| || |
We talk about Stereo all the time, we record it, edit it, mix it, master it and listen to it. Don't we?
In most modern styles of music the vocal is THE selling point of the mix. You'd think that deciding how loud to make it should be easy - make it louder! The problem is that a vocal tack that's too loud often sounds "stuck-on" to the track, and if we push the supporting instruments too far back then we lose the power and drive that moves the track forwards. It's very common for a mixer to provide several versions of a mix, with the vocals a bit louder (or softer) in each case - "Trackname vox +1", "Trackname vox +2" etc. This is good practice, but we still need a starting point. Here's a suggestion.
Once you've done your mix and finished your automation (I can't remember the last time I did a mix without automation to keep the vocal clear against the instrumental tracks), adjust the overall vocal level so that it sounds about right (if you've done lots of automation and you're not happy to "trim" this then send the vocal output to a group track so that you can simply adjust a fader without disturbing the work you've already done).
At this point, you might want to make a quick list of what you think is important in the track - guitars, piano, banjo? Write it down.
Now drop your master fader or your monitor controller all the way so that you can't hear anything. Slowly fade-up the levels until you just start to hear the prominent parts of the track. What are the first things you hear? For most rock/pop type tracks, if it's the vocal/snare/kick drum then you're probably going to be pretty close to where you want to be. What do you hear next? If you've got a piano driven song and you're hearing the strings way before the piano then you might want to revisit some mix choices.
One quick thing to note - if you export your mix as a MP3 or other data compressed format then recheck the vocal levels before you send it out; the data compression can really mess-up the balance of the middle/sides of the stereo field, and as your vocal is almost always all (or virtually all) mid it can end up at a quite different level than you expect it to be. The same is true for aggressive loudness processing on the master - you may find that your vocal is suddenly a lot more prominent than you intended ( I usually check a mix with enough master buss gain to give me an average level of about -15 LUFS irrespective of what level I'm monitoring at).
Drop me a line and let me know that's useful to you.
Choice is good, so more choice must be better? Right?
Well, in the wee-small hours of this morning it wasn't. Despite my attempts to keep a trim plugin folder (TheGreat Plugin Purge) I've acquired a few extra goodies over the past few months, including Waves Gold bundle (what can I say - it was on a "too-good-to-miss offer here in the UK) and a couple of new Neve inspired modules for IK Multimedia's T-RackS (which must be the hardest-to-type software I use).
Anyway, it's been a mix-heavy week (good!) and I was wrapping-up the last bit of work before the long weekend. I found myself (early morning remember) with both the IK and Waves versions of the Neve 1081 console eq on a track, and I was making small changes (sub 1/2 dB small) on each, and, here's the insanity, ABing the different plugins to decide which to use. And that's how good choice can be.
Eventually I gave-up and went to bed. This morning I dipped 2.2dB using Cubase's stock channel eq at 362Hz and problem solved, in about 12 seconds!!!
Now seriously, this isn't JUST about having choice. If you lose your mental image of what result you're actually trying to get-to, and when you stop hearing what you've got (Fighting the urge to "do-it right"), then your stuck anyway, but it sure is easy to fool yourself into thinking youre making decisions whan in realithy you're just playing with the toys!!
Here's a question, what are your "I couldn't live without it" plugins? What if you could have only three (besides your DAW standards)?
Many years ago, when I started to get involved with recording and mixing music, I bought a notebook - it's nothing special, just a small cheap thing with a silver carboard cover and wire rings to hold the pages together. In this book I write useful things that I find in magazines and on the internet or that I discover for myself. I have Andy Sneap's multiband compressor setting for distorted guitar in there, and Mixerman's frequency bands, and Graham Cochrane's magic mixbuss move, and a thousand cut and boost frequencies to make mixes come alive, and compressor settings and goodness-only-knows what. The book sits on or by my desk, and to this day I still write things in it; and by now most of the good stuff is probably in there several times over. It's probably fair to say that it contains an unsorted jumble of just about everything I know about recording and mixing and mastering! Except - I'm not sure, because I don't think that I've ever actually read anything out of it.
I was working on a mix this afternoon, and reached a break-point with drums, piano, bass, 2 x guitars and a banjo all coming -together like one lovely very big and very complex instrument. I had it just about there and was struggling to get the drums to sit properly - eventually achieved with 1.1. dB of low shelving-boost at 330 Hz and 2.2 dB of high shelving cut at 1.5 kHz. using the (at this time) new IK Multimedia EQ81. Just like that - it all came together. I grabbbed a pen and the book, and realised that I actually had no idea what to write. The numbers are just chance - just what happened to work for this particular job. The clever bit (if there ever is one) was to know, or to work-out, or to guess that the kit needed a different frequency envelope - deeper and less bright; but I didn't work any of it out - I grabbed a low shelf, boosted it and slid the frequency centre until it was right, then fine-tuned the boost. Same again for the top-end cut . The actual adjustments took less than a minute and it's one of the most satisfying moves I've ever mixed. And I don't know how to write it in my special book.
The point? Well, it's easy to get caught-up in how clever all the folks writing books and articles in magazines and online are - but the reason that they get good in the first place is because they practice unti they know what they want to hear, and develop some instincts for getting it. When I tell you that I used +1.1 dB of low shelf at 330Hz, don't think "how clever to be to know to do that", I just read the numbers off the dial after it sounded right; the numbers just tell you what you've done, they don't make it happen.
