.