I don't even know where to start with vocals.

[kylendm]

Alright so one of my songs is finally getting vocals on it. The guy is being recorded with a audiotechnica 4040 and the Wavs are being sent over.

Basically I don't even know where to start with raw vocals. All I know is that I'll probably have to cut somewhere in the mix to get them to fit in because the mix is pretty god damn full as is. I'm getting the files by next weekend but I just wanted some input.

Here's the song I have done.

 

[Trevoire520]

First step to any vocal for me is to smash the absolute crap out of it with a 1176 style compressor and send it to a delay. Maybe add some saturation and eq if necessary.

 

[kylendm]

How would you setup the delay?

 

[heshian46]

All I know is that I'll probably have to cut somewhere in the mix to get them to fit in because the mix is pretty god damn full as is.

Mainly in the 3k area.

But yes, flatten the vocals.

 

[Nimvi]

Agree on agressive compression; makes it sound much more solid right off the bat and makes sure that it stays at whatever volume level you set your fader during the entire song. I like doing this in 2 stages with 2 different compressors. Makes it sound more "natural" to me (as natural as squashing a voice by 20 db can be...).

I find it crucial to experiment with the high-pass point, as it can really change the way the vocal blends in the mix. Too much, and the voice sounds tiny and distant, no matter at what volume you set it. Too little and it just sounds disconnected to the mix and doesn't really glue. A bit like someone is singing karaoke over a background track, if you know what I mean. Where to set it exactly depends completely on what works for the mix, but don't be afraid to test the limits. I end up between 100 and 200 hz usually.

Delay and reverb are great, like mentioned, but again are to be set completely to what works for the mix. No presets here, sorry! It helps to listen to the stereo-image of your mix carefully, and find out where there are holes. You can pan the delays or reverb there to fill up the stereo-image nicely, and make them stand out without muddying up other instruments.

Another cool thing to do to vocals, is adding some saturation or even distortion to them, and mix that in a bit. Makes agressive vocals stand out very well, and ofcourse makes them sound even more agressive too.

Listen to the S-es in words. Do they stick out like a sore thumb? In that case, use a de-esser. Most de-essers allow you to listen the sidechain (the part that will be filtered out), so use this to your advantage. Aim for the frequencies that sounds the most annoying to you, and switch back to normal listening mode. The S-es should be a lot smoother now. Be careful though, as de-essing too agressively can make the vocalist sound like he has a lisp.

Like Heshian mentioned, vocal intelligibility lies around the 3khz mark. If you have trouble making out the vocals, chances are that other instruments are clouding up that area. Usually guitars and synths are a good place to start notching. Put up a notch filter on the guitarbus and move it around that 3khz mark slowly, while listening to the vocal. When you hear the clarity of the words improve, you have hit the mark. You may want to make the cut less agressive than a full notch though, as the guitars also need that area to remain clear. Try to find a balance where you can hear both of them well. I like to automate this cut, so the guitars/synths/whatever are only being cut whenever the vocal is active. No need to leave that hole in there when there is nothing to fill it up IMO.

Same goes for volume of surrounding instruments. You may want to automate their volume to be a bit lower whenever the vocal is active, so it has some more room in the mid-stereofield. I usually automate just the inner guitartracks about a db whenever the vocal is active.

Good luck!

 

[B36arin]

One thing that can help A LOT is automating the vocals to sit properly before they hit a compressor, making it the comp work more evenly all the time. It takes a bit of time but usually makes the vocal sit that much better. Riding the level of a vocal into the compressor is often overlooked, which is a shame.

Otherwise compress a lot and EQ to taste, EQ varies with mic choice, the vocalist etc. The most important thing in getting a great vocal sound is obviously in the performance, mic choice and preamp(have you ever heard a U87 through a Studer? With the Neumann it smokes the SSL pres we have in the studio.). It's pretty much impossible to get a badly recorded vocal track to sit well.

Play around with different delay/reverb combinations, depending on the material. Try combining verbs and delays in series and see if you find something which fits the music.

Basically, the best advice imo is loads of automation and compression. Automate before the comps to make them work equally, and automate after the comps to make sure that the vocal always sits at the right volume compared to the music. As has been mentioned it's also useful to automate the other instruments in relation to the vocal to give it some more space.

 

[heshian46]

One thing that can help A LOT is automating the vocals to sit properly before they hit a compressor, making it the comp work more evenly all the time. It takes a bit of time but usually makes the vocal sit that much better. Riding the level of a vocal into the compressor is often overlooked, which is a shame.

This sounds like a nice idea. But remember, if your going to try this, most DAW automation is post fader. So you'll need to do something like print the track after setting up the automation.

Or use this: vocal rider

 

[Skinny Viking]

One thing that can help A LOT is automating the vocals to sit properly before they hit a compressor, making it the comp work more evenly all the time. It takes a bit of time but usually makes the vocal sit that much better. Riding the level of a vocal into the compressor is often overlooked, which is a shame.

That seems kinda backwards to me but hey, if it works for you cool

 

[B36arin]

This sounds like a nice idea. But remember, if your going to try this, most DAW automation is post fader. So you'll need to do something like print the track after setting up the automation.

Or use this: vocal rider

Clip gain is a godsend for this purpose.

Skinny Viking, I'm not flattening everything out completely before hitting the comp, but spending 5-10 minutes on making the input to the compressor pretty even gives you the opportunity to push the compressor harder without it crapping out on the more powerful notes. It's the same principle as staged compression, but I prefer the results that I get with editing the volume of the track pre compression.

 

[heshian46]

Clip gain is a godsend for this purpose.

Skinny Viking, I'm not flattening everything out completely before hitting the comp, but spending 5-10 minutes on making the input to the compressor pretty even gives you the opportunity to push the compressor harder without it crapping out on the more powerful notes. It's the same principle as staged compression, but I prefer the results that I get with editing the volume of the track pre compression.

