how do you record your vocals?

I am wondering how most people are doing this.
Are people using hardware compressors, or going striaght in with some sort of pre?

mic, pre, converter, plugins for compressor and reverbs and such

Probably not the best way, but I go:
Mic In on Behringer UB1204FX -> out of the main out on mixer ->Alesis NanoCompressor set to lightly compress (~2:1) to help control levels, -> Sound Card Lin In.


I guess thats my real question…is a compressor needed for voice? I know that EQ, compress, etc. will be done with plugins in n-track, but would a hardware compressor make a better sound than just running through a pre?

Thanks for all your input

in my opinion YES its better to comp befor rec.<br>you get a better signal to noise ratio and its more easy to sing. i don`t use eq at rec.

1 - Rig up a mic
2 - Carefully set levels, then back off a little bit to be safe
3 - Wail away, do lots of takes.
4 - Fix all the mistakes while making a comp track. Cheat like crazy.
5 - Be happy :)

I always record my voice dry, no fx, then add what i want afterwards, I usually add compression to the main buss, and reverb and eq to the aux busses, then i use the aux sends to detirmine how much color i want on each tack, with the vocals i tend to clone the lead vocal at least once i will then pan them just a tiny bit to ad width, usually just 2 or 3 degrees, then i take one and add it’s own compression that one i turn way down on the volume and then bring it up until it seems ti beging to support the other one, but not enough to make it over power the dry one. this way i have a balance compression on the whole mix, and a light compression on the clone vocal, and the dry vocal then can have reverb added from the aux send as per taste.

This method works well for me, others like to compress the vocal going in, I don’t like that since i can not take it back out if it is not right, since it is in the wave file.

To ech there own, just experiment and … here we go again “trust your own ears”

Condenser mic>Behringer Mixer>soundcard>Ntrack.

I use a monkey rigged spit sheild I constructed out of a coat hanger and panty hose, to cut back the P’s, T’s and general wind motion.
I usually do two takes of a vocal, making sure to sing them the same, with no FX.
I then pan one track -12 to the left or right depending on how the rest of the intruments are mixed, keeping it’s volume slightly lower than the main Vox, one digit.
The other track panned straight down the middle.
I reduce the mid range EQ on one track and increase the high end EQ.
The I raise the mid, and decrease the high end on the other.
This seems to give each track it’s own space in the spectrum.
And the panning of the second track with it moved slightly off timing (about 1/2 sec, gives an almost verby sound without taking away from the punch.
I usually don’t compress, but acasionally I will compress the final mix of a song, and this also does something to the vocal, that sounds good.
Here’s some examples of what I do if your intersested.

Soundclick site

That will give you a better idea of what I’m talkin’ about, no fancy expensive equiptment, no external compressors, mojo’s, ect, just a man and his mic.

keep trackin’


Cheat like crazy.
There you have it! That's me with a guitar in my hands. :)

Serious answer...record at 24 bits if possible, but record low enough that you NEVER hit clipping, no matter what bit dept you record at. 24 bits will give much better results than 16 bits for things like drums and anything very dynamic like acoustic guitars. 24 bits has MUCH more available dynamic range than 16 bits. You have to think about recorded level a lot more when doing 16 bits. Digital is not like tape in that you should record as close to 0 db as possible to keep the signal above the tape hiss. There is no noise down there to stay above, but being down there does have it's own problems, so you don't want to record too low.

But no, never use outboard compression unless you know it's good and that is the effect you really want, except in situations that there might be unexpected spikes in the volume, such as recording a live band. Then it's a very good thing to have some kind of limiters that don't touch the sound except on the very highest peaks (as a safety net). It's easy to add effects later. It's virtually impossible to remove them later.

I am very interested in this question because the only recording I do is of live (yeah - real people) musicians who come to my studio. The music is mostly classical but I did have a hip-hop group here recently. What these musicians want is a recording that preserves their performances. I don’t add anything to the recording (like drum tracks); whatever sounds they made when they were playing is what goes into the recording - and nothing more. I insist (and so far nobody’s challenged me) that everything is recorded at the same time. That is, I won’t let them lay down the piano track and then sing to that. Classical music doesn’t work that way because of the spontaneity that exists and must be preserved.

