Thoughts about doubletracking vocals?

[surferbeto]

I've been listening closely to Beach Boys vocal arrangements as I work on a demo recording of a song that's strongly vocal-driven. I've read about Brian Wilson's use of double-tracking vocals and I'm learning about ways to do that. There are a number of "how to do it" pieces online <http://www.soundonsound.com/sos/apr09/articles/doubletracking.htm> that describe basic techniques.

I wonder if you savvy studio wizards have any general comments about double-tracking for real, and the many and varied ways to simulate double-tracking through digital processing? Are there any nuggets you could share of what experience has taught you?

 

I recognize 1000% the comment that "those guys just really knew how to sing." There is clearly no substitute for those singular skills. The more I practice singing these vocal parts, the tighter they get, and at some point it stops sounding like two guys singing in unison and start to sound like something more seamless. I haven't yet tried listening critically to mixes with doubled vocals panned hard left and right- I read somewhere that is one thing to try.

[WRGKMC]

I can say, if you haven't tried it before, it can be difficult at first. Knowing the song really well having practiced it many times helps a whole lot too. Technically its as easy ad any other multi tracked part. The hardest part is simply getting a good balanced mix with enough of the first vocal track where it neither masks the new part being sung or too weak where its gets lost trying to sing along to the part. You can try panning the playback track to one side and the new part to the other and see if it helps. Adding effects like reverb to the first may also help distance yourself from the playback,

 

Most of these choices will simply depend on your skill as a singer. If you are using to working with other singers, singing harmonies or singing the lead with harmonies going on and you don't get distracted, then you probably wont have any problems doing it in unison multi tracking. If I can do it reasonably well I'd say most singers can with a little practice. I do allot of my own music and often make up melodies and words on the spot. This can lead to issues with word pronunciation, pitch, vibratos, accents on specific beats and unique timings and dynamics which can be difficult to recreate the second and third times around.

 

But it is exactly those small variations that make the parts sound fatter. Otherwise you may as well just copy tracks and add flaws to them to make it sound like the second part was actually sung. So long as the multi tracked parts don't have flaws that are overly noticeable, it should work.

 

I usually sing the part a half dozen times all on separate tracks, then pick out the ones that are best overall. Then if there's a noticeable flaw or two, I can check the other leftover tracks and see if that one part is ok. If it is I can just copy and paste the one part to correct it.

 

The other thing is to make sure you keep the same volume and tone dialed up on the mic. Get a good level and don't tweak it between takes. Try and use the same mic techniques as well. This will help minimize strange things like proximity effect and phasing issues when you mix the tracks. Plus you can always tweak the tracks mixing. You can even pull a word or sentence over if the timing is off. No one is going to know it if you use good editing techniques and choices.

 

The only thing you will notice is your voice is going to be the same. This is true for unison and harmony vocals. Sometimes you can fake a second voice and a singer can produce a variety of tones, but too much of the same thing can be worse then just a single track, especially if that persons voice isn't so hot. Mainly the issue is a persons throat length and chest cavity are the same for each track.

 

They have come out with some new tools that can help here. Antares has a program now that lets you shape the voice cavity. Its especially useful when you're using the program for creating harmonies. The big problem with pitch shifting is the voice box changes size when you change pitch. Any more then a few cents changes and you start sounding strange, especially up in pitch you begin to sound like a chipmunk. This tool helps to change the tonal size of the voice and make it sound like an actual choir. http://www.sweetwater.com/store/detail/AT7Nat-e

 

The rest is all just multitracking techniques, mixing and editing. Unison is one of the easier ones to pull off so long as you have te ability to block the other voice for distracting you. If you find that difficult, just try re-tracking the part mute the first track and hope they align properly. I know people who can do this well, I'm only good up to about 75%. I rely on the other track in the background to get it right.

[Phil O'Keefe]

One thing that really helps is consistency in how you sing the parts. If you release the note on the "and" of 3 on the first take, then you need to count and make sure you release the note in the same place when singing the overdub / double track. If you're not a natural at "feeling" that, or if you tend to do things a bit differently every time you sing something, try counting, and be intentional about trying to remain consistent from take to take in terms of timing and phrasing.

