I've tried finding tips on this, but they generally fall into doubling one vocal track, so here it goes: I've got three vocal track to mix together for the chorus so I'm just looking for advice on how best to mix unison vocals together; levels, panning, reverb, or whatever will help make them blend together in a powerful way.  Thanks in advance (because I may not get to log back in until Sunday).

So three vocal takes all singing the same thing, same notes, same timing etc?

If so, I'd typically choose the best one for centre and pan the other two left and right. How wide depends. Probably wouldn't go 100% as it'll leave a gap which I personally don't like (between centre and hard side image). Doesn't mean it's wrong though. Artistic decision. Then knock the outside levels down a bit. Also be careful with sibilant and plosive sounds. If the timing is slightly off on these it doesn't sound like one big vocal any more and can sound off. Sometimes singing without them or softening it in the backup parts can blend better. Almost sounds like you're singing with a lisp. Or you can manually edit them out, or align perfectly. Probably compress the backup ones a bit harder as they will be lower in level so you'll want them to always cut through. Even using a different mic too could work. Maybe something less bright if you want a warm, thick sound to engulf the main lead. Even a dynamic mic and then a condenser for the lead could work nice. Really a lot you can do. Can you elaborate more what you want? Maybe with examples on professional stuff?

Matt's approach is a good one and one that you might use but I have also done the opposite and fed all three vocals at the same level to a vocal buss and processed all three at once so they all have equal weighting. I agree with Matt too in that panning is probably best in close proximity to each other as well as opposed to far L, Centre and far R etc. That can sound a bit spread and weird. It tends to separate them a bit too instead of bringing them together.

This can sound great but it does also mean they have to be sung pretty tight in terms of timing. This approach will inevitably require some editing of words and phrases of two of them against the main one you might consider to be the lead. In terms of reverb if I use an all three at once with equal weighting I would send all three into the reverb. If you only send one into the reverb you will still get nice reverb but it will have a singular quality rather than the group quality.

Personally, I love the sound of just 2 unison voices, panned no more than about 20-30% L+R, with identical levels, sends, Fx etc

Every approach mentioned here and pretty much any you can think of has probably been pulled off with success. Anything can work really, assuming it's applied well. I think the one Bristol mentions tends to be a pretty common one for pop these days. Used typically in a chorus. Often will take a hell of a lot of tightening to get it to work as well as you hear on the radio. Lots of hours go into it. There might then be a big stack of harmonies below all this too adding meat and thickness to the sound, but with two main leads kicking out at front.

I was thinking before that an interesting thing to do would be to have your main vocal or two, but then have a big wall of underlying unison harmonies all singing with 100 wet verb. And then spread that right out to the sides and keep the main vocal pretty much bone dry. This could make the vocal sound like it's huge, thick and full in a big stadium, but dry and out front at the same time. Because the verb for the main vocal is actually taken from a different take, there is no relationship to the two and it can't pull the main vocal back. No idea if this is common practice or not, but I think I love it! Got to try this one soon... Maybe it will become my sound!!

I've been working on this very thing with the last few tunes I have done. 

On my soundclick site (links on my website) the song Missing Person (2012) has a total of 10 vocal tracks in it. The 2 leads are unison... in several places they (2 different singers) are singing together. They also sing solo in the track. 

When they are solo, I have them panned 10% to a consistent side.... when they sing together, the panning goes to 35% or so to give them distance. 
The 8 tracks of BGV are mixed much lower and panning is a bit more one sided, but there is balance. The 10 vocal tracks soloed as a group, present a fairly wide stereo image. 

Now: I have also used 3 tracks as you describe. I generally use this to "thicken" the main solo vocal in a track. The song is not a "group sing" but is intended to be a solo preformance.  I do choose one.... the better one IMHO to be down the center of the mix. It is set to the center of the stereo field and the level is up where the singer should be in the mix.  

The other 2 tracks are panned R/L by at least 30% up to about 60% but generally no more.  The panned tracks must be as close to perfect as possible to the lead. Timing, pitch, and all aspects of the performance of these tracks is critical. Especially the ending of words with s's in them.  

I normally will send these tracks to a bus with a slight verb.... panned according to taste....and with levels set low. I want the levels so low that when I solo all the vocals I can barely hear the BGV tracks. Once the instrument mix is put into the picture..... that tends to mask the BGV a bit but they are still there, at the same level and the result is a thickening or fattening of the solo vocal track. I used this method on Ode To Wolf on the lead..... there are also harmony vocals in places where I kept the levels low to add the harmony without being too obvious with the singers...

I use envelopes a lot. Volume and panning.  For example.... on Missing person... 2 singers in 2 different studios recording the unison vocal track...... no way will that come out picture perfect. And it didn't.  In one place, my vocal sang a word and stopped..... Robby sang that same word and held it slightly longer. (the second time in the song ... "touch herrrrrr"  right before the chorus)  That is in a part of the song where we sang it together in unison..... the solution was to use the envelope to pull the level down on Robby's word so it ended with mine.  The result sounds good.  Same thing on the "s" sounds at the ends of words.... I use envelopes to make the ones not in sync disappear. One track ending properly will give completeness to that sound.... 3 ending slightly off  from each other sounds sloppy. 

There are no hard and fast rules. I have developed this method through trial and error and I know as time goes on these methods I currently use will evolve as I learn more about mixing. 

Experiment and see what works for you. 

The reason you won't find any generic answer to this question is that how you mix unison parts depends on the desired effect. 

Double-tracking to thicken a lead vocal: pan the double up the middle, roll off the high end around 5KHz and compress aggressively. You can either bring it up just until it starts to thicken but isn't quite audible, or you can go for a full-on Jethro Tull / Genesis chorus effect and put the double at the same level as the primary track.

Simulating a big choir/chorus: pan each part 10 to 50% apart, and if you have a true stereo reverb pan the sends to opposite sides. Alter the EQ so that each one has a different tonality, so they don't all sound like the same singer. You can optionally apply a chorus plugin and/or delay to all or just some of the parts.

In either case, suppressing transients will help meld the parts. Let the primary vocal track carry the "t"s and "ks" and use a de-esser or automation to soften them for the doubled parts. 
Try varying the reverb amount, mixing the doubled parts either drier or wetter than the primary vocal. Making them dry adds weight to the vocal while preserving clarity, while making them extra wet can add width.

Thanks.  You all are brilliant and have given me lots of ideas to try.  I'm dying to get back to the DAW and start working on it (so that means you're inspirational too).

I turned around to toucherrr and she slapped me "hard."

I'm with the doubling and panning, but nudge the doubled tracks to get more of a wide sound. (Stereo Separation/Widening. There are plugs for such a thing, but if you rather not go that route).

Have the doubled voices behind the lead or even amongst themselves off by a few ticks. 5-10 usually works for me as anything more will make them sound obviously off.

This method, to me at least eliminates phase if one comes across it when the voices are stacked on top one another. It's not only good for vocals, but other doubled instruments as well.


You can even transpose the vocals/instruments by a few cents to get "chorus" instead of just applying said effect.

Have the lead sing @ 440, but the doubles @ 450.