Mouth De-click 등 RX 노이즈 제거 플러그인

Mouth De-click

입 소리(쩝쩝) 제거하기

Output clicks only 체크하면 제거될 예정인 쩝쩝거리는 소리만 골라 들어볼 수 있음

De-crackle

Mouth De-click으로 잘 제거안되는 딱딱 탁탁 틱틱 치직 따닥거리는 소리를 제거 

De-plosive

팝핑 제거, 파열음 제거 등

De-Ess

다른 디에서 플러그인도 많이 있으므로 이 플러그인은 옵션사항

Voice De-nosie

배경노이즈가 있는 것으로 보이는 구간을 반복 재생되게 해둔 뒤 "Learn"을 누르고 목소리(보컬)는 "Dialogue", Filter Type은 들어보면서 설정. Threshold는 학습(learn)한 주파수 형태를 얼마만큼의 강도로 적용할지에 대한, Reduction은 깎아내는 정도에 대한 것

https://www.youtube.com/watch?v=DJD8p1jRk3U&list=PLe5TlVjA95zkBIxXQi7RnSZZrdjG9yAFJ&index=12 

영상에서 추천하는 설정은 어댑티브 모드 체크, Dialogue, Gentle, Threshold는 0, Reductiond은 6에서 잡음이 많으면 좀더 높게 조절

 

 

..그밖에

De-hum

험 노이즈(예를 들면 접지가 제대로 되지 않은 기기를 서로 연결했을 때 나는 윙...... 이잉....... 소리) 제거

https://www.youtube.com/watch?v=iDRQXJmC6Eg

(추천 De-hum, De-clip 관련) https://www.youtube.com/watch?v=gWqWGiMeVvs&t=2s 

 

De-clip

클리핑(녹음시 마이크에 넘 가까이 대고 볼륨 크게 녹음이 되어 마이크가 수음하지 못하고 잘리는 현상) 복구

 

De-click = Mouth De-click과 유사

마우스 클릭음, 입 쩝쩝(이건 Mouth De-click으로도..) 소리 등 제거

https://www.youtube.com/watch?v=DJD8p1jRk3U&list=PLe5TlVjA95zkBIxXQi7RnSZZrdjG9yAFJ&index=12 

 

Spectral De-noise

https://www.youtube.com/watch?v=jR2OMTdf1ts 

https://www.reddit.com/r/iZotopeAudio/comments/kehams/voice_denoise_vs_spectral_denoise/

Voice De-Noise vs Spectral De-Noise

They are pretty different.

Generally speaking Voice Denoise is better used in real time. It has vastly lower impact on your delay compensation and cpu usage. It's easily automated. You'd see this on Dialogue Tracks in a film mix for these reasons.

Spectral Denoise can technically be ran in real time, but the algorithm is then limited to a lower quality mode, is more CPU intensive, and has a greater impact on your delay compensation. You'd see this used as an offline process to remove broadband or tonal noise from an audio clip or file. It's typically used destructively, where the tradeoff is a more powerful, flexible algorithm for complex noises. Very tweakable and surgical.

The other weakness Spectral Denoise has in my experience is with non steady state noise- noise that may ever so slightly change over time. If the noise it's trying to remove is now different than the noise you sampled as your noise print, it can run into problems and artifact. In cases of variable background noise, Voice Denoise may work better because it's not relying on a noise print, it's simply using 64 band pass gates to attenuate signal under the threshold set for each band.

So, say you recorded a podcast in a quiet room with a dynamic mic with a bad preamp. You have some consistent hiss. Either would work well in this situation, but Spectral Denoise might have the upper hand.

Say you recorded a podcast, but a phone sat too close to the mic/a bad cable and caused some buzz. Spectral Denoise would probably have the upper hand removing the buzz.

Say you have a recording of someone talking in a busy convention hall. There's a ton of background wala, shuffling, ac noise, etc. Voice Denoise would probably have the advantage as the background noise is constantly changing. (Dialogue Isolate would also be effective here)

Say you recorded some dialogue of a person speaking under a tree on a windy day, with traffic coming and going. The tone of the background noise is constantly changing so voice Denoise might have the upper hand.

Say you want to reduce the room tone of some really quiet foley. Spectral Denoise might have the upper hand. (Voice Denoise would be effective as well, but can tend to cut into the wanted signal)

Technically Spectral Denoise has an adaptive mode, but I personally don't think it's useful/effective in the majority of situations you'd need it. For voice Denoise, it's adaptive mode is pretty good, relatively speaking, but I still much prefer manually setting and automating the band thresholds, for many reasons.

 

Yeah Voice Denoise is probably more dialogue focused with how they placed their bands.

Spectral De-Noise is also really good at more complex noises such as buzzes and hums. You can separate the reduction fader and apply the reduction amount to Tonal and Noise separately, which is great. You can get really surgical with the reduction curve and be precise where you want to focus the noise reduction, whereas Voice Denoise only has 6 fixed bands.

With Spectral De-Noise, I've found there's a few parameters that I have to pay attention to to avoid certain artifacts though.

-With too long a release you can get swishing around the high frequencies at lower reduction. ( I have mine set pretty fast at 10ms as default )

-Too high a synthesis/enhancement can cause the Noise Reduced signal to be harsh and distorted sounding. (I have synthesis at 0 as default, enhancement at 5, and bump synthesis up when I find the NR is causing gaps/dulling in the high frequency signal.) If I'm hearing too much distortion in the Noise Reduced signal, usually lowering synthesis works.

-Musical Noise/Gating: if I hear musical noise in the resulting signal after NR, I may push this toward Gating. Sometimes if set too far on Musical Noise, you can get a chirping/swimming artifact in the signal, but on the other hand sometimes you need the NR to better separate the noise from the signal, so it's a trade off that you just listen for.

-Finally threshold, sometimes after NR, certain frequencies in the background noise might poke through the noise reduction, because the threshold was set too low, and the NR avoided attenuating certain frequencies above the threshold. Raising the threshold usually fixes this, at the compromise of potentially cutting into the dialogue (or whatever signal) a bit more.

Ultimately Noise Reduction just comes down to personal preference and workflow uses. I actually use WNS on all my dialogue tracks, with a Voice Denoise on my Dialogue aux/submaster. Together in that workflow I find it pretty flexible. I'll then use Spectral DeNoise or Dialogue Isolate destructively on specific clips in situations where I would need them. Eventually I'll phase out WNS and Voice Denoise for Cedar DNS.

For podcasts, I'll use a combination of WNS + Spectral Denoise because the background noise is almost always consistent and I commonly get poorly recorded dialogue with hum/buzz/interference. Abstentia DX is also really great for complex hums and intermittent noise, but from my experience is super difficult to get dialed in...

But anyways, they definitely have different uses. Hopefully this helped make clear why you'd want one or the other. It's not necessarily dependent on the type of source, both could work well for Voice or SFX, but more so other factors. But for non-real time processing I usually, personally, reach for Spectral Denoise.