 

Some people I've worked with were just naturals at it from the get-go; even on their first attempt at doubling something, they could do it effortlessly, while many others struggle with it. The best advice I can give you is to practice it a lot - sing along with a rough mix with the part you want to double pushed up pretty hot in the mix so you can hear it clearly. Practice singing along with that until you can nail it, then go in and track it.

[surferbeto]

Thanks guys. I appreciate your thoughtful replies.

 

I understand that there is simply no substitute for exceedingly high quality original signal from the source. It's pretty easy for me to sing a part consistently on successive takes, once I've sung along with demo recordings in my car for a while. Do you ever use video to help a singer duplicate their part?

 

Initially I imagined that I'd mix the two vocals at something close to parity. As I go on, it seems like I get a better effect if one vocal is mixed rather hotter than the other, so that the second one is just reinforcing the first. Is that how you like to do it?

 

I thought the idea was to produce two separate recordings of the singer (me, in this case) singing the parts identically in every respect (at a macro scale). But WRGKMC says "...it is exactly those small variations that make the parts sound fatter." Maybe I'm still missing something here. I thought totally nailing every nuance is what would make the double-tracked vocal sound like one fat vocal instead of two voices singing in unison. Are you saying it would be best to nail all the attacks and releases perfectly, but nevertheless to have some other kinds of "flaws", small but not too big in the vocal stream? I'm trying to imagine what those might be and how to practice them.

 

I've read that it can be good to de-emphasize consonants when singing the second track, because that's where tiny differences in timing will be most noticeable. Kind of makes sense. Singing a song without consonants feels and sounds really weird- like a crazy person singing. But as a blend to reinforce another vocal, maybe it's just the thing. Is this something backup singers practice?

 

I wonder if running a de-essing processor on the second vocal could also possibly yield a good-sounding second vocal track with reduced consonants? Or would that come at a cost of audible processing artifacts best avoided? Will applying autthis process work better or less well by applying any auto-tuning to the vocal parts?

 

I just read that Brian Wilson would often double-track recordings of the Beach Boys singing together, 3-4 voices at a time! I guess that is one way of keeping the mixing simple (and conserving tracks, in the days of 4 or 8 track analog recording) by way of dramatically raising the bar for everyone's consistent performance. Those guys were just so amazingly good at blending their voices in a quiet room around a condenser mic!

 

Do you guys do it that way today? I imagined that with modern tools, infinite tracks, and digital recording, you might want to record all the vocals separately. But singers blend better when they can see each other's lips moving. How might you let a group sing together while also isolating each vocal? Can you get good signal separation recording in one room full of sound-isolating panels with windows? Or do you actually have multiple soundproof booths with windows side by side to put several singers in? A studio friend of mine has sometimes put a singer in another room, on a video monitor with the band. I guess that probably works OK as long as there's no significant latency in the signals.

 

Approximately how many vocal takes might you guys typically record on the projects you produce before concluding "that's as good as it gets"? I saw a YouTube clip of Mike Love saying Brian would make them record vocal tracks many dozens of times trying to perfect qualities only Brian could hear (with his one good ear!). Mike said Brian had "dog ears" that could hear things no one else could perceive. "He'd make us do it over if someone had an impure thought while singing." The results of Brian's perfectionism and their talent speak for themselves: historic benchmark recordings, and (for Brian) a tragic descent into madness.

 

Then no "turd polishing" is necessary. I hope I can achieve some reasonably good results without going crazy or getting too depressed.

[WRGKMC]

 

Thanks guys. I appreciate your thoughtful replies.

 

I understand that there is simply no substitute for exceedingly high quality original signal from the source. It's pretty easy for me to sing a part consistently on successive takes, once I've sung along with demo recordings in my car for a while. Do you ever use video to help a singer duplicate their part?

 

Initially I imagined that I'd mix the two vocals at something close to parity. As I go on, it seems like I get a better effect if one vocal is mixed rather hotter than the other, so that the second one is just reinforcing the first. Is that how you like to do it?

