4 New Tricks for Cleaner Vocals

De-tuned Soothe 2 Trick.

I’m not the biggest fan of using Soothe 2 on vocals since I think it often strips the vocal of some of its musical sound, but this is a good workaround.

To do this, first duplicate the lead vocal. Then, using either a tuning plugin or a tuning function in your DAW, detune the vocal down or up by 1 semitone.

Mute this detuned vocal, and insert Soothe 2 on the original lead vocal. Side-chain the detuned vocal, and be sure to enable side-chaining for Soothe 2’s detection.

With Soothe 2, isolate the detection to the lows up to the high mids and use sharper settings with high selectivity.

So, what we’ve done is find all of the resonances that Soothe 2 would typically attenuate; however, since the plugin detects detuned vocal resonances, any attenuation will occur to out-of-tune frequencies.

For example, if Soothe 2 had originally attenuated a resonance at, say, C#, it will now attenuate C instead. So, instead of attenuating in-key aspects of the vocal, the vocal’s out-of-key notes will be attenuated.

As you’d expect from Soothe 2, the vocal will sound smoother after processing; however, our processed signal will retain the in-key elements that help it sound musical.

Let’s listen to the before and after - although the effect is subtle, the vocal sounds balanced while retaining its original notes and overtones.

Watch the video to learn more >

The Cleanest Compression Possible.

After looking behind the scenes of some of my favorite plugins, I found the settings that result in the absolute lowest level of harmonic distortion.With the FabFilter Pro-C2, set to its mastering algorithm and while using a soft-knee setting, we can achieve a good amount of attenuation without any distortion.

Additionally, 2ms or more of lookahead lowers harmonic distortion even more while making it possible to utilize a fast attack and capture the full performance but without the distortion that would typically result from this.

A 50ms release will return the signal to unity quickly without causing wave-shaping to lower frequencies or the vocal’s fundamental.

Lastly, automatic makeup will amplify the quieter details of the vocal, helping it sound full and impressive, but again, since we’ve reduced the distortion that our compression would have added, the only distortion we’ll amplify is what was present in the original recording.

To be clear, distortion can and often does sound great, but if you’re looking for compression settings with the least noticeable alteration to the vocal’s timbre, this is it.

Let’s listen and notice how the vocal becomes controlled without any alterations to the timbre.

Watch the video to learn more >

Click Removal to Significantly Clean Up a Vocal

For lack of a better term, mouth sounds are non-musical parts of a vocalist’s performance - for example, the sound of the mouth opening, that kind of clicking sound from saliva, etc.

It’s not a commonly discussed thing, but I’ve always found it very distracting.

To get rid of it, try Izotope’s De-click module from their RX software - it doesn’t matter which version. You can do this in the actual RX platform, but it works just as well in your DAW.

Select the Multi-band Random Clicks algorithm and use a higher frequency skew - somewhere between 3 and 8 will work, depending on the vocal.

Use a low-sensitivity setting to avoid cutting out too much from the vocal. Similarly, use very subtle click-widening settings.

If you’d like to hear what’s being attenuated, enable the ‘Output Clicks Only’ function. Just be sure to disable this after dialing in your settings.

One thing to note is that this plugin causes significant latency that the DAW cannot compensate for, at least not with the version I have, so, once everything is dialed in, bounce or export the vocal and bring it back into the session.

You’ll need to realign it to compensate for the latency.

Let’s listen to the before and after - notice that small pops and clicks are gone, resulting in a cleaner-sounding vocal.

Watch the video to learn more >

Parallel Compressed Mids

I’ve been doing this more and more for vocals lately, and I’ve noticed how well it helps them translate to different speakers.

In short, I’ll set up a send from the vocal, and on the corresponding auxiliary track insert a linear phase EQ.

With it, I’ll isolate the low to high mids using high and low-pass filters.

The trick here is to attenuate the majority of the vocal’s fundamental and second-order harmonic with the high pass while also attenuating the majority of the vocal’s sibilance and harsh consonants with the low pass.

This way, we isolate a range that contains a lot of musical info without amplifying any frequencies that could cause a muddy sound in the lows or a harsh sound in the highs.

You could do this more accurately with higher sloped filters, but I’ve found 12dB/Octave slopes keep it sounding natural.

Then, I’ll insert a compressor and use soft-knee settings with some look-ahead and a fast attack to capture the full performance before amplifying the compressed signal with make-up gain.

A slow release really smooths out the mids, but I’d recommend varying the settings to find what works best for your vocals.

Additionally, moderate to high amounts of compression work well; I’d say anywhere above 10dB of compression will create a dense mid-frequency range.

Lastly, blend in this parallel compressed mid-range with the original vocal via the auxiliary track’s channel fader.

Let’s take a listen to the before and after - if you’re listening on a laptop or over the phone, you’ll notice the vocal translates a lot better on your speakers.

Watch the video to learn more >

Combining These 4 Tricks.

The first thing I’ll do is set up the de-clicked vocal - I’ll use the one I bounced out earlier in the video.

Then, I’ll set up the Soothe 2 trick, which means I’ll create a de-tuned track from the de-clicked track and side-chain it for Soothe 2’s trigger. I’ll use the same settings as before.

After Soothe 2, I’ll insert the compressor and use the Clean Settings I covered earlier - that is, a mastering algorithm with a soft knee, some lookahead, a quick attack, and a 50ms release.

Lastly, I’ll perform the mid-range parallel compression I just covered. This time, I’ll use the same compressor settings I just used but with a longer release.

Let’s listen to how these 4 tricks come together to create a super-clean sounding vocal that’s ready for temporal effects like reverb, delay, and any other creative processing.

Watch the video to learn more >