Best way to even out volume dynamics on a vocal track?

JRGariano

I use Maschine+ exclusively in standalone mode. I have a clip that's about 4 minutes in length containing the recording of a singer singing the lyrics/melody of a song. In some parts of the clip, the singer is singing closer to the mic and is louder, and other parts of the clip the singer is singing away from the mic and is quieter. It's never so extreme that there is clipping or no sound, but there's enough variation to sound unprofessional in it's current state. I've already applied compression to the clip to even out the peaks and valleys, but it's not enough, and I have a hunch that compression can only get me so far in addressing this. I'm assuming I need to go into the particularly quiet phrases of the recording and increase the gain/volume.

What's the most efficient way to level out the volume of minutes long vocal recording (loaded into a sound via the audio plugin) that's already had compression applied to it? Should I go into the Sample Editor and somehow increase the gain in the quiet parts of the waveform? Or would it be more efficient to apply volume automation to the clip somehow? Or something else?

Any tips and tricks on how to do this in the most efficient and effective way would be much appreciated.

DeepThumb

Tame your singer, tell him/it/her how to perform professionally (don't vary the distance unless intended) and redo your recording.

Fixing garbage in post isn't a good idea, sorry!

JRGariano

Unfortunately, in this case, redoing the recording is not an option.

Kubrak

It depends... How important it is, how often it happens, how much time you may spend on it and so on....

And also there may be beside changes in volume also changes in frequencies.... SplitEQ by Eventide might be good tool to handle the frequency changes.

https://www.youtube.com/watch?v=Jo7eGeuclck

It it is important and you have time, probably best would be to do it by hand. Automate the gain and if needed also adress frequency changes.

tetsuneko

Volume automation is the time-tested method for such issues.

Don't use compression on the vocal track until you have written the volume automation in fully. Once the volume has been automated, export the vocal track as a new samole and reimport to the session for adding compression.

In order to keep the noise floor from creeping up, only lower the track volume when automating. Its better to turn the louder passages down, than to turn the more quiet passages up. You might still need to use a gate or an expander before applying compression to clean things up a bit.

ozon

@tetsuneko wrote:

In order to keep the noise floor from creeping up, only lower the track volume when automating.

Where’s the logic in that? You’ll have to turn up the level of the whole vocals in the end, which will raise noise floor. It will happen now or later, either way.

However, if you first lower the parts with higher level, you’ll lose dynamic range in those parts and therefore could introduce additional noise or digital artifacts.

I do agree on the recommendation to use a gate, though.

ozon

@JRGariano maybe this excellent video by Jef Gibbons might give some hints:

https://youtu.be/O-bSfWv3Vhk

D-One

There's no "best way" or everyone would just do it that particular way, It depends, a propper vocal chain is sort of it's own science.

It's pretty normal to have a lot of volume variation in vocal performances. Obviously, the vocalist moving too much is not ideal but for the takes to be unusable they really need to move a whole lot, many vocalists move around to some degree.

It's hard to tell why the volume variation is still too high after compression without knowing what compression settings you used and what type, "it's compressed" doesn't say much, if anything. How much gain reduction are you seeing? What Ratio? Whats db difference in loud VS quiet parts? etc..

Tons of mixing engineers use more than one compressor or combine traditional compression + upwards compression/expanders. "Fader ridding" automation is very, very common in professional mixes, so much so that there's plugins to emulate it and do it semi-automatically. Manually separating particular pieces of audio and changing volume is also common (some folks call this gain staging but that could also mean other things)... this last approach is very impractical in Maschine - I would definitively not go about normalizing particular parts in the Sampler, that's a recipe for a crapy vocal.

TBH i would just open the vocal track in a DAW, even if a free one and work on it with most techniques above depending on each part then export it back to Maschine, use the right tool for the job and save yourself a headache.