 

I thought the idea was to produce two separate recordings of the singer (me, in this case) singing the parts identically in every respect (at a macro scale). But WRGKMC says "...it is exactly those small variations that make the parts sound fatter." Maybe I'm still missing something here. I thought totally nailing every nuance is what would make the double-tracked vocal sound like one fat vocal instead of two voices singing in unison. Are you saying it would be best to nail all the attacks and releases perfectly, but nevertheless to have some other kinds of "flaws", small but not too big in the vocal stream? I'm trying to imagine what those might be and how to practice them.

 

You can copy a track, Try it. Every detail will match. All it achieves in doing is making the part louder. Or use two mics at the same time. You may have a little phase variation or frequency response differences but the parts will still be coming from the same voice with no variations.

 

You can add different effects to each. One trick is to copy the track twice. Pan the copied tracks slightly left and right then use a pitch shifting plugin like Autotune and pitch the one copied track up in pitch a few cents, and the other down in pitch a few cents. This provides some fattening but it still fails to achieve the same results as actually singing the parts twice.

 

If you could sing the parts exactly the same it would be no different then copying the track, but because this is physically and mentally impossible, the best you can do is get it as close as you can so no major bloopers draw the attention of the listener unless they are listening very carefully. This is where you have to decide when the second part is "good enough"

 

Even if your timing is perfect, word pronunciation dead on, mic distance exactly the same and all the other variables match, there are still going to be variables you cant match that are going to set them apart. Unless you're a zombie, What you think moment by moment when you sing cannot be repeated and those thoughts influence the parts emotional content.

 

No human can perfectly match sine waves. They may be able to maintain a steady pitch, they may even be able to match vibratos, but the actual sine wave at say middle A occurs 440 times per second. Think of it this way. When the sine wave goes positive it pushes the mic diaphragm in and pushes the playback speaker out. When the sine wave goes negative, the mic diaphragm pulls out and the speaker sucks in, all at 440 times a second.

 

Even when you sing in pitch, the two tracks sine waves being produced will only match one out 360 degrees. That means the second track may begin at zero volts on a positive swing while the first track is already on a downward swing at 80 degrees. The slightest variance in pitch of either track can cause the waves to catch up and sync or drift to reverse polarity.

 

What this causes is note beating. This is a key element the ears use to tell the two parts apart. The beating may be a super slow drift, or a faster beat that can be heard cycling. Even if all the other parts of the voice match perfectly, Pronunciation, timing etc. this note beating is what's needed to make the voice sound fat. Its the voices natural chorus effect, the same thing you get with a chorus pedal for a guitar and shift the waves phasing by using a delay mixed with a dry signal.

 

If the chorusing between two voices singing in unison (or harmony) are pleasing the listeners will hear transparency. If there's too many beats happening in the wrong places, those faults will draw listeners attention away from the groove of the music and instead of following the flow second by second, they stop and think about that flaw and loose the groove.

 

You as a singer must plow through the parts as close as you can and hopefully any flaws that are there are below the normal listeners concentration threshold and remain subliminal. This way they absorb the emotional content of two tracks hearing them as one, instead of detecting the flaw in one track and have their attention split by it.

 

Musicians ears are usually pretty good at pulling this off and you'd be surprised how many minor flaws go completely unnoticed. Its just takes practice. By the way, I've always found it best to do your second take immediately after the first while you'd still in a groove riding that same emotional peak. Waiting for days to come back can be a more difficult, not only getting back in the groove, but because your ears sensitivity isn't going to be the same. If you're like me you wind up redoing the first part again. Again, try it an see what works best for you. Maybe the differences are exactly what the music needs. Only you can make those decisions and know through trial and error.

[onelife]

Interesting vocal on this track...

 

 

 

The instrumental is mono so if you playback with the left and right sides out of phase then you will only hear the doubled vocal part.

 

I don't know if they used the ADT (automatic double tracker) on this or if it was just sung that way but it certainly is a distinct effect.

[Anderton]

If you're in a hurry, this technique is surprisingly effective. Something like VocAlign is great, but expensive. I've been know to cut parts and slide them to match other vocals...

