How to tame multi Vocal tracks, VOl, EQ

pure.fusion

New member
Hi all,

I've got a bunch of raw vocal tracks, with quite widely varying volumes etc. Yes, It would have been better to have it done properly with real singers and real engineers but that's not what happened!

It's a case where very high and very low vocal parts are soft and are less constant in volume. And of course, the same vocal part in the next phrase is a different volume than the previous.

So I'm about to launch into sorting this mess in to sets of harmonys, main vocals. I'll have to change volumes in the harmony parts to balance that mix, then balance this with main vocal.

I posted previously here and I think the advice was to ride the fader (set up fader automation) to balance all of this. Should I do this before or after I set up EQ?

Any suggestions here? (I can tell this is going to get messy from the get-go .. and I need all the help I can get)

FM
 
What DAW are you using? In Reaper, if my vocal tracks are all over the place (usually only happens if I have tracked them in different sessions), I will normalize them first, so they are all at roughly the same volume. Then automate the volumes on them separately,,if needed. Apply EQ separately, if needed. Then I'll put all the BU vocals in a folder, balance the volumes with the sliders in that group, then automate volume on the group to synch in with the lead vocals. I won't apply compression to the individual BU vocal tracks, but will add it to the group folder.
 
No offence intended, but personally I would never normalize anything! Ever. It will only normalize all the noise, breath spits, esse and vocal all to the same level and sound bad. But each to their own.

You could cut up all the vocal tracks into bits and automate the volumes of each bit in turn to get an average level throughout and then use a compressor to tame the highs and lows. That's what I do when I have vocals that go up and down in volume. It takes a bit of time but you'll get much better results that using a normalizer.
 
no normalize.


group all the vocals into 'loud', 'soft', and 'lead'.

send them to separate stereo 'loud', 'soft', and 'lead' sub busses.

mix the pans from the individual tracks, but mix the levels from the busses.

put compressors across the busses to help further dial it in.
 
Not to jump on the No Normalize band wagon, but even if I did think there was any use for normalizing, I don't see how normalizing a bunch of tracks accomplishes anything in this case. They're all still going to be proportionally as far apart from each other. If the highest peak comes in at -10db (for example), and you normalize the tracks to hit 0db, all you're doing is raising all the tracks by 10db. They're still going to be as far apart from each other as they were pre-normailizing, they're just now all 10db louder.

If you meant that you normalize every track separately, then disregard most of what I said above.
 
Last edited:
I'm pretty sure (yes, just checked) I didn't say normalize them together! Normalizing them separately gets them to roughly the same level so when mixing, you are not cranking one up to +10 on your automation curve while keeping others at -5 (or whatever level). yes, noises gets boosted too, but that's what you automate for - they are going to get boosted when you crank up the track volume regardless.

As to using a compressor to equalize volumes as Mr Clean suggests - there's been a long thread here recently on this very subject, I don't need to comment more. Not sure why you would need to cut up the tracks into multiple pieces, either, that's what automation is for.

Also not sure what Gonzo was suggesting with the 3 buss idea, except I guess if you have a TON of tracks to put together, it saves doing each one separately.
 
I cut tracks up because it is easier for my workflow. Especially in Reaper. If the first word of a line is too high, cut it after and reduce the volume of that slice. Done in seconds. Quicker than clicking Envelopes, ticking boxes, adjusting the timeline of automation. It's how I work. It's still automation but a different way about it.

However, I didn't say use a compressor to equalize volumes, I said use a compressor to tame the highs and lows. This would be after doing the above and there'd be very little to tame but it would help to make the track sound like a unit. Subtle compression. Not squash and suck the life out it. That would sound equally as bad as normalizing it.

But as for normalizing, I learned the hard way what normalizing can destroy and wouldn't use it again. It's too hit and miss and I wouldn't want any additional noise added and more automation work.
 
normalizing doesnt really destroy anything and could be useful in this scenario... if you recorded in a poor room with a lot of floor noise then maybe... but even still.... its a decent tool when used correctly... why else would it still exist?
 
