This slider controls how much the noise signal is reduced. It is calibrated in decibels. A setting of 0 dB specifies no noise reduction. Increasing negative amounts specify more noise reduction. The default value is 0 dB.
In many cases, as much as 20–30 dB of noise reduction can be used to good effect. However, because higher amounts of noise reduction can generate unwanted audio artifacts, you may want to avoid setting the NR Amount slider to its maximum value.
This slider adjusts how quickly the downward expanders and noise reduction process responds to the overall changes in the noise in milliseconds. Depending on the character of the noise, different settings of this control will produce varying amounts of artifacts in the signal, as the modeling process attempts to track the noise signal faster or slower.
The Response speed ranges from 0 ms to 116 ms. A setting of 116 ms (slow) specifies that the modeling process should not attempt to track very fast changes in the noise character. A setting of 0 ms (fast) specifies that the modeling process should attempt to follow every change in the noise character very closely.
A faster setting can yield more noise removal, but it may generate more artifacts. This is similar to how a noise gate produces chatter when attempting to track highly dynamic material. A slower setting will allow slightly less noise removal, but will generate much fewer artifacts.
This slider is used in conjunction with the Response slider. It controls how quickly DINR reduces the amount of noise reduction when the amount of noise present in the audio diminishes. Release times range from 0 ms to 116 ms. Like the Response control, a faster setting can yield more noise removal, but it may also generate artifacts. You may want to avoid setting this control to its slowest position, since it will cause the noise tracking to slow to the point that the other controls seem to have no effect.
This slider controls the rate at which noise reduction occurs once the threshold is crossed. It lets you reduce the audibility of any artifacts generated in the modeling process, at the expense of noise reduction accuracy. This is done by limiting the rate of change of the Response and Release controls to the specified Smoothing setting. As soon as the frequency threshold is reached, the full NR amount value is immediately applied according to Response and Release settings. When the frequency threshold is reached, DINR will ramp to the NR Amount level. Settings range from 0 to 100%. A setting of 0% specifies no smoothing. A setting of 100% specifies maximum smoothing.
The High-Shelf EQ (Hi Shelf) is a noiseless filter that can be applied after noise reduction has been performed in order to compensate for a perceived loss of high-frequency content. It is unique because it operates only on the signal, not on any remaining noise. The Freq slider controls the center frequency of the filter. Values range from 20 Hz to 22 kHz.
The Gain slider controls the gain of the filter. Values range from –12 dB to +6 dB. The High-Shelf EQ can be enabled and disabled by clicking the Enable button.
You can also use the High-Shelf EQ to reduce the amount of high frequencies in a signal. This is particularly useful if you are working with older recordings that are band-limited, since the high-frequency content in these is probably made up of noise and not signal.