Layering Vocals Tips

Hey guys, I'm currently working on a hard rock project, it's for a very good friend of mine so putting in the extra mile to really make it as good as it can be.  The singer has a very Alice In Chains / Soundgarden sound, and the vocals are definately the highlight of the music.  I'm looking to do some serious layering (more than I normally would put the time into) to make them ultra huge, and this is new territory for me.  Just looking for any general tips anyone might have.... panning, etc, and especially opinions on editing, as I don't want to be editing for the next 3 months, so trying to decide what the most practical way to do this is.

Vocal chain was SM7B > G9 > SA-3A > Profire 2626.  This guy can nail it, so not going to have to be doing much pitch adjusting.  I'm running Pro Tools, normally would just use the warp feature for timing corrections, but wondering if I should maybe invest in melodyne, if it would be worth it or not.

In any case, just curious what some of you guys are doing, any help is greatly appreciated.

I don't want to be editing for the next 3 months

Click to expand...

The best way to do avoid editing for the next 3 months is to not edit.

Seriously.

Don't go into the project expecting to edit anything, much less expecting to buy a new program and learn it just to edit the vocals.

The quickest and best sounding way to work, providing the singer can sing, is just to do takes until he nails it, then repeat for the dbl, tpl etc

Then you might need to nudge the odd word or syllable, but you shouldn't need to cut it into a thousand pieces and stretch every waveform blob to line up - and surprisingly, it'll actually sound exciting and full of energy if you do it right.

If this is a band, and the singer can "nail it" you shouldn't assume it's a good idea to be doing intensive editing right off the bat. People made HUGE vocals before melodyne.

A lot of the "HUGE" comes from things being a little bit spread out in terms of pitch and timing - making it look "perfect" isn't always the best idea.

I've clocked many hours in melodyne, autotune, elastic audio & beat detective etc - they are amazing tools. But they take time and experience to know how to use them and more importantly when to not use them.

I would concentrate on:

1. getting great performances
2. layering them, and listening to how they gell
3. try different positions on the mic, and performance styles (i.e. combine far shouty with close growly etc)
4. experiment with panning, levels and other basic mixing techinques
5. do anything that makes it exciting and engaging to listen to.

These simple things are far more effective than getting bogged down in computer land.

Then if you or the band are still not happy, look into doing some intensive editing, but at least you'll have something great and exciting to work with.

Hope this helps! good luck!

Oh, and in my experience if all you need to do is "stretch/compress time" of a mono voice, melodyne often isn't the best sounding tool... you can use it for this but I wouldn't suggest buying it for that purpose.

I'd use Avid's X-Form (in monophonic mode), which once you own it you can assign to your stretch tool in pro tools (check PT preferences).

But there are others that also sound good, just experiment, get the demos and try them out - some stretch algorithms sound much better than others for different applications...

also maybe something to note, depending on how it all goes together...

sometimes one honest true blue vocal track is WAY more impactful then a bunch of FANCY over dubs and yadda yadda.

Exponentially when your dealing with talent.

experiment to the moon, just don't loose sight of the ground.


edit: ok something maybe more constructive. I think the right delay/delays on a vocal can make a huge difference.

I like to layer vocals and other things.  I also work in protools.  Obviously you'll need a good comp to layer things under and it is very helpful to work in the grid.  Once I have a good lead I'll go to a new track and start play-listing the layered track.  If your vocalist has some time with a reference CD of his lead it would make it much easier.  I usually will come up with 5 or 6 playlists and start building a comp, copy paste to an empty playlist, of the most exact phrases to the lead.  Sometimes if the performance is great but the timing is off a bit I will time-stretch the phrase/word or sometimes I use a plugin called VocAlign.  I rarely will use the original vocal track playlists because I usually end up using a phrase that is already in the lead.  Another thought is that I usually record more things than would be used in the end.  It seems like a waste of time to do this but I like the idea of presenting options.  Things don't always come out on tape like they are in your head so don't be surprised if you don't like the the layers in the end. The most acceptable layer is one that thickens and does not sound like an "effect", IMHO... Also I tend to use an autotune on layered tracks even if they sound perfect, I figure it's there just to insure that things don't jump out at the listener.  This becomes more important to me when you add multiple layers.  Timing, Pitch, Dynamics and VIBE are my major concerns when layering.