For vocals I use a Neumann TLM-103 and I set it behind a screen. The singer faces the pianist with the microphone about 8 inches in front of him/her. On my piano I put two Neumann KM-184’s. These are very close to the strings. Two more Neumann KM-184’s about 12 feet from the piano in an “X” configuration. About 25 feet away I place two Audio-Technica 4033a’s about 12 feet apart from each other. I record these 7 tracks with absolutely no effects added. All I will do is insure that the volume on the pre-amp keeps each track to a maximum of -9dB as the loudest signal.

Of course most of the work has to be done after the musicians go home and the last thing I want is to have to call them and tell them to come back because some setting or other was wrong. By having nothing set, I can’t make a mistake.

The mix process may take weeks (as you well know), but since I have the raw tracks all I have to do is start over if I really do something bad. My challenge is to mix these 7 tracks to a stereo sound field with the singer and piano placed in the field such that it sounds like we’re in a concert hall and the singer is standing in front of the piano. This is not always as easy as it sounds. If necessary I will compress but don’t tell anyone because I don’t want anyone to know I do it. The reason is that when classical musicians play loud they want the recording to be loud. When they play soft they want the recording to be soft. I may add some reverb to the mixed tracks. I have to insure that the piano and the singer sound like they are in the same hall and if I added reverb to the raw tracks this might not be the case. The final touch is to normalize to -.5dB. This process must be done against the entire CD, not against each individual track. If I normalized a track at a time, a very soft piece of music would be just as loud as a very loud piece of music and this would not be acceptable.

If anyone has other techniques for live recording I be grateful to hear them.

Andy, has experience taught you to set your levels as low as -9 db ?
I suppose you’re recording at 24bit resolution going that low ?

Well now that you ask, I think that I just started recording no higher than -9dB and have just stuck with it. My thinking is that when the wave files are combined that I need this much headroom to insure no clipping. If you think this is too low, I’ll try at a higher setting next time. But I do normalize when all is said and done, so the final recording is as loud as it can be. I haven’t heard any noise in the recording, so I’ve just stuck with the -9 number. Maybe I’ve been wrong all this time. Thanks for pointing this out.

I’d think you want as much as possible when recording. I always try to hit -3 or so.

I almost always use compression on vocals going in, because hardware sounds better to me than software. I don’t have fancy comrpessors either. A Bellari LA 120, a couple of dbx models, and the behringer no one wants, but which I’ve decided is really not all that bad for drums. Of course, “printing” with the effect sort of commits you, but why is that bad? Think about what we had to do with 4 and 8 track tape - you sort of get to know what will work and what won’t, and the recording that you end up with, well, it is what it is. I think that’s good.

FYI, for software compressors I’ve used the Waves stuff extensively, which is pretty state of the art, and PSP stuff, including the vintage warmer with I think is pretty glorious. But I’d still rather use a decent hardware compressor. Actually I’d rather use a really good one, but I can’t afford that sort of thing… :)

Last night I was hitting slightly over -1 db going in without clipping, but by then I was ‘vocalizing’ rather loudly :)

Andy, Tom is correct. There is no technical argument to justify recording at a lower level. The ideal is to peak just under zero – that is, as high as possible without clipping. The “headroom” argument is irrelevant; headroom for mixdown has to do with the fader levels when mixing and nothing to do with how hot the recorded tracks are.

However, at -9 dB, you’re only wasting 1 1/2 of your 24 bits, in the digital realm. More practically speaking, though, you’re wasting 9 dB of the range in the weakest chain in your analog gear, whatever that is. If you’re recording in 16 bits, those 1 1/2 bits may be significant, but not for 24 bit recording.

