What’s the purpose of mixing a song? Is it to create an accurate record of music that an artist has played? Or is it to create a sound recording that is larger than life? As an engineer who records only in-studio performances, my answer to this is influenced by the unique circumstances of recording live music for broadcast. For an example, we’ll look at a recent session with Gary Clark Jr.
Evaluating the Original Mix
Here’s the first version, which was mixed and recorded live during the performance.
Some notes I made while listening to one pass:
- Mix is mostly mono - Guitars very prominent in low midrange, masking bass - Bass and kick sound muddy..low end rumbles along - Vox could use more dynamic control - Vocal effect is recorded and committed. Different acoustic space from everything else - Gtr 2 changes tone and arrangement role, gets lost in the mix - Foundation doesn't push me forward enough - B-section and solo sound special - the music is awesome - Series of ascending builds which plateau, then start to build again - Drums recorded in wide stereo, doesnt support the otherwise narrow mix - No phase coherency issues noted, tracks seem reasonably well recorded
You can tell that instead of featuring a very hooky chorus, it relies on groove and musicianship to keep listeners interested. Not an easy thing to do. Let’s listen again to only the first minute.
- Rhythm guitar generates lots of low end energy - Can't hear intro guitar 2 solo - Cymbal crash has a pleasant sheen to them - I'm feeling the groove in the heavy 1&3 kick, sharp 8th note rythm guitar, and gtr2 doing an 8th note suspension on 4 leading into a staccato chord on 2. At least it sounds like gtr2, hard to tell with everything masking each other. - Bass does some fills, buried
Developing a Vision
By this point I get the sense that the song is awesome, but the mix isn’t supporting it as well as it could. The mix for this song needs not only to portray great sounds, but it also needs to convey a feeling which suits the song’s message. You can hear that the elements are in there, but they’re not communicating the artist’s message as effectively as they could – they’re not pulling me into Gary’s world. This is where vision begins. Personally, I start sensing an emotion in my chest of what I want the song to be… some feeling that drives all your actions while you mix as you try to make the song evoke that feeling.
I ask myself an array of questions, in no particular order, before and during the mixing process. What will the song sound like? What should it feel like? When you close your eyes, what kind of space are the instruments in? What is the song about, and does your vision complement it? What subtle changes can you make throughout the song to keep the listener’s brain tuned in at the subconscious level? What’s most important in the arrangement? What is driving this song? Who has center stage at a given time? In my view, mixing is all about using technical and creative knowledge, skills and equipment to achieve a sonic vision. So, where do we start?
From Vision to Reality
Like most things, its best to get chores out of the way first. Unnecessary audio was removed, undesired resonances were tuned out, I made sure each track had enough gain to hit the console at normal operating level, cleared up blatant masking and low frequency rumble issues… basically, I took care of as much of the annoying stuff as possible on the front end so that the rest of the process could be freely creative. This was all done in mono.
Now you can really listen to song without getting distracted, and you can listen to it from different angles. I start listening to groups of instruments, and the parts they play. Are they fighting each other? In my vision, is the instrument or part closer on the sound stage or farther away? Should it be distressed and urgent? Or is it an afterthought? The answers to these questions and many others will inform your decisions on EQ, compression and effects processing. Once you really understand the song, its message, its form, the arrangement, and the source material at your disposal, you’re ready to start.
Here I’ll usually throw up some faders and get rough levels going (in mono), but after that the procedural aspect of mixing starts to devolve as you wind your way through various cyclical decisions. It’s hard for me to describe this part. You engage both the technical and the creative sides of your mind simultaneously while you use equipment, tools and techniques to create sound. Here is a partial list of some of the tasks during mixing:
- Corrective and creative equalization
- Corrective and creative dynamics management (including compression, expansion, triggering, gating)
- Decisions on levels
- Decisions on panning
- Application of reverbs, delays
- Automation of any/all of the above
The following thought process is typical:
Let's throw the vocal up and see what kind of bleed it contains. That's the focus right now, but I also need to establish the foundation and groove; this song relies on it. Let me mute everything and pull up the kick and bass tracks and work on them until I have the balance, tone and feeling that I want. Do they need EQ or dynamic correction? Or should I let them fly? Definitely need to compress the bass - I want it to be thick but clear, and rock-solid dynamically when he goes into higher registers for fills. Will need to pull some fundamentals out of it to let the subkick through, though. [I EQ and compress accordingly]. Unmute channel and listen in context... does it fit the song? Did my last action bring the mix as a whole closer to the vision I have?
