Persuading a problematic vocal recording to play nicely with the rest of your mix can seem like a futile task. Well-recorded vocals and poorly-recorded vocals both need to be correctly prepared, and the processes we're going through today will help you turn your untreated vocal take into a polished and commercial sound.
Note: this tutorial contains embedded audio that will not display in a feed reader. Click back to the site to read the tutorial with audio or download the Play Pack at the end of the tut.
1. Get Rid of Background Noise with Gating
The first thing you’ll probably notice about any vocal recording (good or bad) is the presence of background noise, ranging from subtle to annoying. This can be caused by anything; a computer in the studio, traffic, background conversation and, of course, headphone output bleeding onto the microphone.
Before the core of the vocal signal can be tackled and processed efficiently, as much of this background noise as possible must be removed. You have a few options here as there are plenty of techniques available, including manually editing the areas of noise or even using automation to lower the level of these regions. Although perfectly valid techniques, the most convenient tool for tackling the issue is a noise gate.
Noise gates are supplied in one form or another with pretty much every DAW (digital audio workstation). Products such as Cubase, Logic, Ableton and Digital Performer all include feature-rich gate plug-ins out of the box. I’m using Logic Pro 8 here, but the techniques are generic and can be used with any software and similar plug-ins.
Once you have selected your gate plug, it's simply a case of setting the threshold so that the body of your vocal is heard but the areas of silence or noise are muted. Even with the most precise setting here, you may notice that your vocal sounds unnatural when coming in and out. This can be solved with some tweaking of the attack and release settings. Some gates will have extra controls such as hold time and gain reduction amount to get an even more natural sound.
2. Dealing with Low Frequencies
Once your vocal is playing back without any noise it’s time to start thinking about a chain of processing that will compliment the vocal and bring out the best in the performance. The order that I have here is by no means set in stone and is simply one way you can work. Once you have a good understanding of each process you can mess with the order of the processors as you see fit.
It is important to filter out low frequencies and it is always a good idea to do it early in your chain as this will allow any plug-ins after this point to operate at optimal capacity since they won't have to process large amounts of unwanted low frequency energy. Many problems due to vibration, further noise during the recording and so on can be eradicated with this simple process.
I have used an EQ plug-in with the high pass filter engaged. Watch the slope and frequency you use here, as these will need to be adjusted depending on the type of vocal you are working with. For instance, I was able to filter quite a lot of low end as I am working with a female vocal.
3. Controlling Dynamics with Compression
Of course, it’s likely that in any vocal performance the singer is louder at some points and quieter during others. This may be because the piece requires lots of dynamic variance, or just that they were too close or too far away from the mic at times. Whatever the cause, if your take contains these issues you are going to want to fix them and this is where dynamics processing comes in.
The use of compressors to control dynamics should be no secret to even a beginner in the field of production. Even the most simple production software will include some form of compression plug-in or limiter that is capable of evening the peaks and troughs in any audio signal. Vocals react especially well to compression and once the correct attack and release settings are applied, pretty large amounts of gain reduction can be used without introducing any negative side effects to your sound.
Again, if you are not sure how to set up a compressor, start with an initiated patch and try some subtle settings to begin with. You can resort to presets to get the ball rolling and tweak things as you go. Once you have a compressor working correctly on your vocal, you should be able to enjoy an even and more powerful overall sound. If you still feel there are some wayward dynamics leaping out of the vocal channel, some light limiting can be used to put a cap on things.
4. Take the Hiss Out with De-Essing
If “ssss” and “ch” sounds are causing spikes in your vocal recording, these need to be processed out. This unwanted artefact is called sibilance and is pretty hard to avoid during the recording process.
You can tackle mild cases of sibilance with a high pass EQ or low pass filter. You may even prefer reducing the gain of the problem areas manually in an audio editor, but again, the easiest route is to use a processor called a de-esser, an automated process designed especially for the job.
De-essers are pretty common plug-ins with a simple set of controls. Usually one would set the frequency of the problem area and then move onto the sensitivity or threshold setting to ensure the process only kicks in when needed. If you are in any doubt try using the plug-in presets as a guideline and starting point.
You can hear the de-esser working on the ‘its’ section of the vocal here. Be careful - over use of the process can effect the signal in undesirable ways and make words sound unrealistic and illegible.
5. EQ and enhancers
After using compression and de-essing, it is possible you may have lost some of the important higher frequencies in your vocal. Although dynamics processing has obvious bonuses, it can leave the signal sounding slightly dull. These frequencies can be dialled back in with the use of subtle shelving equalization or the use of an enhancer.
In this case I have used Logic’s own EQ to add some upper mids, just to claw back some of the ‘air’ lost in previous processing. Be careful here as you can end up re-introducing the sibilance you spent time removing in previous steps. ‘Less is more’ is the way forward here. Keep things subtle.
6. Keep Organized with Grouping
Although there was no use for it in this example, as there is only one vocal track being used, setting up groups or busses can be extremely useful. It is not uncommon to have several similar vocals playing on numerous tracks in any one project. If this is the case, send them all through a master buss and apply your chain of vocal processors to it. Not only will this give you a more cohesive vocal sound but it can save CPU resources as you are only applying the plug-ins once.
Of course backing vocals and ad libs may need to be sent through their own buss to avoid confusion and masking. It can also be useful to apply different more subtle processes to these secondary vocals to create some contrast between them and your lead part.
7. Adding Reverb and Delay
Finally, to add space and an ethereal quality you can use some delay and reverb effects. These can be used as inserts, but become especially useful when used in a send/return configuration. This way any automation on your vocal track will not effect the processing and any tails will stay intact. Also the delay and reverb can be used on any channel in your mix. Try not to swamp the vocal with extra effects as it will lose presence and clarity.
8. Pitch correction and other trouble shooting
There wasn’t a huge amount left to do to this vocal as everything was working nicely after all this processing was applied, but in some cases extra work may need to be done. For instance, if the singer was a little shaky (or tone deaf!) you might feel the need for some pitch correction. This process was made famous by the Antares Auto-Tune plug-in, but there are plenty of other options. Logic 8 has a great bundled pitch correction plug.
Another final alteration you may want to make is the use of automation to reduce the level of any remaining problem transients. If you aren’t happy with any sibilance or pops that have managed to find their way though the chain of processors, you can simply apply a small dip in the volume automation lane. This will really help to soften the area and eliminate the problem.
So, after all that processing, what does our final product sound like? Have a listen for yourself: