Quite often in the world of audio engineering we record unwanted sounds with our target recording, especially background noise. Let’s face it, the real world is full of mixed noises and different sounds, so unless you’re recording one sound source at a time in a super quiet studio environment, you will end up having unwanted sounds in your recordings.
For music and band scenarios, the most common source is leaking from other instruments, especially when trying to record the whole band playing together. A close second is possibly other noises in the environment, such as dogs barking, fans, air conditioner, objects being dropped, and your mum calling you for dinner.
Yes most of us mortals have home studios and record in less-than-ideal environments when it comes to acoustic isolation. So we are exposed to the unwanted capturing of these additional “features” that we may later want to cut.
Other recording scenarios are also exposed to unwanted background noise and artefacts. Outdoor interviews, live sport events, sounds of nature, public speeches, the list goes on and on. We went as far as helping a writer who had recordings of interviews with lots of background noise due to a uncontrolled environment and moving subject.
In a not so distant past this had little or no remedy. If it was annoying enough you would record it again. If not, just live with it and try to mask it in the mix. Sometimes the stellar performance of the recording outweighs whatever extra audio artefacts it has. Or the recording of a speech or interview is not something that can be repeated so you have to deal with it.
But in these days of modern DAWs and intelligent algorithms, a lot can be done to improve or even completely fix these issues. We regularly use two different tools which we would like to present in this article. They may help you in solving your background noise issues. And we’re here to help if you don’t want to source them or don’t have time to learn how to use them.
To quickly drive the point home, we’ll demonstrate both using a simple example of a speech recording with some fan noise in the background. These tools are way more capable than that, but it’s a good and consistent start.
Accusonus ERA Noise Remover
This company creates and distributes a suite of tools aimed at improving audio recordings. The whole suite is called ERA bundle and consists of a de-clipper, de-esser, noise remover, plosive remover, reverb remover and voice leveller. There is a more advanced bundle with a more advance restoration and repair tool as well.
In this particular case we will focus on the noise remover, but in our voice overs and even youtube videos we’ve used the whole lot and they are quite amazing.
This tool installs as a plugin for your DAW. We are using the AU version on Logic Pro X, and it can be added to any track. The key here is really how simplistic and effective this tool is. There is one main control. Yes you read correctly, one main control.
Additional minor controls allow you to set the level of the output and the preferred focus of the tool (flat, low frequency, high frequency, mid frequency or high low) so you can direct it better to the noise you are trying to reduce.
But the main control, which sets the amount of processing to be employed, is just one big knob at the centre of the tool, obvious and intuitive.
The experiment we did was to fiddle with this to reduce the unwanted fan noise in the speech recording while trying to keep the speech audible and intact. Of course there is a trade-off, the more you remove the more it will affect the recording of interest.
Loop sections of your recordings, start low with the controls and work your way up until you’re happy with the trade off achieved. Save the settings as a preset for future use in similar situations.
The single setting works well as a one-size-fits-all if your noise source is constant throughout the recording. If it isn’t, then you may have to cut the recording in multiple clips and apply different settings individually, or use some automation to vary the settings as the recording goes.
We have included a sample of the before and after recording later in this article.
Izotope RX 7 Voice and Spectral De-Noise
This company is quickly becoming a behemoth in the area of audio restoration and repair. While their toolset does include DAW plugins as well, the main benefit of paying the extra premium they command is their audio editor, which allows detailed editing in the time or frequency domain, down to the level of scrubbing individual audio artefacts out one by one. While the review of the editor is not the subject here, it is worth considering if you have a lot of audio restoration and repair work to do like we do.
They provide a voice de-noise tool aimed specifically at leaving the voice intact as much as possible when removing the noise, and a more generic spectral de-noise tool option. In this example, we’ve used the voice de-noise as we have speech and a fan noise only in the recording.
The main feature of this tool is that it has an adaptive mode and a learning mode. This means the tool can ‘learn’ about the noise by looking at a specific section of the recording, and then apply what it learning to the whole track or the segments of interest.
This is achieved by looping the section where you want the learning to happen, and is useful if you have lots of variations in the background noise and want the tool to learn about each segment instead of applying a general setting. Once the learning is done, you can preview its results and adjust the frequency domain parameters as well as the threshold of detection and the amount of reduction wanted.
Needless to say, there is a trade-off here, so go gently and work your way up until you find the sweet spot.
Another option is to use the adaptive mode, which is what we did in our example as the noise source was pretty constant. That way you only have to worry about the threshold and amount of reduction wanted, and the algorithm will do the rest.
We trust your ears as much as we trust ours, so we’ve included recordings with results below. Our take is that both are VERY competent in what they do so we utilise both depending on the situation. For a quick DAW focussed job, especially in some of the voice overs we work with, the ERA bundle solves a lot of problems. If more detailed editing is needed, we bounce a high-res WAV out to Izotope RX 7, do the editing and then import it back into Logic Pro X. It really depends on what you are looking for.
Both tools do an amazing job at what they propose and your decision should consider cost and how much you would benefit from the dedicated editor that comes with Izotope.
Other important mentions go to Spectralayers, Zynaptiq, Waves and Adobe Audition. They all have audio restoration and repair tools in different forms and shapes. It’s worth downloading a lot of demos and experimenting before making a decision.
Of course full blown restoration and repair takes time, experience and dedication. It is one of the services we offer, so if you want to know more and get a quote please reach out via our contact form.
Did you like this article ? Sign up to our newsletter to get awesome articles like this regularly: