DeRoom Pro 2 by Accentize

By Anthony Gofton

Most of my post-production work has been quite small-scale but nonetheless important to the folk concerned, in my case, I used DeRoom Pro 2 to clean up dialogue recorded during a dress rehearsal in an empty hall.  I’ve encountered numerous tools claiming to mitigate poor acoustics or reduce unwanted reverb. DeRoom Pro 2 genuinely delivers on these promises.

DeRoom Pro is a real‑time plug‑in predicated on machine‑learning and it can be authorised to a dongle or to up to two separate machines. Mac and Windows are supported, an AAX, VST3 or AU host is required, and it can process mono or stereo sources at 44.1, 48, 88.2, 96 or 192 kHz sample rates.

I used it to clean up dialogue recorded during a dress rehearsal in an empty hall, the mics were a pair of 414s on stands quite close to the lip of the stage, angled upwards.  The audio was recorded on a Zoom recorder at 44.1kHz.  The hall was obviously very reverb-heavy being empty.

DeRoom slotted in well into my workflow using it with DaVinci Resolve to begin with, then taking the track out to Logic Pro for a final shaping before syncing it in with video again for the final pass.

Drawing from another user’s advice, I opted not to try and eliminate all the reverb in a single pass. I found that whilst this works there can be some artefacts with extremely reverb-laden dialogue.  Instead, I processed the audio with moderate settings, printed the track, and then applied a second pass. The results were remarkable: the dialogue became significantly clearer and more intelligible without sounding over-processed.  I have to say even just cranking it to the max on the first pass was very impressive.

I found DeRoom’s interface both clean and intuitive, featuring four primary controls for precise adjustments:

Sensitivity:  Adjusts the balance between complete room-sound elimination and potential artifacts. A higher sensitivity aggressively targets reverb but may introduce artifacts, while a lower setting preserves more of the original ambiance, in this case I generally kept it below 20%

Reduction: This determines the depth of processing applied to the detected reverb. It works in conjunction with Sensitivity to fine-tune the amount of reverb removed.  Too much too soon can produce artefacts but balance with this Sensitivity and it really does sound remarkable.

Attack and Release:  Being just dialogue I didn’t have anything with a very fast transient so leaving these both a 0ms seemed to work for this project.  

The ability to listen to Room Only and hear what has been taken away was a useful feature, especially for more ‘shouty’ actors.

I didn’t make much use of Spectral Focus in this case as it was for dialogue.  This control lets you target the frequency bands where reverberation is most problematic.  By narrowing the focus, you prioritize cleaning specific parts of the spectrum—often midrange frequencies where vocals or instruments are most affected.  

Bottom line – I’m starting to seek out old recordings so I can use this more, it is that good!