I've been an intererested spectator to the "loudness wars" for several years now, and I'm a fan of what Ian Shepherd and other have campaigned for with initiatives such as Dynamic Range Day. I've made it part of my workflow when mastering projects to check the Dynamic Range stats and I even used to state that I'd work to a certain DR level by default.
I haven't actually looked at a Dynamic Range figure for several months now, and I think I'm starting to realise that I just don't care any more. When I'm mixing I mix for the song and the sound - always: I don't give a damn about level, Just balance! If a mix has a low DR then live with it - that's how it's meant to sound.
Here's a funny thing. I've been working on some video content for the site. In one of the videos I'm illustrating the effect of an EQ move - showing a change between 4 and 2 dB of gain on a high-shelf. This is the sort of move that we make all the time, usually leaning forwards between the monitors or clamped into headphones, tongue sticking-out between teeth, all full of concentration so we don't miss any tiny detail.
I am absolutely mortified. I have a copy of Toontracks EZmix for a magazine review. I've been mixing a country flavoured piece today and I've been struggling with the vocal. In a coffee break I fired-up EZmix and surfed a few presets. The 4th one I tried is perfect - it absolutely nails what that track needs. I now have 2 small devils yelling at me - one says "leave it on and move ahead" and the other says "you HAVE to stop and figure out what it's doing"..
What do you think? Is it OK to "just use the preset"?
I well remember the excitement of the first time I opened a set of audio files that I hadn't recorded myself; that I'd never previously heard, and set-them-up for my first ever 3rd-party mix. I remember the tingle of nervousness as I listened through them and tried to figure where the heck to start and what I should try to do. I also remember stacking-up every book and article I'd read and being afraid to make a move without checking-in with the gurus. I made notes about everything I did, made many, many mixes, and looked-forward to the flash of inspiration, the eventual "got-it" moment. I evaluated and trialed and tried hundreds of plugins, waiting to find "The One". Well, no flash, no "The One".
Along the way I've found a very small number of gems, things that add significantly to the pot. Ermin Hamidovik's "Systematic Mixing Guide" appeared a bit late in my story but consolidated many pages of my own notes and added a few more besides, and Graham Cochrane's Recording Revolution is almost worth having the Intenet for on its own: Mixerman's "Zen And The Art of Mixing" is irreverent and insightful and funny and anyone who wants to mix should read it beforfe their first mix and following their 50th. But still no flash - perhaps I'm just a little slower than most people.
Now many years ago, when I had a real job, I found myself in a pub with a colleague, trying to figure (over a pint of warm, flat English beer) how we managed to learn to do what it was that we did (we were both Enterprise Sales-people). We couldn't find an answer and eventually forgot the question. Today I had a "revisit" to that moment. I finished a mix this afternoon and at-last, "the-sort-of-flash" happened. That's a lower case "t" and a lower case "f".
What's the sort-of-flash? It's that mixing is truly a game of a thousand mundane moves. It's actually quite unexciting and there is no boom. It's about learning some basics, then forgetting that you've learned them and applying them over and over and over again until your sensibilities start to take shape. It's about knowing that something's wrong or right before you know what it is, it's about learning techniques then forgetting that techniques even exist, it's about accepting that the music is more important than the sound, that excitement is more valuable than perfection, it's about developing your senses, about finding nuggets and not mistaking them for answers, and it's about realising that you ain't gonna find the answer in anybody's "Black Friday" sale.
I just read a statement online that said "wav unquestionably sounds better" than mp3 (sic) - I don't disagree with the sentiment but I figured perhaps I'll test how noticeable it really is.
Here are 2 x files. Both are 44.1/16 wavs. One was initially exported as a 16 bit dithered wav from the initial 32 bit floating file, the other was a 320kbps mp3 that was encoded from the same 32 bit source file. Both were imported into a 16 bit Cubase project and then re-exported as 44.1/16 bit files.
Check 'em and let me know how close you judge them to be?
For the record, I chose this track because it's simple and exposed (OK, and because I had it handy); I know it doesn't test dense or loud or swirly cymbals - you can't have everything :-)!
I flipped a coin to decide which track became "1" and which became "A" and then flipped again to decide which order to place them on the page, and then again to decide which way-round to ask the question below, then again to decide which way round to put the options, so it should all be pretty free from bias on my part.
I was working recently on a couple of tracks of solo vocal that I recorded with peaks at around -12dBFS. They were single takes, unacompanied, and I really didn't want to do much with them, just a 1/2 dB or so of compression and a bit of reverb, so there was no "mixing" to be done, skip straight to the master stage.
Now I'm a great believer in proper gain-staging in the DAW and I prefer to not bang into plugins with hot levels. I use Cubase which has trim at the beginning of each channel, so it's trivially reasy to set a healthy level into a plugin chain. Setting plugins for near-as-damn-it unity gain gives a relatively low output level, and smashing the guts out of it with a limiter then takes us to the expected listening levels - but in this case I wanted to have NO limiting. My peaks were rather lower than I needed for the final version, so I needed to apply clean gain - but which way?
Overall there are several ways to skin this particular cat. This particular performance really needed to be clean and so I ended-up re-testing a number of assumptions about how gain works in the DAW. The different aproaches are really about workflow rather than effect - they all work. I'd be really interested to know what works for other folks?
All original content copyright The Dustbowl Audio 2019
Other material copyright acknowledged