Perhaps Normalize tools have changed in the last 10 years and I'm out of touch and they work magic now. I still wouldn't use it.
 
Perhaps Normalize tools have changed in the last 10 years and I'm out of touch and they work magic now. I still wouldn't use it.

lol mr. snippy... for someone that doesnt know how to regulate vocal volume you sure know a lot about how peoples advice WONT work...

Good luck my friend.
 
Where exactly have I said that someone's advice won't work? Get back in your box with your normalize button.
 
... ok. So lets assume that normalize is not what I'm going to do. Actually thinking about it, that still wouldn't get me to where I need to be because the 4 and 5 part harmonys will need different volumes on different parts any way.

I'll come clean. This project was sung in different stages, on different days, examined and the part that were too rough gone over again. So (for example) one out of the four parts of a harmony may have a quiet first line, and loud second line and a medium third line.

It's actually not as bad you *could* imagine but the above issue does exist and If I have to address the problem in one bar or the whole song, I still need a method.

Currently I have 10 vocal tracks. I could probably do it with 5 but this would take more automation and mucking around and PC resources isn't my limitation so....

Mr Clean, "You could cut up all the vocal tracks into bits and automate the volumes of each bit in turn to get an average level throughout". I think this was my go to method, except I wasn't going to cut anything, just Volume automate (Cubase) each part until it sounds balanced. Loooooong process though.

Gonzo, if I group into Soft and loud, I'll still have to mix each line of vocals so the harmony sounds balanced - so I may as well not group because a soft part might actually be needing to stay soft for that bit? I dunno.

Mr Clean "I cut tracks up because it is easier for my workflow. Especially in Reaper. If the first word of a line is too high, cut it after and reduce the volume of that slice.". So, you edit the WAV file and make a permanent dB adjustment? I don't use Reaper.

Whatever I end up doing, I'm (obviously) tryinbg to make it so that I don't have to go back and adjust balances again.

...Oh, and reading back my original post, I don't know about hte EQ. If I balance all these volumes and then apply EQ, will I have to balance again?

So options are

EQ individual tracks, then balance volumes, then group for mixing.
Balance volumes (to get rough mix), EQ them, then balance again?

Cheers.
FM
 
Not sure what normalize would 'destroy', but the great thing in Reaper is that everything can be reversed. 'non-destructive editing' is what they call it. :D
 
Where exactly have I said that someone's advice won't work? Get back in your box with your normalize button.

i actually never normalize because i never need to, but "get back in your box" ??? you can do better than that mr clean.

anyways.... you have 10 vocal tracks.... a capella? is this a choir type setting? full music?

basically each vocal range should be "one" track... make each track of a particular vocal range and adjust the individual volume of each take to make it sound like 1 consistent take... if it doesnt sound like 1 consistent take then .... good luck fixing it without retracking.... if its a question of dynamics, that what compression is for.

the 2 tools you want to even them out is automation and compression.

Normalizing could work because it takes the highest peak of a track and brings it to the hottest level it can get to without clipping, so if the track is broken up correctly, then normalizing can achieve what your doing.

Normalizing got a bad name from 90's music... everything was normalized and over compressed and limited and raised the floor noise and blagh.... but thats not to say normalizing is always bad, as i stated before it can be a useful tool if used correctly... however its rarely reached for in my case. and as stated before nothing is permanent or irreversible.

is it possible for you to bounce the part so we can hear it?

if you want you can send me the tracks and i can run them through melodyne... which allows you to adjust the volume of each transient individually (along with correcting pitch if necessary or desired)

this is definitely something that you can fix... rather easily
 
Mr Clean, "You could cut up all the vocal tracks into bits and automate the volumes of each bit in turn to get an average level throughout". I think this was my go to method, except I wasn't going to cut anything, just Volume automate (Cubase) each part until it sounds balanced. Loooooong process though.

