We take a deeper dive into the new features of iZotope RX9 in our full review, where we put to the test features such as Restore Selection and the time-saving Spectral Editor plug-in that operates directly in DAWs. The Dynamic filter type uses up to 1,024 dynamic notch filters that adapt to incoming signal levels, resulting in a more natural-sounding reduction of unwanted sounds in a manner that’s less likely to affect vocal tracks. The De-Hum module has also been upgraded with a new Dynamic filter type, which covers a much wider range than the original’s Static type (which covers the fundamental frequency and up to 16 harmonics). Essentially, this helps you give ADR performances from a studio the same ambience as recordings made on location.
EG I'm only automating the reduction around and threshold.Another leading feature from RX 9 is a Complex mode for Ambience Match, which lets you match the noise floor between two different recordings with ease. Then individual really noisy clips will be rendered at source to match the rest. Its great, try it! I also feel the same about EQ - EQ the dial so it sounds right, then have the plugin react to that reality. I also find riding faders into the plugin is very helpful: A low line needs to be pushed up, which means more noise comes up - the plugin on the bus reacts to all of this and keeps the noise level smooth. We work this way more with documentaries.
If all your scene's dialog is on one track, then having the plugin on a track as the same effect. Given I've got lovely crossfades between clips on separate tracks, making the noise lovely and even, noise reduction on the bus works better.īut then I do a lot of chequer boarding. I found, with the plugin on tracks, each time the playhead reached an active clip, you can hear the plugin kick in making the first 1 sec of each clip very noisy while the plugin analyses the noise. Didn't like it on individual tracks but then this may be because I do a lot of chequer-boarding. It's not like a dynamic EQ or a de-esser where it pushes down harder the harder you hit it, it pushes down a static amount based on the noise profile it's learned, and then backs off when the signal is different from the noise profile.
It can't really "attack" noise upon first hearing because the dialogue denoiser doesn't have an a priori concept of noise, noise is whatever you tell it noise is. It's like a Dolby SR decoder: the expected input level has to be static relative to the learned signal otherwise the dynamic expansion will have the wrong gains. It may work differently for different processors, but izotope at least is expecting the incoming signal to have a stable noise floor level, its reduction bands are relative to the incoming signal gain. I will say on the other hand, I do do a lot of automating of the reduction and thresh parameters in izo, this actually works if you want to scrub up little bits and pieces frame by frame If you have a dynamic gain structure leading into a denoiser like that it has to move slower than the integration time of the adaptive logic. If you are making the kind of nudgy move you're talking about and you're using it in adaptive, it's never going to catch up because the integration time on the automatic detector is something like a second or more. RandallIt may work differently for different processors, but izotope at least is expecting the incoming signal to have a stable noise floor level, its reduction bands are relative to the incoming signal gain. This is a nice way to work if you are setting noise reduction for a scene as opposed to a bite. Having the Noise reduction plug in post fader means you are riding into it (so the processor is theoretically evening out noise floor issues as you raise a bite of dx). Don't take this too personally, but I don't really agree with this assertion.