Making all words and vocal parts equally loud

[Leander]

Hi!

How can I get my vocals with all words equally loud..about the same level without having to spend hours on manual automation?

I have Vocal Rider by Waves, but it never works...and I don't know how to use compressors either...I have a couple of plugins, but they either don't do anything or not what I want.

Do you know how I can level my vocals in an easy and fast way, e.g. "reference is -6db, everything below should go x db up, everything too loud should go x db down"?

[Base 57]

Melodyne has a volume leveling macro that works pretty well. 

Clip Gain envelopes are the most precise way and shouldn't take all that long with a little practice.

[Leander]

Thanks! Is there a way to use clip gains automatically without having to draw the line myself?

[Glenn Stanton]

perhaps learn the tools? learn about compressors, downwards and upwards "compression", learn the vocal rider - it's a subtle effect. Melodyne is also very nice (but more subtle than the compression approach). it might be several FX to get a solid leveling of the vocals - but be aware of lost dynamics in doing so.

another option, use an EQ on the instruments buss which is sidechained to the vocal(s) to trim 1-2db off the frequencies used by the voice (or could be a solo instrument as well) to let it fit better.

and definitely learn to use the clip gain envelopes - this is a skill which would be useful in any modern DAW. as nas Base57 noted, won't take long to grasp and do.

for me - i mainly start with a) good vocals; b) melodyne and RX cleanup with some leveling applied; c) clip gain adj where needed; d) moderate levels of compression in a couple of stages; e) the EQ carve out in the final  submix stage on instruments. keeps the vocals mainly natural sounding while keeping a good level throughout.

 

[mettelus]

17 hours ago, Leander said:

How can I get my vocals with all words equally loud..about the same level without having to spend hours on manual automation?

A couple quick questions for our clarification. Are these multiple clips from multiple sessions? What is the underlying cause for the variation in your assessment?

To add to the above, one thing I harp on with folks is repeatability in the production phase. Identical setup and positioning for every session, but even then there can be variability between sessions (although as minimal as possible).

In addition to the above options (just another option to try), when I am doing such "for time," using clip gain (CTRL - drag up/down) on the clip will give a visual redraw of the entire clip's waveform (may need to split clips to use effectively), but only the clip being actively hovered over will be affected... one at a time. From there, a compressor (or multiple passes of a compressor, you want to keep them subtle <2:1 per pass) at the track level will be more effective since the gain insertion/removal process will make the compressor threshold much more consistent for each clip. If things are noticeably different, the clip gain mentioned above is preferred (for me) because of the automatic waveform redraw feature and can be done quite quickly visually once you get the hang of it.

[Leander]

Thanks. It is different sessions, but also a variety of articulations. I have a metal project that uses growls, screams, theatrical passages, whispering...and when recording everything alone at home, it is impossible to have even volumes for all words and all articulations. 

[Mark Morgon-Shaw]

In order :

Performance / Mic Technique

Compression / Limiting on the way in

Comping / Editing

More Compression

More Limiting 

Automation

Best way is to dive in and get your hands dirty

[mettelus]

2 hours ago, Leander said:

Thanks. It is different sessions, but also a variety of articulations. I have a metal project that uses growls, screams, theatrical passages, whispering...and when recording everything alone at home, it is impossible to have even volumes for all words and all articulations. 

Roger that, and thanks for the feedback. Chopping up clips into sections/phases that have similar volumes make using the clip gain (CTRL-drag up/down) easier, but when you mention screams/whispers, those very much should have different levels IMO (see second item below)....

A couple quick items here... first is you mentioned compressors in the OP, so setting aside time to learn/understand compressors will pay for itself 1000-fold in years to come. They are basically ways to tame dynamics in a more automated fashion than manually adjusting faders (aka, "riding the fader"). Understanding their function, when to use them, how to use them, and when NOT to use them is important knowledge to have in your tool kit.

Second is the dynamics of your song. When you mention "whispers," it is common practice to have repetitive elements lower in volume (sometimes as much as 3-6dB), because the listener has already heard them and can still "hear" them at a lower volume, but that also gives "space" for new elements to be introduced and have focus. A massive dynamic like a break in a song (or outright "ducking") puts immediate focus on what is left... in the case of "whispering" you want the listener to know it is a whisper, but also gain their attention so they listen to it. Essentially, repetition can get boring (always keep this in mind), so movement/changes in a piece can help keep the listener engaged... the advice here is more not to focus 100% making the the vocals "the same," but also realize that you can also make the song conform to the vocal track. The adage "if you want to get someone's attention, just whisper" is true... they shut up and focus on listening, so definitely use that to your advantage.

Edited Friday at 09:35 PM by mettelus

[Mark Morgon-Shaw]

13 minutes ago, mettelus said:

The adage "if you want to get someone's attention, just whisper" is true... they shut up and focus on listening, so definitely use that to your advantage.

