Maintaining real-time communication flow can be difficult when unwanted sounds and echoes interfere with the audio. Fortunately, noise reduction software and hardware are available to help.
To use noise reduction, you first need to identify the types of noise in a track and its waveform. It is called getting a noise profile.
Identifying the Source
You can try a few things if you’re having trouble locating the noise source. It is how to remove background noise from audio: first, pay close attention to the sound and note any characteristics that might help you identify where it originated. Does the noise, for instance, happen during specific hours of the day? If this is the case, you can link this to outside factors like neighboring streets or outdoor equipment.
The adaptive noise reduction effect (effects > noise reduction/restoration) is a good starting point. This effect quickly removes variable broadband noise, such as background rumble or wind. It can also help remove constant noise, such as a 60-Hz hum. It can help eliminate hiss and buzz sounds that can’t be removed with the standard noise reduction effect.
You can adjust the parameters of this effect to fine-tune its performance. Generally, decreasing the smoothing parameter allows for more significant noise reduction at the expense of more bubbly artifacts; increasing the smoothing parameter reduces these effects while maintaining more excellent noise removal.
You can also listen to the residue option, allowing you to hear what would be filtered out if you select “reduce.” Setting the wrong parameters is an excellent way to ensure you don’t damage any audio you want to keep. If you hear recognizable bits of the desired audio in the residue, your settings are too high, and you should decrease both the sensitivity and noise reduction.
Identifying the Frequency
Identifying the frequency of the noise helps to focus your efforts. For instance, reduce frequencies below 100 Hz to minimize the impact of fan and air conditioning sounds or use the DeHummer effect to remove the power line hum. Alternatively, increase the frequency volume between 500 and 1000 Hz (1kHz) to make your vocals and instruments sound more prominent.
If you need clarification on the precise frequency, consider using a sound level meter with octave bands to analyze the noise and its components. Then, you can apply an appropriate notch filter to remove the offending frequency.
Another way to address frequency-specific noise is to use the hiss reduction effect. It significantly lowers the amplitude of any frequency range that contains hiss but leaves audio in frequency ranges above the threshold untouched.
To adjust the sensitivity and threshold, click the menu button at the upper right of the graph. It displays a blue control curve, which allows you to vary the amount of reduction in different frequency ranges. The higher the sensitivity setting, the more hiss you can remove with less resulting loss in audio quality. The menu button also lets you select between linear and logarithmic spectral displays. A logarithmic display more closely resembles how human ears perceive sound. It’s worth remembering, however, that low frequencies mask higher ones more strongly.
Identifying the Level
All primary digital audio workstations include noise reduction tools that can cut unwanted sounds from a recording without refraining from the source material. These tools are designed to isolate specific unwanted sound frequencies and filter them out while leaving desirable audio components alone.
The sensitivity setting determines how much of the noise signal is processed by the effect. Higher values increase the amount of noise reduction but can introduce undesirable artifacts and reduce overall audio quality. A value between 40% and 75% generally works well.
A noise profile setting specifies the frequency bands from which unwanted sounds are removed. The range can be narrow or broad depending on the application’s needs. A more limited range is better for eliminating single-frequency artifacts like chair squeaks or power-line hum, while a broader range is more effective at reducing ambient background noise.
A processing focus button focuses the noise reduction process on a given area of the frequency spectrum. The spectrogram view can help remove breathing sounds or other subtle noises that are difficult to isolate. Finally, a noise threshold setting controls the amplitude above which unwanted sounds are removed by the noise reduction tool. Lower thresholds reduce the amount of background noise removed but may also reduce the clarity of desirable audio.
Identifying the Type of Noise
Noise is a part of every signal. It can be a random variation in voltage or current or unwanted interference in data collected by a sensor. It can even be the “snow” in video and television signals or digital recordings made with compression artifacts.
It is helpful to categorize noise into frequency-based and non-frequency-based. Frequency-based noise is often described by its color. For example, pink noise is high-frequency white noise, and blue noise is low-frequency white noise. Non-frequency-based noise includes pops, snaps, and crackles that can be heard in audio or the random dots of static appearing in video and television signals.
Identifying the type of noise is necessary to ensure that the correct effect is applied to reduce it. That is because removing specific types of noise can result in audible damage to the remaining audio.
For example, if too much of the high-frequency part of the waveform is removed with the noise reduction effect, it will sound like a hiss. It is also essential to use a sound level meter with third-octave band analysis to determine the low frequencies that make up the noise so that it can be removed properly. The spectrogram view in the audio editor is an excellent way to see this.
Arman Ali, respects both business and technology. He enjoys writing about new business and technical developments. He has previously written content for numerous SaaS and IT organizations. He also enjoys reading about emerging technical trends and advances.
