Raw Vs Processed Screams

Screams can sound drastically different in person, recorded straight from the mic, and after effects processing. Often when you first start recording screaming vocals, you'll be pretty shocked at how dull they sound compared when you practiced. Its important to realize that the scream sound you've come to expect from your favorite bands was at some point in the exact same situation as your sound - it needed some studio magic to bring them to life.

Screams typically sound good in a live scenario because the vocalist is using a dynamic microphone with their mouth right up against it. Typically dynamic microphones have a reduced frequency range, and some mics can give the effect of natural compression. When the vocalist puts the microphone right up against their face, they're further helping to compress the signal by making all of those small 'passing-sounds' right up in your face. These factors are really what influence the need to add dynamic and frequency processing to your screams when recording in the studio - its just how people like hearing screams!

Check out my video on this topic below for some comparisons:

If you want to download the raw, speech, and processed tracks I used for these sounds and the video, click here. Note that the processed track only has the screams, and the raw and talking tracks also have the talking for the video.

In the images below i'm going to show you the signal chain for each of these 3 comparisons. The first one is somewhat of a 'typical' metal screaming vocal chain. I'm using Slate Digital's VMR plugin to apply some harmonic treatment, 3-6dB of compression on a 4:1 ratio, tonal EQ (mostly adding air and removing rumble), and a second compressor that offers a couple dB of compression only on the peaks at an 8:1 ratio.

Next we're adding a Waves CLA-2A compressor. For some reason this compressor just makes everything sound better, and i'm using it here because I really want to slam the heck out of this vocal phrase. Every word should be a uniform volume, and every nuance heard - since this is screaming and metal music, most people aren't expecting 'realistic' or 'natural' sounds. Next I use an excellent plugin by Soundtoys called Decapitator, one of my secret weapons. I'm using it to distort the vocal sound relatively heavily, and then I mix it back in with the original at around 60%. Honestly I typically only have to use presets in this plugin, and just fool around with the gain and wet/dry knobs to taste. Adding distortion to your screams really helps create a finer grit, and does wonders to thicken the sound - however, you should be careful to apply it sparingly or it sounds super lame.

Next in the signal chain we have some limiting, DeEssing, and Reverb. In this case all by Waves plugins. With the limiting you're again just trying to suck even more dynamic out of the vocals. The DeEssing will help to reduce the 'S' and 'T' sounds in your vocals, which the Waves DeEsser plugin does an excellent job of doing (even just by using their presets). RVerb is just giving the vocals some space and dimension.

Going to the next signal chain, you'll notice that its mostly the same. In this case the first plugin I use is Logic's Channel EQ where i'm band-passing the screams pretty hard. I then go into Slate Digital's Virtual Mix Rack again for some harmonics and compression, into Wave's CLA-2A compressor for more lovely compression, SoundToys Decapitator with more or less the same settings as before, Waves L1 Limiter for dynamics, and finally to the next new addition of Waves H-Delay.

I've said for a while that H-Delay is my favorite delay plugin. Here i'm using it to create a band-passed delay sound on my already band-passed vocals. The delay is also very slightly modulated, with the mix being sent out as a Ping-Pong delay. One feature of this delay is the ability to model some analog characteristics of delay pedals, and I didn't notice it for the longest time. I'm using Analog setting 2, which gives the sound a very slight amount of saturation. For more aggression you can use analog setting 3, which for some reason from memory I remember it being more distorted than analog setting 4.

This just absolutely makes this vocal phrase come to life. Check the image below to see all the plugin settings.

Finally we have our New Millennium Cyanide Christ vocal part by Meshuggah.

Again, the signal chain starts with Slate VMR, CLA-2A, Decapitator, L1 Limter, and DeEsser. The real magic of this signal chain happens with the Waves Reel ADT from the Abbey Road collection. Traditionally I would double track almost all of my screaming parts to give that wide stereo image effect, and achieve some thickening on the sound. However I wanted to experiment with using 'fake' doubling for this sound. In the past i've used a technique where I copy and paste a screaming part, drop the copy down an octave or two with a lower format, crush the heck out of it with distortion, and mix it in slightly.

You can see the settings below for each plugin.

That about wraps it up for this blog post. I hope you've all learnt a little something about how screams sound raw vs processed, and about some techniques you can use to process them. I know that almost every single plugin I used here is a paid plugin by either Waves, Soundtoys or Slate Digital, but you can truly do all of this with stock Logic Pro X plugins and some hard work. Having all of these fancy plugins doesn't really make you a better producer/mixer, but it helps to automate a lot of processes that would be really hard to do with stock plugins, and it gives you a lot more creative freedom.

#vocals #screaming #mixing