 

But at the risk of being heretical, more and more I find myself NOT doubling things and reducing the number of tracks so each track has more impact. Don't get me wrong, layering can definitely be fun and useful: I did something like eight tracks of vocals to do vocal backgrounds for the "title line" on my most recent music video.

 

[video=youtube;F33jgLw-uJI]

 

However I still put another vocal, mixed like a "real" vocal, on top of it. And when it went into the verse, single vocal with a little slapback echo...nothing more, except when another harmony line came in, which was also a single vocal with slapback echo.

 

Regardless of how many vocals you do, I think it's still important to have at least one that can really connect as directly with the listener as possible.

 

 

[surferbeto]

Wow- lots to think about and practice. Roger on the comments about recording the two vocal takes back to back in the same session, WRGKMC. I've done my best work to date in that way. Thanks for sharing the recordings y'all. I'll listen and think.

[MarkGifford-1]

IME, double-tracking a Ld Vox is something which is used for effect - to emphasize a chorus, or something. It's immediately discernible as being doubled, and can work very well in the context of the song.

 

Not a big fan of using it to try to fix a bad lead vocal.

 

Very effective use on the verses in this song - they're phrased very loosely and it really creates a cool vibe:

 

 

 

Double-tracking on BG Vox is another artistic choice - do you want the smooth, produced BG Vox of Queen, or something rougher, like what Mick and Keef do? Up to the artist/producer.

 

As far as the actual technique goes, if you don't do an actual double, you don't get the true effect, IMHO. A good singer will be able to do it as tight or as loose as you need.

 

When mixing, I'll generally put a Ld Vox double about 6dB down from the main vox, which gives it a more subtle effect. Bring it up, if you want it more pronounced.

 

MG

[the fall]

If you really stink at it…there is always this.

 

http://www.waves.com/plugins/reel-adt

[Lee Knight]

I've noticed that different voices take two different doubling techniques. The guy am working with now sounds fantastic with the old fashion straight up double and all the imperfections of pitch and timing and tone present. Very old-school 70s sound he's got going. I wouldn't change it for anything. My voice, on the other hand, never really sounds that great with this type of treatment. I've got to double ferry very precisely, and even fix pitch to match the original so it is almost spot on. The pure and more fluid like, or rather sine like, The better the loose and easy double will sound I found. Bob Dylan on the other hand is going to sound funky as hell doubled. Then again they're all sorts of exceptions to that rule, if in fact it's a rule which it really isn't.

[Lee Knight]

Please excuse the Siri auto corrections above. :-) two = to

[Anderton]

I've also found some voices like to be spread a little bit in the stereo image, while others seem to want to go straight up the center.

[Phil O'Keefe]

Don't forget to play with the levels of any doubles relative to the original vocal and see what works best for you. It's often better to pick the stronger performance, then use the other 6dB down or so for the thickening...

[witesol]

As far as gear or method technique ideas, there are a few things I've done that have worked over the years. Of course nail your parts like everyone has mentioned. My particular voice doubles nicely but my consonants pop out and I'm "Ssss'y" . If there is time I will break out my darkest, often crappyest( don't know I've ever written that word) to do the double. If I've used a nice LD mic on the lead, I'll break out a beater sm58 for the double. Lately it's been an EV PL10, the darkest mic I own. Also I've tried the Fathead with good results. Even something weird like a harmonica mic can add personality.

 

On my double track I tend to use HPF and LPF to narrow the freq range, especially the LPF can really soften the sibilant consonants. Sometimes I'm running the double totally dry, sometimes it's wet and the lead is dry. By having two tones you can choose what the reverb or delay sounds like. If you put delay on the softened double it can sound more analog ish and tuck into the track while still being felt.

 

Try standing back a couple feet from the mic when doing the double. That can bring some interesting room tone into the piece. Then it can be added only during a bridge or chorus and the subtlety of widening the space can be nice sans plug in... Just ideas, hope they help.

[AlamoJoe]

Double tracking was employed by The Beatles on a majority of their recordings. On pretty crude equipment. To very great effect. Listen to the wisdom imparted here, and you'll get where you want to go....