The kind of headroom you need to worry about is avoiding clipping during recording. So, planning to hit 9dB is probably a good safety margin. I hope you don’t worry if it hits -1dB! If the top peak hits -18 dB, though, do you consider that a good take? Even if the typical level in the track is in the -24 to -30 range? That’s starting to get dicey, asking a lot from your analog gear to get a good crisp signal when you’ve wasted 30 dB in its limited range. If your gear is superb and your gain structure is optimal, that still might give you as much as 60dB, which is fine for most mike recordings.

The original question was whether to use hardware compression when recording vocals.

Experts argue about this. There are good arguments both ways. Bottom line is it’s not necessary, but some folks like the results they get that way. Experiment and use what works for you.

That said, here are the common arguments.

Many folks feel that, anything you can put off until the digital stage, you should – because then you can always change your mind later, or readjust the settings, or whatever you want. This is a pretty strong argument.

On the other hand, without compression or limiting, you might ruin a recording of what would have been a peerless performance, due to clipping during conversion. Anti-hardware folks say they can just leave more headroom to avoid that. Pro-hardware folks counter that this would reduce the quality of most of the recording just to deal with a problem that might not happen. Anti-hardware folks reply that the gear is good enough to sacrifice a dozen dB and never miss it – and meanwhile, they don’t have to worry about hardware mangling their signals.

And then there are the folks who say that the software compression doesn’t sound as good as their XYZ hardware compressor.

But most of all, the argument is “I’ve been doing it that way forever, and it works for me.” And this is the argument that counts, of course.

If I were recording live acts I’d be sure to use limiters (not compressors) to save my butt when I failed to leave sufficient headroom. In the studio, where I’m recording things that can be done over without too much fuss, I don’t don’t use limiters. If I had clients paying big bucks for studio time and who’d resent do-overs, I might slap those limiters back on. I wouldn’t do compression outboard, because I like to dial in the right amount of compression on each track (where I use it) while listening to the whole mix, just like any other effect. I don’t give a rat’s patootie how good an individual track sounds, it’s the mix that counts.

HTH :)

These are all interesting points and I appreciate hearing everybody’s thoughts.

One that stuck out was the case where I was recording at too high a level and during the session the signal clipped on one of the tracks. Yes, I’d have to ask the performers to start again and they would certainly not be happy about this because they are paying for the studio time.

To try to avoid this situation I always ask them to play the loudest section that they are going to record so I can set the levels correctly. They will do this, but interestingly enough, when they are actually recording, in the heat of the performance, they are very likely to play much louder than they did during setup. This has happened so often that I’ve come to expect it. What makes this so I don’t know. So planning for -9dB has saved my bacon a number of times, because during the recording that -9 magically becomes -1.

There’s another (somewhat related) issue that I have to wrestle with. Always true of singers, but some wind players have the same problem which is that they are physically unable to sing (or play) a particular piece an unlimited number of times. Consider “In Questa Reggia” from Puccini’s Turandot. This is a killer aria (and has ruined more than a few soprano voices). I must not ask the poor singer for another “take” just because I’ve set the levels too high. If the singer and accompanist decide they want to do another, then we do because that’s their decision, but most often they just sing portions of the aria again and I’m required to edit these repairs into the final mix.

I guess I forgot about this aspect of editing in my first post when I said that all I do is “preserve” the music that is created by the musicians I wasn’t being entirely truthful. I have been asked to patch together sections from multiple “takes” to create a final performance. I’d rather not have to do this, but I will do it, of course.

So, would it be more critical to have 24-bit without compressiong versus 16-bit with hardware compression?

Sorry for all the tough questions.

I do agree with the -9dB = -1dB during the performance. No matter how hard the drummer bangs (or me on bass for that matter), it always clips without a few dB’s of headroom.

Random-id, from my point of view that’s a false dichotomy. 16 or 24, I still think it’s better for some kinds of music to record vocals with some compression. Opera strikes me as a different creature - limiters but not compression. A good classical recording needs to be done in a good room, and the mics are going to be some distance from the performers. That in itself tends to even out dynamics. But for close micing of pop vocals, I still say print compression.