It might be easier to speak about technique rather than typing it.  Drop me a PM with contact info if you wanna geek out...

Hi,


    I approve of the dont-think-youre-gonna-edit approach, but I would strongly recommend getting Melodyne to touch up the odd lapse in tuning. You will hardly notice the tuning, but  overall perfomance can improve by leaps and bounds.


  DOnt tune all the bvs if they are layered. they will seem to phaSE with each other, especially if panned together, making it smaller not bigger. If I have a block of 4, I comp the best pair of vocals  for tuning, and tune the other two, and pan the tuned ones  on the outsifde left and right.

I agree, making them too perfect can easily make them sound smaller.
Also, If you are layering for a chant or crowd effect it's better to have several voices, as
one will start to cancel itself out .... too similar tone etc.

MM.

I've been listening to Lenny Kravitz first album a lot lately and his vocals are all over the place. Lots of plosives, timing on the doubles is not tight, certain phrases are way out of tune. I mean it's not a mess but it's definitely not super tight and it works great.

Best,
jonathan

Hey Ruckus,

Well I usually compress the heck out of any backing tracks, reduce their volume level until they start to disappear, pan them out a little, then add a nice high shelf boost to bring the presence back that I lost with the volume reduction.

If the music is more old school country-ish / old rock n roll I'll use a high shelf CUT without so much volume reduction.

Well if he can nail it then yeah - try mic positions and experiment with stuff, you'll get the most amazing
things outta just simulating 2-3 different room circumstances and seeing how they come together.

We'd fake these huge sounding orchestras with it and a lot of it had to do with the depth perception that
would happen when we intentionally went for a more diffuse thing for backing - mic about a meter away,
big room, you know - and then getting the lead to gel into that (awesome word)

One fave was to use stereo micing in a relatively big room and then put the musician once to the left, once
to the right, so you'd have kindofan ORTF stereo background picture with a certain amount of room verb
already on it. That way, you can create way more illusion of depth and position because there actually IS
depth and position. Mic pair was about a meter or two away for that one, at the center of the room.

There's also the old technique of doing multiple takes, moving the performer each time, while keeping the mic stationary...

This is great for simulating a big ensemble of strings with only a couple of players - but it definitely works for vocals too.

This technique works especially well with stereo micing, but also works with mono micing - it's more obvious with omni patterns or more reflective rooms. All the reflections and spatial cues change with each take - bring up the faders and you should get a really deep sounding ensemble.

Try doing the lead vocal right on the mic, and then two DBLs a few steps back and far on either side of the original performance.

I believe there's somewhere on the web where Bruce Swedien talks about doing this on Michael Jackson records...

EDIT - Oh, I think this is kinda what livingnote is talking about - if i understand correctly?

Yeah,


  FOr instance, I used 4 singers to make a 64 strong choir, setting up for a 64 member choir, and moving the people for each take(16 times). - Sure saves on session fees!


    ANdyP

Right on, I was thinking we're on the same page when I read your first post.

One of the most awesome recordings I've ever heard was an entire orchestra AB mic'd with no processing after that (was from the 20s or 30s).

Thing they did was: Orchestra plays for 6 hours straight while the engineer gets to fiddle with mic positions.

This led me to crawling around orchestras for 8-channel recordings at the main rehearsal, just listening for positions with one ear closed. You'd get this one exact position where the harp was deep and crystalline with the winds floating beautifully creamy over top. 30cm left right forward back, the harp would needle, the winds would spike, cello would rasp, and so on.

Ah Andy you too and big - yeah, I don't know how your experiences are but sometimes I've noticed that the
simulated-by-doubling choirs and pads had a way to them that was distinctly different than recording
"the real thing", but especially for songs had this unique flavor to them that fit in a lot better in the end.

  for sure! for one thing, the character of four individuals is NOt the same as 64 differenent peeps. they are often far tighter, since they are repeating themselves.


