Mixing so you can understand the lyrics

[djgout]

This is something that was burned into my brain very early on in college. During one of the first production classes I had to take the professor comes in and says alright let's listen to this song. He then puts in a cd and presses play, at the end of the track myself and everyone else in class were ready to sing back the melody and point out cool parts in the music. So the only question the professor asked us about the song was "okay, who can write down all the lyrics or tell me what the song is about?". Fuck me was my answer, being a musician I've always paid more attention to the instruments. The whole lesson was that the lyrics are what is telling the story, the music is the setting. That's something that always sticks out in my mind as it's so obvious but also very very easy to overlook, especially when we think we're working on a cool track.

Anytime I'm driving with my fiance and we're listening to classic rock it drives me insane, she isn't a musician but knows the lyrics to every song on the radio, while I typically have no clue what the lyrics are but I can sing and play all the drum fills no sweat. Even the wizardy dungeons and dragons Rush-y shit. It absolutely kills me that she knows them and I have no clue.

The story can not get lost in the background.

As far as techniques I like to start a mix with just the vocals. I want to get them all dialed in and compressed if needed to actually get into the story before the band puts up the scenery. From there the rest just adds all the support.

[RefD]

This is something that was burned into my brain very early on in college. During one of the first production classes I had to take the professor comes in and says alright let's listen to this song. He then puts in a cd and presses play, at the end of the track myself and everyone else in class were ready to sing back the melody and point out cool parts in the music. So the only question the professor asked us about the song was "okay, who can write down all the lyrics or tell me what the song is about?". Fuck me was my answer, being a musician I've always paid more attention to the instruments. The whole lesson was that the lyrics are what is telling the story, the music is the setting. That's something that always sticks out in my mind as it's so obvious but also very very easy to overlook, especially when we think we're working on a cool track.

right on, well said.

[snoopy23]

I'll be brief (for once). I had a similar problem with a band I recently worked with, a basic punkk trio of drums, bass, guitar. The vocals weren't cutting through as well as I would have liked, and it seemed like any tweaking I did left the whole thing sounding out of balance. I finally decided to run the vocals through my Aphex 104 Aural Exciter (with Big Bottom: the best name of any piece of studio gear ever!!) After fucking it up with too much bass enhancement at first, I ended up getting a nice vocal sound that blended into the mix but maintained it's definition, rather than adding eq and making it too sharp, which was what was happening previously. You might want to try running the vocals through some sort of enhancer or exciter to add some second and third order harmonic distortion, it may help you a bit. Just be careful of adding too much or it turns to mush pretty quick.

Just a thought...

[DrummerMan]

cool suggestion. I'll try it out for sure. Thanks.

[Dakota]

I'm actually thinking about it now because I just made a demo of a song that will be the basis for a satirical music video.

...and if you don't understand all the funny references being made, most of what you'll see on the screen will be meaningless.

...I ended up having to crank the vocals through a decent comp plugin, EQ'ing them so they cut more, and raising (with automation) the gain until there could be no doubt what was being sung. It worked, but, again, I feel like the impact of the song as a whole is a bit lost as a result.

...a better singer on the mic, but in the interest of making this thing as good as it can possibly be when the time comes, what specific recording techniques do you use that are different when this approach is needed, so that the lyrics can stand out without the vocal track taking over everything?

Yeah, if music has a comedic/theatrical element, it totally falls flat if the lyrics aren't crystal clear to the average listener.

Ideally, yes, tracking with a singer who by training or talent and attention can enunciate more clearly than a typical rock singer, while hopefully not sounding too stiff.

But even just having to work with whatever vocals you have, the mix approaches are the same. (I've dealt with this kind of thing plenty).

Lead vocal will need to be edited and ridden/automated a lot, line by line, word by word, syllable by syllable in places. Looking to "tag" each moment the listener uses to understand the words.

The beginning of each lyric phrase usually needs bumped up maybe 3db, maybe even as much as 6 db higher than the "sustained" level of the line, to make it briefly pop through, then back down. Often enough, consonants like B's, D's and P's need briefly turned up. (Sometimes even pasting in a substitute B, D, or P from some other word in the performance, sheesh). If there is dumb low frequency popping in the consonants, they each need hi-passed individually, sharp knee, at somewhere around 50 to 125 hz.

Sustained vowels don't need to be loud. Sibilants might need *both* de-essing and gain riding to make sure they are just right for intelligibility. I don't like to leave a de-esser plug on a track all the way through. I usually dial in a de-esser, print another track through that, then paste back into the lead vocal only the moments that needed it. Or volume edit each one by hand.

Automating a gain plug before the compressor, and maybe after it as well.

Choice of attack time is pretty critical on the compressor. Let through enough of the transients at the top of each word.

Find any "bulky" resonances in the lead vox (could be between 200 and 800hz), dial in a fairly narrow Q and cut as needed.

Air boost w/high shelf, find whatever gets above the sibilant band. Might be 10k or 15k. Maybe automate that as well.

A side send to a plug-in chain to fake up an "aural exciter" kind of thing can help. Bandpass eq letting some choice of mids through --> distortion --> bandpass or hipass eq. Blend a teeny bit of that back into the vocal mix.

