RX11 Advanced's mouth de click capability limit

I'm using the Mouth De Click feature in RX11 Advanced.
Its purpose is to remove the smacking and clicking sounds from the mouth.
Applying it too strongly causes consonants to disappear, while applying it too weakly leaves noise behind.
Does this mean Mouth De Click can't meet my needs, and manual editing will often be necessary?

Find more posts tagged with

RX11

Comments

Vocalpoint

https://community.native-instruments.com/discussion/48879/rx11-advanceds-mouth-de-click-capability-limit

Depending on what you are editing - manual edits are always needed.

RX11 is not a magic wand - it is a tool like any other.

Best tip I can offer (as a voiceover talent for 35+ years now) is to address mouth clicks while recording. You need to get a good downward expander into your workflow and you will soon find the need for RX is almost non-existent.

Jeremy_NI

@Satoshi Yamamoto Mouth De-click in RX11 Advanced is an invaluable tool for automatically removing common mouth smacks and clicks, but its primary limitation stems from the fact that non-speech clicks and speech consonant transients (like 'p', 't', 'k') share similar, sharp spectral characteristics. When the tool is set too aggressively to eliminate all noise, it inevitably begins to misidentify and remove parts of the desired speech sounds, leading to dull, disappearing, or "mushy" consonants. This necessity to protect dialogue means the automated tool cannot be set high enough to achieve flawless results across an entire track, often leaving behind smaller, more complex clicks.

Therefore, for professional-level dialogue cleanup, a combined automated and manual approach is essential. Editors commonly utilize Mouth De-click for an initial, gentle pass (or multiple gentle passes) with low sensitivity to safely remove the bulk of the consistent, less complex noises without damaging consonants. For the more dramatic and tricky noises, such as large lip smacks, editors rely on the Spectral Repair tool within the RX Editor. By visually selecting the noise on the spectrogram, Spectral Repair allows for surgical, manual replacement or attenuation of the offending sound with surrounding clean signal, which completely bypasses the automation's inability to perfectly distinguish between desirable speech and noise.

The approach normally very much depends on the file your working with. Hoping this helps!

Satoshi Yamamoto

It was very reassuring to learn how professionals handle such situations.
Thank you for the excellent advice.

Quick links

Unanswered questions