  . . . . but the budget often dictates . . . .

I remember coming from doing layered stuff and then hearing the real thing on a recording

"oooh...thin."

Hey guys, sorry I've been MIA, so much crap going on.  Thanks for all the input so far, lots of good bits. 

My main concern isn't so much backing vocals but rather doubling/tripling the lead vocal for a thicker sound (less monoish, but still center focused?), without it starting to sound like 3 guys singing.  The backing vocals obviously still play a big role as there's alot of them as well so the input regarding that is still greatly appreciated, never really thought about the issue of tuning all the tracks causing phasing, good stuff to know.  For those I've been having him stand back a few inches off the mic (didn't take much to really change the sense of space).

As far as the leads, I've tried messing with using a doubler, it helps some but as soon as I push it until it really starts to make it full it also starts to cause a noticable phasing effect.  I did go and do some quick tests tripling the lead (Main lead center panned, other 2 hard panned and compressed pretty hard and quieter in the mix).  Just used the warp feature to tighten up all the words so they're dead on and bam - there it was, that thickness (or pretty damn close anyways).

Slightly off topic, but still relevant to the matter at hand, any tips regarding delay?  Better to use mono delay or stereo?  I'm struggling with it to say the least.  Trying to get it to duck while he's singing so it really only kicks in between phrases but can't seem to get it right.  Have the vocal triggering the sidechain on a compressor on the delay channel, but doesn't seem to be working too well, maybe there's a better approach?

Again, thanks for all the help guys.

ruckus328 said:

Slightly off topic, but still relevant to the matter at hand, any tips regarding delay?  Better to use mono delay or stereo?  I'm struggling with it to say the least.  Trying to get it to duck while he's singing so it really only kicks in between phrases but can't seem to get it right.  Have the vocal triggering the sidechain on a compressor on the delay channel, but doesn't seem to be working too well, maybe there's a better approach?

Click to expand...

I love mono delays on vocals. I use a separate return channel and ride and automate a lot.  I love to keep the listener guessing, keep the delays moving and changing without becoming a distraction.  It can be nice to roll off top or bottom on the delayed channel, or subtly distort it or wash it out with verb, whatever. One more thing, DAWs these days seem to want to lock everything to the session BPM!  Go manual and move the delays around, just a little early or late can really add energy or create a groove.

Really wanna build one of your 4000 jobbies, need time!

Enjoy the mixing,
Ruairi

I'm generally a fan of mono too - whenever possible (not just for delays!) - leaves more space for everything in the stereo spectrum if it's made up of lots of individual mono sources, panned around - rather than a bunch of wide stereo tracks all hard panned and stacked on top of each other. It depends on how many tracks there are - in dense arrangements I would go for mono when possible. For very sparse tracks there is more room for true stereo things - IMHO.

But this is a generalisation - there are no rules. Whatever sounds good is the right choice.

ruckus328 said:

Slightly off topic, but still relevant to the matter at hand, any tips regarding delay?  Better to use mono delay or stereo?  I'm struggling with it to say the least.  Trying to get it to duck while he's singing so it really only kicks in between phrases but can't seem to get it right.  Have the vocal triggering the sidechain on a compressor on the delay channel, but doesn't seem to be working too well, maybe there's a better approach?

Click to expand...

Try compressing the vocal on a BUS, rather than as an insert, and bus all the vocal effects to the same compressor - and "mix into" the compression, i.e. push the vocal fader up so that it will "push down" the level of the effect during the vocal, and then the effect will rise to the top after the vocal line is finished. I think this is a more organic way of achieving what you are trying to do with the sidechain - although that method can work too.

Combine that technique with fader rides for the delay (and pans etc if you like) - and things should be work well. But personal taste and the taste of the band will inform you if it's the right thing or not.

H

Hi,


  I often like to use an h3000 on the lead, with something like plus 8cents panned left and minus 5 cents panned right, no delay




      often much better than doubling the lead.


ANdyP