If you examine the user interface in the SoundSoap window starting from the left and moving right, this outlines the basic recommended workflow. Let’s assume that we’re cleaning an LP recording which contains clicks, crackles, needle hiss, and a 60Hz hum. In this example, we’ll use multiple noise reduction tools, in a suggested order, so as to eliminate as much of the unwanted noise as possible, while preserving as much of the desired audio signal as possible.
Following this basic “left to right” order of operations will allow you to achieve quick and easy noise reduction results. In cases where you only have one type of noise (ie, clicks and crackles, broadband noise, hum), simply use the tools that are needed, and leave the others in the “Off ” position.
SoundSoap's declipper produces stunning restorations of clipped audio. Clipping is a form of audio distortion that occurs when the signal is overdriven beyond it's maximum volume range, and reduces the sound quality and intelligibility of material. Use this setting to automatically fix distortion on audio that was recorded too loud, using SoundSoap's advanced digital signal processing algorithms that detect and restore clipped audio signals.
Remove Clicks & Crackles
The first step is to remove the random clicks and crackles. With the media file in your host program and SoundSoap active, first initiate playback, and adjust the Click & Crackle slider until you no longer hear this type of noise.
Remove Broadband Noise
Next, find an area in the recording (such as the very beginning of the record, before any music starts) in which SoundSoap can “learn” the profile of the broadband noise. The host program’s playhead should be positioned at the beginning of this section. Once playback is started, press the Learn Noise button and release it once the section with noise only has finished playing. SoundSoap will analyze the audio material, determining a noise profile, and broadband noise reduction is turned on automatically. The next step is to switch the Noise Reduction to “Noise” to hear what is being removed. If any of the desired audio content is heard, turn the Noise Tuner Knob slowly to the left (counter-clockwise) until only the undesired noise is heard. Next, set the Noise Reduction to “On”. Start with the Noise Reduction knob all the way to the left (counter-clockwise). Slowly begin to turn the Noise Reduction knob to the right (clockwise). You should hear the broadband noise begin to fade away, leaving behind only the desired audio signal (music, voice, etc). It may be helpful to engage the "Track" control to automatically lower noise reduction during the louder parts of the material.
Tip: Sampling (“Learning”) Broadband Noise
Often, the noisy media that you encounter will have areas where there is just the undesired noise present, and none of the audio signal that should be kept. A simple but very common example would be video footage of an interview – this situation would contain spoken word, interspersed with silence (or what should be silence, except for the undesired noise). In cases such as this, the “silent” areas will contain the same noise that gets in the way of us hearing the voice of the person being interviewed. These silent sections of the file allow us to sample just the noise itself, essentially telling SoundSoap what to remove from the entire file. The end result is that the undesired noise is reduced equally in all parts of the file, making the desired audio signal much cleaner.
In other cases, there may not be an area present in your media file that contains just the undesired noise. In media files that contain the desired audio and undesired noise throughout the entire length of the file, it is often best to first experiment and learn the noise in different parts of the file to get the best noise reduction results. This can be done by clicking the Learn Noise button in various parts of the file as it plays. When you find the settings that work best, you can then apply, or render them, to the entire file.
Once the ideal settings are made, the plug-in must be applied, or rendered to the media file, in order to make the changes permanent. In audio editing software, the term “Bounce” is commonly used to describe the process of applying plug-in settings to a file.
Another approach to “learning” noise in a part of a media file which contains both the unwanted background noise and the desired audio is to locate a section within the file that has the lowest signal to noise ratio. For example, in a digital recording of an audio cassette, there will likely be a low level hiss throughout the file. If this file does not have a section with the cassette hiss by itself, then the next best option for “learning” the noise is to use SoundSoap’s Learn Noise feature in a part of the file where the desired audio content is at its lowest level, compared to the level of unwanted noise.
Next, click the 60Hz Hum Removal button to eliminate the low-frequency hum. If the audio content you’re working with is just a person’s voice (dialogue, a cappella singing, etc), you would also click the Preserve Voice button to automatically help filter any noise that lies outside of the frequency range of the human voice. Much of the unwanted noise should now be eliminated (or greatly reduced). In some instances you may wish to use the Enhance slider to help restore some of the "sparkle" that was present in the original recording. The Enhance slider is recommended as a final processing step, when all other settings have been made.
Finally, if at the end of your adjustments to SoundSoap's noise reduction settings you find the volume is too low, use the boost slider to increase the volume on the track. While the boost control does not cause distortion/clipping, it can cause dynamic compression to the audio. Compression provides you with more volume, but alters the range of loudness in your material, making the softer sounds louder, so there is less variety in volume levels. Be aware of adding dynamic compression to your audio when using this control, and use it carefully and sparingly.