Vst that removes noise but leaves 'sound' behind

[studiowaveform]

Anyone know of a VST that removes noise but recognises what is and isn't 'sound'/oscillation etc?

Looking for VST so I can batch a lot of samples (or something offline with batch options)

I know there is stuff like Z-Noise that removes the captured hiss, I'm looking for something that resynthesizes a new clean version of input with zero noise so I can boost treble or spectral compress (a lot!) without the noise showing back up
__________________

[Benedict]

This is really in the realms of Star Trek.

Yes there are spectral algos that can do quite interesting things but whenever they take noise, they must be taking signal too. The human brain is the best filtering device for separating wheat from chaff so use noise gates and the like as much as possible and let the listeners' brains do the intricate filtering. Batching is the same as it will not know when to hack and when to pare with care.

Resynthesis just isn't at the level the movies lead you to believe. Probably the best are those in iZotope RX-7 but I have never used myself so for all we know, their videos are 'cooked'. There are free noise removers in Audacity & ReaPlugs but good luck getting great results as when doing post for Audio Program, I learned never to use them as they leave the speaker sounding worse than before.

Oh and get decent quality recordings first time as that will save you a world of trouble as a the more you process, the weirder it sounds.

You don't say what your situation is so I can only give you broad strokes advice.

[plexuss]

Sonible entropy:EQ

[zacha]

RX works great, though of course it always varies with the material. But it usually needs user intervention to sample “hiss” or other noise.

Both Spectral Denoise and Voice Denoise have an automatic setting, which tends not to be as precise as working from a sample, but can still work well where you have uneven noise patterns and no time to work by hand.

If the samples all have a similar noise pattern, you might find that you can learn the noise pattern on one and then use that capture to process the rest. But Benedict is right that there is always a trade off, even if you’re lucky/skilled and it’s a small one.

[Debutante]

thats called sound editing

[Fleer]

Zynaptiq to the rescue.

[deastman]

I’ve mostly used Izotope RX and Adobe Audition, and always with mixed results. It isn’t going to be an automated process, because each sound will need individual care and tweaking to get the best out of it. It certainly won’t be like the Star Trek scenario described above.

[studiowaveform]

Thanks everyone, it's for a janky hardware synth that is terribly noisy, and after adding distortion that is even louder. , and then after sampling it (which sounds great) any filtering just catches that noise.

Oh yeah, i don't care how synthesized it ends up sounding, just want them clean and a lot of top end for the filter to bite into

I'll try sonible and zynaptiq, I guess you're talking about the suite, not intensity ? (Or are you talking about intensity? I'll have to ask for another demo, tried it a while back but just for compression duties)

[ReverendLove]

Isn't a noise gate the usual tool? Every electric guitarist with single coils has one.

[whyterabbyt]

Probably the best are those in iZotope RX-7 but I have never used myself so for all we know, their videos are 'cooked'.

please dont extrapolate your specific lack of knowledge into a general one. some of 'we' have used izotope.

[Unaspected]

If you want something more rough and ready, SpecOps has an algorithm for knocking out lower level partials.

[bibz1st]

Thanks everyone, it's for a janky hardware synth that is terribly noisy, and after adding distortion that is even louder. , and then after sampling it (which sounds great) any filtering just catches that noise. ...

Then just add the distortion after recording the synth rather than on the synth itself.
EDIT
Realised thats probably what you did, ignore me.

[Smapti]

If you want something more rough and ready, SpecOps has an algorithm for knocking out lower level partials.

Seconding the Spec Ops recommendation. You can even tune the low-level partial filtering algorithm (or any algorithm in Spec Ops) to various frequency bands, which is nice if you don't want to apply noise reduction at the same level across the entire spectrum. You can also change the sample rate and FFT algorithms, which is great because it allows you to pick the style of noise reduction that works best for your project.

[Benedict]

Probably the best are those in iZotope RX-7 but I have never used myself so for all we know, their videos are 'cooked'.

please dont extrapolate your specific lack of knowledge into a general one. some of 'we' have used izotope.

I hope that making that post made you feel good about yourself. From here it just made you look like a child, what with the deliberate taking out of context and twisting the meaning to give yourself a platform to attack someone apropos of nothing that added nothing of value to the OP or conversation in general. No need to respond, you've done enough for yourself already.

As for the OP. Noise Gates or by the time you have stripped all the "noise" freqs out, your synth won't sound the same anymore as the freq profile is no longer the same.

[Dillinger]

Brusfri

Noise reduction, no phase altering. If noise profile is the same across recordings, I can envision ways to batch process with it. No resynthesis