The Song Fight Bulletin Board

I'm learning about Reaper and also about VST's, and loving it, but I can't seem to find the VST to fix my vocal track. Sometimes it's a little too loud and sometimes it's not loud enough.

In the past I'd either run the mic through a compressor/sustain pedal during the recording, or destructively alter an uneven track in Goldwave using the "Ampify Lite" function. That's still an option, but In the spirit of trying new stuff, I'm looking for a VST-based approach.

So far the things I've found are more about controlling stray transients than correcting an erratic recording level. There's a blinding number of VST's to look through, and I can' get any of the likely ones to do what I want. What VSTs and settings do people use for this?

first off, why is your vocal track so erratic? are you an erratic singer or did you splice together disparate takes?
-bill

You could click the track(s) that you want to even out and click "normalize" if reaper has it.

As far as a VST for full dynamics, I use the one that came with Cubase.

Wouldn't you just want to use a VST compressor? That should boost the quiet parts and cut the loud parts. Of course, if you've got too much variation in volume during one track, a compressor isn't going to completely fix that problem. Ideally you would control your volume level while recording by controlling the volume of the noise coming out of your mouth and/or controlling your distance from the mic (farther away for loud parts, closer for quieter parts, though don't get <6" from the mic or you'll get an unwanted low frequency boost due to the proximity effect, not to mention loud plosives if you don't have a pop filter or wind screen).

You may want to cut the quiet parts of the current track into their own pieces and drag them all onto another track. That way you can apply different gain, EQ, compression, reverb, etc. settings to them. You may need to do that then boost the overall gain of the track with the quiet parts -and- compress them, then you may need to cut the overall gain on the track with the loud parts, and compress those too. It sounds like you're relatively new to this, so I'll also say that you might want to apply a hi-pass EQ to the vocal tracks to cut out the really low frequencies, like <180-200Hz or something.

I use a bit of compression on vox, but going overboard never sounds good ... and I have yet to find a cheap/free vst that magically edits my vocal tracks to the proper levels. So I change the volume of low/high points manually. Here's how I do it (pretend I actually have an audio track loaded - this isn't being typed from my recording computer).







Does that help? It doesn't take too long (unless you're really picky).

Reist wrote:Here's how I do it

Or just select a track and press 'v'.

I use both volume envelopes and compression, though I find that I need less compression lately.

Thanks for all the suggestions. I did get the ReaComp plugin to do it -- I just had to get more extreme with the controls than I expected. But I'll definitely look into "normalize" next time -- that option does exist in the Reaper menus. I'll also try 'v' and see what happens. I don't think I'll use the envelope method for this purpose though.

HeuristicsInc wrote:first off, why is your vocal track so erratic? are you an erratic singer or did you splice together disparate takes?
-bill

I'm an erratic singer. You got a problem with that?

Seriously, this may have to do with mic-ing techniques (I have no idea how to spell "mic'ing"). I use a good-qualty dynamic mic made for stage use (an ATM41HE, if anyone wants to know), and I'm usually two or three inches from it when singing. At that range, an inch makes a big difference. Most performers without operatic training move in and out to compensate for their singing level, but I guess I'm not so good at that.

If anyone wants to comment on how I record vocals, I'm here to learn.

2-3 inches sounds too close. Do you have room for a pop filter in between you the mic that way? You're probably going to have bigger issues with 'p' and 'b' sounds (aka plosives) if you're that close. Also you're going to get an unwanted bass boost from being so close.

Also normalizing doesn't seem to be something that would be useful to you in this case. Normalize just raises the gain of the entire track equally (the same boost for the loud and quiet parts) until the loudest moment of the track is boosted to a certain level (usually 0db). It won't actually affect the variation in volume during the track at all. That's what compression is for.

You can use a volume automation for the track as suggested but that's a fair amount of manual work and it's only going to partially address the problem. Compression also helps your vocals sit in your mix more easily, because they take up less dynamic range. You'll usually also want to cut the really low frequencies so that the vocals aren't competing with the bass guitar or kick drum, etc.

Thanks. I had to retrack the vocals to add another verse for Robot Ninja Zombie Bear (I figured I'd never duplicate the original timbre). This time I stayed about six inches from the mic, moving in for the bassier part. It improved the tonal quality, but I still had to even off the level with a bit of compression. I guess I'm just an erratic singer.

I have a foam-cover pop filter on this stage mic. I try to control my plosives when I sing, but sometimes one gets through and then the foam screen helps.

I think the best use of envelopes is for when there are compositional reasons for the volume of a track to change. Like, for instance, to move an instrument into the background when it keeps playing after it's "solo" is over. Picture something like a bluegrass band recording with only one mic for all of them. After the banjo solo and before the fiddle solo begins, the banjo player steps back to make room for the fiddler to step forward. That's how I try to use envelopes, when I use them at all.

Btw, you know Reaper can record your fader movements and save them as envelopes, right? Set the "track automation mode" to "write", as shown in the screenshot. Play the song while "riding" the fader, like you were doing an mixdown in real time on analog tape or something. Change it back to "read" or "trim/read", and play the song again, it will remember all the volume changes.

(And yes, btw, you can do this with effects parameters too. Like tweaking the cutoff frequency of a lowpass filter to simulate a wah-wah pedal)

I knew you could do that in theory, but I didn't now the exact steps. Thanks.

I agree with you about the use of volume envelopes. I've used them to bring up the intensity of one or more instruments gradually throughout a song, or to take one instrument down a bit when another comes in.

I used to use them in Multiquence as a kind of dubbing mechanism, to kill a short section of track in favour of a fix on another track. In Reaper there are more sophisticated ways of dealing with dubs. One thing I really appreciate is the way you can collect multiple takes by looping while recording, chop the takes into sections, pick the sections you like from the multiple take lanes, and then merge them into a final track -- slice-and-dice, they call it.

If you do decide to try a compressor again, set your attack and release times as high as they will go, and turn on whatever limiter and auto-gain functions you have. You'll still want to use volume envelopes to keep your overall dynamic range intact, though (your song DOES have dynamic range, right? Gotta keep those ears interested!).