It happens; sometimes you end up recording noise into your vocal parts. There are a lot of reasons why this may happen. Often times you will not even realize there are noisey vocal parts until you get into another room. Unfortunately, current vocal production is very fickle about these kinds of mistakes.

So let’s start off with why this can occur. It is very simple. The good and bad thing about how our brain works is that it adapts. Our brain is actually wired with what some call as “novelty detector neurons”. These neurons have been found to quickly stop firing if a sound is repeated. So, if you are sitting in the same control room everyday, and there is an HVAC noise in that room… your brain is most likely not detecting it anymore. Infact even if you’re working in the same room for a short period of time, your brain will start to ignore the surrounding sounds. That is where noise removal comes in.

The most basic process of noise removal is to put a gate on the vocal part. This does a great job of eliminating all noise that is below a threshold. There are also often settings where you can also adjust the amount of attenuation. For me personally, I like to have a little bit of life in my vocal tracks. So I’m often not completely attenuating below the threshold, but I will frequently leave a little bit of life sneak past the gate.

Gates can work very well when the noise is either covered up by the vocalist, or not present when there is singing. However, there are times when you want to catch some of the breath from the singer, or when there are quiet parts that have some room noise in them. That is when you need to analyze and remove room noise. This is a similar process to using a gate, but it utilizes frequency bands. The first step is to make sure you have a good sample of the noise that needs to be eliminated. Then analyze that noise and start to adjust your threshold. What happens is the frequencies that match what was analyzed start to get eliminated. So, you are essentially eliminating the noise that was analyzed… however, there is a catch.

Remember that you are technically not removing a noise, but you are removing frequencies… that’s right… some of those frequencies are also associated with the audio you are going to want to keep. Just like with many audio techniques, it may be best to have a light touch with multiple stages instead of one heavy handed approach. The best thing to do is to make a little adjustment and listen. If you are getting what you want, then keep going; if you are getting artifacts, then it is time to split the work between a couple different processes.