And so, you wind down the rabbit hole, layering in the foundation, leads, pads and fills, and making a series of decisions that affect decisions you’ve already made, until eventually you reach a point of balance. Not just balance in levels – you reach a point of artistic balance where the song is communicated in a compelling way. Listener interest is retained, if not through changes in arrangement and structure, then through subtle changes in your mix. Mix engineers are not automatons who just read manuals, turn knobs and push buttons. Anyone can do that. Mix engineers create a vision, and possess the technical knowledge, musicality, foresight and experience to connect with the musical art and achieve that vision.
Once the mix is balanced, I set about making it even stronger. How can you make the mix support the song even more? Usually around this point I’ll go to a prominent section, and automation ideas start coming to mind. (In this case, I started rehearsing fader movements). And throughout the process you listen through different monitoring systems, at low level, at high level (briefly), from down the hall, through a crack in the door, in mono and in stereo. Each different angle you listen from can prompt you to make a change, and pull you back into the cycle once again.
It should go without saying that the previous few paragraphs don’t fully describe the mixing process. At least for me, there’s a lot to it that I can’t even find words for.
The Final Product
We’ll listen to both versions. Here, again, is the original 1st minute, followed by the newly mixed first minute.
- Tone is different on rhythm guitar, and reverb/delay effects bring it closer to effects on vox - Gtr2 solo is prominent but pushed back in the acoustic space - Rhythmic groove is established with kick/bass complementing Gtr2 suspensions on beat 4. Complements the vocal melody and the agitated rhythm guitar. - Vocal compression pulls up the amp's reverb tail, make it more musical with regards to tempo
Here’s the whole song. (The accompanying video is at the end of this post).
If you listen closely, you’ll hear the live automation as guitars tuck in for vocals, then back up for solos, etc. One important decision I remember revolved around how loud Gary’s guitar solo should be at the beginning of the B-section. From there the song plateaus 2 more times while it continues to build. It was a tough call but I decided to hold him back at first, and leave room to really push the guitars forward towards the end.
Considerations and Limitations
To be honest, I avoid doing secondary mixes of our live performances unless the recording is really messed up (but correctable) or I feel very strongly about the song or artist. There’s a lot that happens in the moment, and many of your decisions are based on the auditory and visual cues you pick up on while the performance is happening. Taking the tracks back to the console is tricky because while you want to preserve the spontaneity and vibe of live recording, mixing requires you to deconstruct the song and build it up from the ground floor.
Secondly, as a recordist of live performances for radio and web, I don’t have the same type of creative license as a producer/engineer may have at a commercial recording studio. While I usually obtain the artist’s trust during the session, once it’s done they’re out the door to the next engagement. So when mixing them long after they’ve gone, I need to be judicious about using any production techniques that may significantly change the sound of the first recording. Additionally, since we record in one 24’x24′ room, bleed is inevitably part of the equation, and it limits some of your options – especially in the area of dynamic processing. While most artists I’ve worked with wouldn’t care at all as long as the product sounds good, the last thing I need is backlash.
Finally, it’s important to note that we work within the realm of diminishing returns. In reference to an “ok” mix, spending 5 to 10 hours mixing a song from the ground up could yield a result that’s only 10 or 15% “better.” That said, it’s often just that 10 or 15% in quality which makes the difference between a track that sounds good, and a track that sounds polished, professional and compelling.
No matter how much you want to go on, there comes a point when you have to put the mix down and commit, forever, to your work. In the digital realm, you can save, quit ProTools, go play in the sun for a few days, then listen to it again with fresh ears. This isn’t as easy to do with analog mixes. You can log your settings, write down effects routing, knob and fader positions, hell you can even take pictures. But the second you pull those faders down, you’ll NEVER get that mix back. In this case, I’m reasonably happy with the result.
For comparison, here’s Gary Clark Jr. at KCRW in mid-February. Really nice long version of the song, which kicks in at 0m45s.