Reducing Reverb with the RX De-Reverb Module

Description:

The De-reverb module in iZotope’s RX audio repair software can help reduce reverb components immediately surrounding source audio where conventional tools, such as a noise gate, are less effective. Also this module can boost the source audio to increase intelligibility and presence. In this recipe, we will look into using the De-reverb module to reduce the reverb without losing the character of the dialogue or ambience.

Sample:

This dialogue track from a documentary film has a substantial amount of natural reverb from the cathedral it was recorded in. De-reverb can help improve the intelligibility without sacrificing the ambience.

Standalone Workflow:

1. Open the audio file in the RX Audio Editor or send it via RX Connect (or open De-reverb as an offline plug-in in your DAW or NLE).
2. Open the De-reverb module [Shift + 8]. 
3. Select 5–10 seconds of audio that begins with a direct sound and contains segments of both the direct sound and the reverberant tail (the sample audio file used above would be a perfect example).
4. Click Learn inside the De-reverb. This will automatically detect a multiband reverb profile for the audio, as well as the length of the reverb tail. 
5. Now select the entire section of audio you wish to process.
6. Click Preview to listen to the audio with De-reverb applied. The master reduction amount will still be set to the default value of 10.0, but this setting is likely too aggressive for this example. Reduce this to around 2.0 or 3.0. Any value above 0 will reduce the reverb tail, but any value below 0 will enhance the reverberant signal. You can also tailor the reduction amounts at certain frequencies using the multiband sliders.
7. If you hear the reverb tail getting louder after the initial reduction, adjust the Tail length to accommodate the longer decay time. For this example in St. Paul’s Cathedral, we’ll use a Tail length of 2.9 seconds.
8. Check the box marked Output reverb to listen only to the components of the sound that are being removed. Continue to tailor the settings to remove the desired amount of reverb without affecting too much of the voice. For this example, we’ll reduce the High frequency attenuation to 2.5 and the general Reduction slider to 2.5 as well. 
9. Choosing Enhance Dry Signal can help boost the direct signal while continuing to reduce the reverberant tail. 
10. Click Process to render.
11. Repeat the steps above two or three times, relearning the reverb profile each time and processing with moderate reduction settings to tackle complex reverbs.

Plug-in Workflow:

1. Open De-reverb as an offline plug-in in your preferred audio or video editing software.
2. Select 5–10 seconds of audio that begins with a direct sound and contains segments of both the direct sound and the reverberant tail (the sample audio file used above would be a perfect example).
3. Click Learn inside the De-reverb. This will automatically detect a multiband reverb profile for the audio, as well as the length of the reverb tail.
4. Now select the entire section of audio you wish to process.
5. Click Preview to listen to the audio with De-reverb applied. The master reduction amount will still be set to the default value of 10.0, but this setting is likely too aggressive for this example. Reduce this to around 2.0 or 3.0. Any value above 0 will reduce the reverb tail, but any value below 0 will enhance the reverberant signal. You can also tailor the reduction amounts at certain frequencies using the multiband sliders.
6. If you hear the reverb tail getting louder after the initial reduction, adjust the Tail length to accommodate the longer decay time. For this example in St. Paul’s Cathedral, we’ll use a Tail length of 2.9 seconds.
7. Check the box marked Output reverb to listen only to the components of the sound that are being removed. Continue to tailor the settings to remove the desired amount of reverb without affecting too much of the voice. For this example, we’ll reduce the High frequency attenuation to 2.5 and the general Reduction slider to 2.5 as well.
8. Choosing Enhance Dry Signal can help boost the direct signal while continuing to reduce the reverberant tail.
9. Click Process to render.
10. Repeat the steps above two or three times, relearning the reverb profile each time and processing with moderate reduction settings to tackle complex reverbs.

Tips: 

• Use multiple passes of De-reverb. If the audio you are working with has a very complex reverb, such as a reverb with apparent early reflections, you may get better results after trying a few passes of De-reverb.
• Combine De-reverb with Spectral De-noise. A combination of De-reverb and Spectral De-noisecan be used to tame very reverberant signals. It does not matter whether you process with De-reverb or Spectral De-noise first.
• Enhancing reverbervant characteristics. The De-reverb module can also work in reverse to add more of the reverb tail, which can be an interesting effect for sound design or creating additional ambience. Just bring the Reduction slider to a negative amount to hear more of the reverb.