Nice use of it in the song Parasite Eve by Bring Me The Horizon

[Leander]

Thanks! I guess there is no simple solution to have a plugin adjust the vocals automatically.

No, I have wasted many years trying to understand compressors etc. and learning the dozens of expensive equipment I have. I am someone who writes songs and lyrics and does not understand technical stuff. I've followed YouTube tutorials, some of them doing the opposite of each other with no result and always afraid to "destroy" something in the recorded files. It's been giving me 20 years of nightmares with "I want to achieve something vs. I don't understand it and get depressed".

I'll stick to manual automation then. Takes hours per song, but better than no result at all.

Examples of vocals:

 
 

[Base 57]

4 hours ago, Leander said:

growls, screams, theatrical passages, whispering...and when recording everything alone at home, it is impossible to have even volumes

I would put these disparate parts on separate tracks routed to a Group Bus. Then do lots of ear candy things. Pan, EQ, FX, DISTORTION and Volume are among many things that could be treated differently for each section without using a lot of automation.

Edited Friday at 11:43 PM by Base 57

[mettelus]

6 hours ago, Leander said:

I've followed YouTube tutorials, some of them doing the opposite of each other with no result and always afraid to "destroy" something in the recorded files. It's been giving me 20 years of nightmares with "I want to achieve something vs. I don't understand it and get depressed".

I can definitely understand the minefield that YouTube can become; all tutorials are certainly not created equal. The number of ways you can destructively edit a recording are fairly limited inside a DAW (but using external wave editors will unless you save as a new file). More often (but not 100%) inside a DAW a new file is created, leaving the original file untouched (and also unused by the project). Regardless, the truest "safety net" here is to make a backup project folder with the original audio so you always have that to fallback on for major oopsies.

Another idea that came to mind is have you ever shared a project with someone so they can modify it and send it back to you so you can see what they did (and why)?

[Mark Morgon-Shaw]

9 hours ago, Leander said:

I'll stick to manual automation then. Takes hours per song, but better than no result at all.  
 

It's also very genre dependant. In a modern pop song going in word by word or even sometimes down to a syllable can hep keep the vocal front and centre but you wouldn't really need to do the same in a folk song where you can often just work phrase by phrase.

Vocal automation is also a good chance to unearth emotion in a performance and bring it to the fore. Little breaths, sighs , fading up the ends or longer words where you can sometimes get little vocal cracks. These are gold in certain genres when it's more about emotion than technical prowess.  

Edited Saturday at 08:35 AM by Mark Morgon-Shaw

[Glenn Stanton]

Billie Eilish - her brother (& producer) did an interview showing how in one of her hits, they had almost 200 edits, some at the syllable level, to comp the lead track from some 80 takes. that was before applying the usual FX and automation...

scarey thing is that this isn't the most fanatical someone has done for vocal edits... LOL

[Amberwolf]

I do this sort of thing (compositing several sounds into one) with my music not just for vocals, but for many of the parts.  I am usually not recording my own audio these days, but instead playing with sample packs, and since they don't have the exact thing I want I build what I want from what's there, often enough by using parts of one sample from one pack, and parts of samples from other packs, to build each little part, and various automation and effects to blend them all together.  I'm no expert at it, but it isn't terribly tough--just time-consuming.

The most recent example is Drywater, Mars; another more drastic example is the "elven rock" section of Behind You Lie Many Unseen where it's not just the vocals comped togehter it's also all the other parts in there.   I even do it to the percussion in some cases, where the sound I want doesn't exist in anything I have, so I take two or more sounds to create the hit I want, and blend them together, etc..    

I'm still working on creating "lead" vocal tracks with words out of existing phrases that weren't intended for what I'm doing with them; this is harder than regular comping....    But it's an interesting process. 

 

[Glenn Stanton]

i did that a lot of comping disparate vocals as well, then i discovered Synth V and feeding that into AI voice renders (just got the SoundID Voice AI on sale to keep my processes now local to my PC). problem (mostly) solved 🙂 

Edited 34 minutes ago by Glenn Stanton
replaced the name of the product

[Amberwolf]

2 hours ago, Glenn Stanton said:

Sound Reference Vocal AI

I may be missing it but I couldn't find a product with that name... The closest stuff I found was on this site in their sale but there are separate products for each half of the name....

https://www.sonarworks.com/autumn-sale

 

Which one(s) did you get?

 

[Glenn Stanton]

sorry - SoundID Voice AI

https://www.sonarworks.com/soundid-produce/voiceai

my guess is the sale is over 😞 mine came with like 30 voices and a dozen instruments. 

Edited 40 minutes ago by Glenn Stanton