Very true. Oh, and cubase has a "clip gain" type function built in as well.

 

[Headof75]

Are you getting clean or extreme vox?

 

[Skinny Viking]

Clip gain is a godsend for this purpose.

Skinny Viking, I'm not flattening everything out completely before hitting the comp, but spending 5-10 minutes on making the input to the compressor pretty even gives you the opportunity to push the compressor harder without it crapping out on the more powerful notes. It's the same principle as staged compression, but I prefer the results that I get with editing the volume of the track pre compression.

oh I get why you do it and like I said, if it works for you thats cool, thats your workflow. It just seems like an extra step to me and a little backwards but thats based off of MY workflow

compressing the vocals properly or at least well on the way in kinda handles this step without me having to spend extra time doing rough volume automation on the vocals BEFORE hitting it with compression. I'm assuming that after you do that you then print the vocal to a new track or bounce/import and then place your comp on the new one right? Thats the part that seems like an extra step that if compressed properly when tracked you could avoid. Again, just a difference in workflow, not saying your way is wrong or mine is correct

 

[kylendm]

This is also for some extreme vox. All screams and growls but my vocalist does mostly highs. No singing at all. Still this is all great stuff. I really appreciate getting me started. Can't wait to get the vocal file.

 

[guitarguru777]

For me its straight up 1176 Compression smashing the shit out of it fast attack and slow release so it STAYS smashed, into a delay set at about 50ms jus to thicken it a bit, then high pass at 150, from there you can do whatever you like, that should get you in the ballpark.

 

[B36arin]

oh I get why you do it and like I said, if it works for you thats cool, thats your workflow. It just seems like an extra step to me and a little backwards but thats based off of MY workflow

compressing the vocals properly or at least well on the way in kinda handles this step without me having to spend extra time doing rough volume automation on the vocals BEFORE hitting it with compression. I'm assuming that after you do that you then print the vocal to a new track or bounce/import and then place your comp on the new one right? Thats the part that seems like an extra step that if compressed properly when tracked you could avoid. Again, just a difference in workflow, not saying your way is wrong or mine is correct

I've only recently started compressing on the way in, and I'm always very careful that I don't over-compress the loud parts of the vocal, I'm a bit scared of it! I work in PT10, so I just slice up the take into regions and set the volume of the different regions with the clip gain feature(which is pre insert). An easy way of doing it is putting on a compressor which works like you want it to and just back off the clip gain when the compressor starts squashing the vocal too much.

Like you said, there is no right or wrong here. I've been working with a lot of mellow music lately where the artist is almost whispering and singing loudly in the same take. I haven't managed to get a compressor to even it out enough on the way in without it sounding unnatural/overcompressed on the loud notes. If it's a straight-up growl take which has been compressed correctly on the way in there's no need to go and automate any volumes before hitting the comp, but the technique is very useful for very dynamic vocal performances. As far as I know it's a pretty common thing to do when producing dialogue for TV/commercials to make sure that the vocal is up in your face all the time.

 

[thefalloftheheretic]

For me, my extreme vocal chain is this:

Gate
EQ: shelf at 75hz, boost at 200hz, cut at 1khz, cut at 4.5khz, boost at 6khz
Compressor
De-esser
Reverb

 

[Skinny Viking]

that could be quite true as far the tv / broadcast stuff is concerned. Can't comment on it one way or the other having never done that kind of work. However, given what the OP was asking about in regards to his particular situation not sure tv / broadcast techniques are the way to go

You're absolutely correct in that you don't want to over compress on the way in ... you can always take away more dynamics later if you need to, can't add back what wasn't there Still, if all you're doing is reducing SOME of the dynamic range with the pre-comp volume automation, just to smooth it out for the compressor to react more in your favor when you do put it on there, you can always just go with light settings compressing on the way in ... maybe just shoot for like no more than 3-4 db gain reduction to start with and see if it allows you to skip that step. May save you time in the long run and help you get better results more efficiently. Just a suggestion. You can always experiment later to see how far you can go with gain reduction in a take

 

[B36arin]

that could be quite true as far the tv / broadcast stuff is concerned. Can't comment on it one way or the other having never done that kind of work. However, given what the OP was asking about in regards to his particular situation not sure tv / broadcast techniques are the way to go

You're absolutely correct in that you don't want to over compress on the way in ... you can always take away more dynamics later if you need to, can't add back what wasn't there Still, if all you're doing is reducing SOME of the dynamic range with the pre-comp volume automation, just to smooth it out for the compressor to react more in your favor when you do put it on there, you can always just go with light settings compressing on the way in ... maybe just shoot for like no more than 3-4 db gain reduction to start with and see if it allows you to skip that step. May save you time in the long run and help you get better results more efficiently. Just a suggestion. You can always experiment later to see how far you can go with gain reduction in a take

Yeah, I know. The thing is that if you aim at that low levels of compression on the stronger parts it doesn't even touch the softer passages of a dynamic performance. I was recording an RNB singer a few weeks ago with about 30-40 dB of dynamics from the soft parts to the really loud parts.

Oh well, it's just two different ways of working and we both seem pretty content with our own workflow. I personally like having full control of the levels going into the comp, since I'm very sensitive to how compressors react to powerful notes.

 

[amarshism]

Smash with distressor 20db gr.
Smash with ssl e channel comp 20db gr.
Done.

 

[ArthurD]

For me, my extreme vocal chain is this:

Gate
EQ: shelf at 75hz, boost at 200hz, cut at 1khz, cut at 4.5khz, boost at 6khz
Compressor
De-esser
Reverb

Why cut at 1khz? I found some clarity of vocals there.