PreSonus Blog

Friday Tip of the Week—How to Gain Better Vocals

Vocals are the most direct form of communication with your audience—so of course, you want your vocal to be a kind of tractor beam that draws people in. Many engineers give a more intimate feel to vocals by using dynamics control, like limiting or compression. While that has its uses, the downside is often tell-tale artifacts that sound unnatural.

The following technique of phrase-by-phrase gain edits can provide much of the intimacy and presence associated with compression—but with a natural, artifact-free sound. Furthermore, if you do want to compress the vocal further, you won’t need to use very much dynamics control because the phrase-by-phrase gain edits will have done the majority of the work the compressor would have needed to do.

The top track shows the original vocal. In the second track, I used the split tool to isolate sections of the vocal with varying levels (snap to grid needs to be off for this). The next step was clicking on the volume box in the center of the envelope, and dragging up to increase the level on the sections with lower levels. Although you can make a rough adjustment visually, it’s crucial to listen to the edited event in context with what comes before and after to make sure there aren’t any continuity issues—sometimes soft parts are supposed to be soft.

The third track shows the finished vocal after bouncing all the bits back together. Compared to the top track, it’s clear that the vocal levels are much more consistent.

 

There are a few more tricks involved in using this technique. For example, suppose there’s a fairly loud inhale before a word. A compressor would bring up the inhale, but by splitting and changing gain, you can split just after the inhale and bring up the word or phrase without bringing up the inhale. Also, I found that it was often possible to raise the level on one side of a split but not on the other, and not hear a click from the level change. Whether this was because of being careful to split on zero crossings, dumb luck, or Studio One having some special automatic crossfading mojo, I don’t know…but it just works and if it doesn’t, you can always add crossfades.

That’s all there is to it. If you want to hear this technique in action, here’s a link to a song on my YouTube channel that uses this vocal normalization technique.

 

  • Laurynas Gaidys

    I believe that there is some kind of licensing issue of that particular function.

  • ScooterMac

    Yes! I totally agree with you on that. That’s one of the “big things” that I missed when moving from ProTools 3 yrs ago. I thought we would have that in S1 by now. 🙁

  • Gavin Steiner

    I use the Mixtool to automate the gain changes and then print them. This avoids having to split the clips, makes for smoother transitions and no chance of popping. 🙂

  • Milton

    I hope Studio One implements a “Split at zero crossings” feature like Cubase Pro already does in an upcoming version. Also, a “Stationary Cursor” like in Cubase would be useful as well.

  • Yes, StudioOne does crossfade. (sometimes automatically when splitting or moving, I think? newish feature maybe) BUT, to be certain, select each section of audio on both sides of the split, and press “X” key = CROSSFADE. zoom in and adjust as necessary.

  • Anderton

    Yes, as I said “Although you can make a rough adjustment visually, it’s crucial to listen to the edited event in context with what comes before and after to make sure there aren’t any continuity issues—sometimes soft parts are supposed to be soft.”

  • Jo Trebi (Josh)

    That’s a great confirmation right there. I have been using this technique. but I always have the feeling it is not the right thing to do, Thanks soo much. Appreciated.

  • Laurynas Gaidys

    It should never be used automatically without listening and feel of a song.

  • Malika Powell

    Thanks for the advice I feel like I’m one step closer to being a better producer.

  • Anderton

    The first time I wrote about using phrase-by-normalization, scores of “professionals” said I was completely wrong – that doing normalization was a dumb idea, and it should never be used this way. I’m glad to see times have changed since then to where someone can consider it “nothing new.”

  • Jeff Evans

    Thanks Craig. Another great tip from you. Here is another idea. The Waves Vocal Rider is an excellent plugin and right now it is on special for a fraction of its usual price. It does a brilliant job of further riding the gain of a vocal track. Even better after some prep work such as your editing suggestion. Once Rider is tweaked nice it just levels things out to another refined level. Can be used over a track or in side-chain mode where it takes its cue from the music signal.

    Also note too that clip gain levels/adjustments in Studio One can be input from the QWERTY keyboard numeric keypad too.

  • Laurynas Gaidys

    Nothing new. It is a number one tip for a beginner or pro editor. However, this program should have a clip gain line function like Pro Tools because sometimes clip gain function in S1 is very inconvenient.

  • Anderton

    Thanks for chiming in, Jeff! Good comments as always 🙂

    Just remember that the visual changes in the waveform *approximate* the result of the gain change, which is why you can have waveforms that appear not to hit the headroom limits but when you bounce, there’s an error message that clipping occurred…all the more reason to depend on your ears. The way Sonar gets around this is that you can normalize to any value, but the tradeoff is you get zero visual feedback until you bounce as opposed to Studio One providing visual feedback during the process itself.

    What I don’t quite understand with either program is how I can cut in the middle of a continuous sound, change levels on one side but not the other, and not hear a click. I chalk it up to some kind of software engineer mojo mere mortals cannot understand 🙂

  • Jeff Evans

    I have been doing this for years. It is a much nicer way to get a very natural vocal sound as opposed to slamming a compressor over the whole track. This is where also seeing the waveform changes as you change individual event gains with their event gain handle is very useful (unlike Sonar!)

    You must listen as Craig suggests though because sometimes bringing up a word or phrase to match the rms parts of adjacent words/phrases can sound unnatural. It is easy to hear though so just pull them down 3 dB at a time until they sound natural again.

    Sometimes with a large inhale before a word or phrase you can separate it and actually take it down by 6-10 db or so. Then when you do put a compressor over the whole thing performing light duties the inhale will often end up just being audible and at a nice level again. There is often silence either side of a loud inhale too making it easy to split and modify.