I don't know about Cubase but in reaper when you split a file, you can adjust the volume of the split clip with the mouse pointer at the top of it, click and pull down. It shows you the volume change on the wave as you do it. Does Cubase have a similar thing? Item processing? Trim for clips? I've not used Cubase in years so I have no idea.

Mr Clean "I cut tracks up because it is easier for my workflow. Especially in Reaper. If the first word of a line is too high, cut it after and reduce the volume of that slice.". So, you edit the WAV file and make a permanent dB adjustment? I don't use Reaper.

It's not a permanent adjustment in Reaper as you can just turn the volume back up or down further on the clip. As in pic.

example.JPG

If you were to do something like this with lines of a song, rather than single words, it could be quite quick. Single words however would take an awful long time.

Personally, I would get the levels sorted before EQ.
 
i actually never normalize because i never need to, but "get back in your box" ??? you can do better than that mr clean.

I could of indeed done better than that but for the sake of the forum and being personal over the internet being absolutely pointless, I left it at that :D
 
anyways.... you have 10 vocal tracks.... a capella? is this a choir type setting? full music?

It's everywhere. Think something like Bohemian. Multi vocals, panning, overlapping, some split stereo, some centred, some with backing, some no backing. You name it....

basically each vocal range should be "one" track... make each track of a particular vocal range and adjust the individual volume of each take to make it sound like 1 consistent take

Yeah, I'd like to have one track per part, but that becomes a but harder to manage when that one part serves different functions (as mentioned above). It was easier (at this stage anyway) to separate off the same part into another track.

... if it doesnt sound like 1 consistent take then .... good luck fixing it without retracking.... if its a question of dynamics, that what compression is for.

Ha! Retracking is my middle name. I'm a home recorder, for no monatary gains. My motto is that there's nothing I can't record, if you have unlimited time, and that best 1 take out of 58 is the perfect one to use. This thread is discussing one of the down side of this method. The info you should take from this point is that *withought doubt* my standards of output will be less than the standards of those studio-engineer-types offering advice in this forum. ... remembering of course that I'd like to think that my home recording standard is "ok" and not utterly crap. Samples in my blurb if you're keen.

the 2 tools you want to even them out is automation and compression.
Got it. That much I got!

Normalizing could work because it takes the highest peak of a track and brings it to the hottest level it can get to without clipping, so if the track is broken up correctly, then normalizing can achieve what your doing.

Normalizing got a bad name from 90's music... everything was normalized and over compressed and limited and raised the floor noise and blagh.... but thats not to say normalizing is always bad, as i stated before it can be a useful tool if used correctly... however its rarely reached for in my case. and as stated before nothing is permanent or irreversible.

is it possible for you to bounce the part so we can hear it?

if you want you can send me the tracks and i can run them through melodyne... which allows you to adjust the volume of each transient individually (along with correcting pitch if necessary or desired)

this is definitely something that you can fix... rather easily

Yeah, I don't see normalizing as a great go-to for me. After normalizing, I'd still have to balance multi part vocals. I have to do this with or without Normalizing, and after the heate discussions I'll assume that it's something you only go to ifd you have to, ir if it really solves your problems.

I can do the adjust of volumes etc. I have mixed down multiparts before with success (depending on your standard of quality of course).
.. and yes I could bounce a sample for you, and I'd be happy to but as you see from above, it'd be a big job to get a bounced, "rough-in" track to post.

I'd be doing ok on my own, thank you. The method of a few years ago would be to start at one and and finish at the other and spend an enoumous amount of time and pain getting the finaol outcome right, usually at the expense of audio quality and final result.

This post is all about best method to minimise the pain and maximise results, since there are many more parts than usual to deal with.

Oh.... and the EQ question, and when I apply it.

Cheers,
FM
 
Back
Top