+1 on short delay. Maybe stereo, 21ms left, 34ms right, or anything in that range where they are different times and not in a simple ratio to each other. Maybe hipass and lowpass those too.

Each of these things add up. Often enough so that the vocal just becomes very intelligible and present, and doesn't need anymore to be turned up so loud in raw volume that it makes the instrumental tracks sound small.

[Nick Sevilla]

Sometimes the vocal needs to take over the track.

Sometimes it's ok not to hear every single instrument individually in the mix, but have it be masked by the vocal.

Sometimes the lyrics are more important than the music.

Sinatra - it's a good thing.

[newfuturevintage]

As far as techniques I like to start a mix with just the vocals. I want to get them all dialed in and compressed if needed to actually get into the story before the band puts up the scenery. From there the rest just adds all the support.

agree fully. Start with the most important instrument, then bring in the rest, making sure the rest of the instrumentation doesn't kill the main focal point.

[Professor]

I have to disagree with Professor's philosophy somewhat. Ultimately the listener is going to walk away humming the MELODY, and I'd be willing to bet that they are going to sing along with lyrics that aren't exactly correct.

No problem there, it really is my philosophy and approach most of the time but that doesn't mean it is the only philosophy or approach to take. And like I said, I often have to reconsider whether I'm thinking 'radio mix' vs. 'album mix'.
But I do find it interesting that it seems many folks here may hear the music more than the lyrics. I really think this observation says it all...

Anytime I'm driving with my fiance and we're listening to classic rock it drives me insane, she isn't a musician but knows the lyrics to every song on the radio, while I typically have no clue what the lyrics are but I can sing and play all the drum fills no sweat.

That is such a perfect encapsulation of why I happen to take the approach that I do:
Most people who listen to and enjoy our musical creations are not musicians.
Or at least, the people we (or at least I) usually want to reach are the non-musicians.
They hear the words and the stories and that is what reaches them. That's why they love the recordings of Sinatra, or Elvis, or the Beatles, or Neil Diamond, or Garth Brooks, etc., etc.

But I also don't want to clog this up with discussions of whether one approach trumps another when really we all need to be able to handle both and know when each is appropriate. Well, and I guess it's good to be able to give & take with the clients to find a balance between what they expect and what we might expect.

What's more important here is getting back to the OP's issue of making the lyrics stand out.
I wish I could give a blanket statement like they try to do in most text books or product catalogs like, "give a 3dB boost to 4kHz to increase vocal clarity". Unfortunately that is a fantastically generic solution that I've really never found to work. And what's more, it is built-in to damn near every microphone any of us will ever own. That's great if we were only ever recording solo voice, but when we stick a mic with a built-in 4k boost in front of a guitar amplifier, then we increase how much that guitar will interfere in that "range of vocal clarity". Plus that range is really just the home of articulation, not so much the tone and timbre which seem to come more from the fundamental frequencies and first few harmonics. And a lot of mics can over-emphasize the higher frequencies up around 8k-10k which brings out the sibilant 'S' & 'T' sounds that can be almost the opposite of vocal clarity.
But with all that in mind, it can help out with bringing vocals to the front when that is what is needed.
Personally, I start by picking the best microphone match to the voice I'm recording. I look for one that will reach in and grab all of the best qualities of that voice and reject all the worst qualities. That sounds great in abstract, but it's not the easiest thing to do. Some voices need me to reach into the low-mids, others the mid-mids, and others the high mids. Some voices need the clarity bump, while others down. Some people have lots of lispy 'S' sounds at 6k, some at 9k, some at 12k, and others have barely any at all. Sometimes I can find the right tone but need to turn the mic off axis to reduce the sibilance. Every voice is quite uniquely different, and although I might use the same mic on several people, it might be for quite different reasons, and once you figure on a couple preamp options it gets even more complex. Either way, I take that approach so I can maximize the good stuff and minimize the bad before I even hit the recorder.
Now once the stuff is recorded, it's important to make a hole for the vocals among the instruments, and as it happens, I think a big part of that hole has to be from about 1k to 10k, and two of the biggest opponents in that spectrum are guitar distortion and cymbals. If I can catch those on the way in and tame or thin the distortion a little at the amp or the mic, and choose the overheads carefully it will help. Otherwise, EQ is helpful, and I try to cut those areas before hitting any compressors. Actually for really dense mixes I might EQ, then compress, then EQ again and compress again to get the vocals right where they need to land without sounding like anything is too obvious.
Another helpful thing I've found is separating my reverb into one feed for close, early reflections and another for the long reverb. Reverb in general tends to make things sound more distant and/or like it is in a bigger space, and while I might want the vocals to sound like they are in a big space, I don't want them to sound distant, and I don't want the reverb to cloud the intelligibility. So splitting between two reverbs allows me to dial in just the right amount of send to the early reflections, and just enough to the long reverb (which has just enough pre-delay to have it hang behind the dry vocals. Then the blend of dry, close, and long are tweaked along the way as the mix comes together. Oh, and those aren't ever 'exclusively vocal' reverb feeds, I use the same ones for the rest of the band to give the illusion that they are all within the same soundscape.

I know that's not really much of an exact formula; it can't be. But at least it is a set of tools that I've found will work to accomplish this task fairly well.



-Jeremy