One of those "how did they get that sound" questions

Could have been a flanger or chorus used very subtly. More likely it was just double-tracked very tightly, with both tracks mixed at equal volume and possibly pitch-shifted a few cents apart. It takes a very precise singer to pull the trick off.

Great vocal yes. Bass line is cool though. It might be too good for double tracking but maybe some very subtle modulated effect for sure.
 
Ben could well be right with the Leslie concept below. Feeding a vocal out to a Leslie and micing it and mixing it under the original vocal track is a nice idea. It can give that liquid movement effect to anything that it is applied to. You also get that very slight pitch shifting going on too as the horns come near and move away from you or the mic.
 
Choral seetings are usually a little faster than the effect on this track maybe (not sure) I modified my Leslie to run slower in chorale when I had one.  Fantastic for synth pads. I had two choral settings.

Leslie speaker eh? Well that's interesting. I might give it a whirl, I think I have a Leslie simulation in Amplitube. 

I get flangers and phasers mixed up. I really need to sit down and listen to the difference between them sometime. I guess as a guitar player I never really used stomp boxes - if I had, doubtless my ears would be more attuned to the character of each. For instance, the backing vocals on Bryan Ferry's "Don't Stop The Dance." Am I right in thinking that's a phaser? 
 

Yes, a phaser is a phaser set on stun? A flanger is a phaser set to kill... Van Halen got some interesting effects using both. 

The Band19
Yes, a phaser is a phaser set on stun? A flanger is a phaser set to kill... Van Halen got some interesting effects using both. 

That has reminded me, I have the Waves Enigma plugin (as part of a bundle) and after all this time I've never really bothered to check it out. It actually sounds very cool. 

bitflipper
A flanger or chorus would be the easiest way to get the effect. The primary difference between the two is the length of the delays: flangers using longer feedback delays than choruses. 

Other way found. Flangers use a shorter (modulated) delay time than chorus, plus a variable amount of feedback sending the wet signal back through the input. The pitch rise and fall comes from varying the delay time in a bucket-brigade delay circuit, same thing as an analogue delay's repeats changing pitch if you set the feedback high and alter the delay time. Flangers generally work with 5-10ms delays, chorus is typically 20ms or so.

The feedback in a flanger circuit means that blending the wet signal back into the dry creates an automatically sweeping comb filter, which gives flangers the "jet plane" sound. Flangers where you can turn the LFO off (e.g. Deluxe Electric Mistress) are comb filters, with the number of "teeth" set by the amount of feedback.

This is in the analogue world.... Some digital choruses simply clone the dry sound and pitch-shift it up and down a bit, repeat for as many metallic, icy voices as required.

And tape flanging, running two tapes side by side and alternately slowing each down by touching the flange on the spools then letting it speed up to pitch, which is how the whole thing is supposed to have originated, is something a bit different again.

Leslies are strange things. Such a simple concept in hardware yet so difficult to emulate at all convincingly.

The univibe and its' various clones/sound-alikes were the original emulation attempt of course, and are still amongst the best things for doing the job in my opinion. Way better than chorus effects attempts at a rotating speaker sound. Perhaps because the univibe "chorus" setting isn't really chorus at all but a primitive phaser with a "lumpy" LFO waveform that's blended with the dry signal. I suspect because, unlike chorus, phasing doesn't smear transients or timing slightly.

Not the same as a leslie of course, but much, much cheaper and you don't need a fork-lift truck to get them on or off stage either.

Started watching the Kenny Gioia delay course. That was one of the ones I flicked past and thought "hmm delay....don't need any lessons there!" Of course all of these Groove3 courses are excellent and give you a ton of ideas. 
 
Didn't really have much luck with the Leslie-type simulation in Amplitube. But what I have done is finally pay attention to some of the Waves modulation effects which I have as part of a bundle but have never really checked out seriously. I'm getting some really cool sounds by slapping MondoMod on a bus and mixing it in with the dry vocal. Enigma is very interesting too. This is just what I was looking for. The autopanning in MondoMod is getting me close to that Chrissie Hynde sound!