The noise gates of PodCleaner is a noise gate which analyzes clusters of speech and not the individual samples, allowing you to produce a cleaner audio track, more natural and less “dirty” due to possible clicks and background noises.

Unlike the traditional noise gates, which are not by their nature predictive PodCleaner analyze all files, identifying those that are clusters in which there is sound and then analyzing each individual cluster based on the duration and the RMS power.

And if a cluster is “insignificant” then it is interpreted as a background noise and “cut off”.
As you can see in this picture the cluster-based noise gate is more powerful than a traditional noise gate: it just more “spike” noises:

But there’s more: the attack of cluster-based noise gate is softer and helps to dissolve to the spoken part in a more natural way.

Even the tail of speech is most sweet.

Auto threshold
Select this flag if you wish PodCleaner to analyze the audio file and set the threshold for you. It should work 99% of the times.

Set the threshold of the noise over which trig the noise gate.
It should be 5% most of times.
Try to lower it (down to 1%) if the person speaking has a quiet voice volume.
Try to increase it (up to 10%) if there’s a lot of background noise in the room in which the person is recording.

Once the filter is triggered (because it encountered samples which volume is above the threshold) PodCleaner goes back in time to better identifying clusters of spoken audio.
Values between 0.2 and 0.4 seconds are optimal, there’s no need to go above.

The time that should pass after the end of the identified cluster after which close the gate. It’s the same as hold for traditional noise gates.
Values between 0.5 and 1.0 seconds are optimal for spoken audio.

Min. Duration
The minimum duration (it includes Preload and Decay) in seconds after which a cluster is identified as “speech” and not a simple “click”.
It must be a value greater than Preload + Decay, and 1.0 seconds is optimal in most cases.

Gate Fade in and out
Fade in and out duration of the cluster. Imagine PodCleaner as a minion who rises and decrease a volume slider before and after the cluster od spoken audio.
Fade In should be lesser than Preload, and Fade Out shorter than Decay.
The longer the fades, the smoother the transition between silence and speech, but increasing those values make Min. Duration increase and you should not go above 1.5 seconds, in which cases you could lost simple sentences such as “Yes!” or “No!”, so shorter that can be confused as clicks…