When to apply noise reduction?

[thesteelydane]

I have tracked a recording of Bartok's 44 violin duets with a couple of vintage ribbon mics through a Manley tube preamp, while recording the reverb in the hall with a couple of sdc's, and I'm now mixing the album. the result is absolutely gorgeous but needless to say with that setup, I do have quite a bit of noise that I need to do something about. So my questions are these:

1) When in the process do I apply the noise reduction? I'll be using the Izotope RX plugin. Before or after I start balancing EQ and levels? It was recorded on 3 different sessions, so there are unfortunately small differences that needs ironing out.

2) Do I take a noise print of all 4 tracks at once, or do I do the main ribbon mics at once, then the reverb tracks?

Disclaimer: As I'm a professional musician and just an amateur sound engineer, I will probably get a lot of flack for posting in the wrong forum, but I'm hoping a pro will have mercy on me, and spill a bit of wisdom. It is for a commercial release, so my friends (the performers) have placed a lot of trust in my self taught abilities.

 

[audiokid]

What Manley pre did you use? I was thinking about the TNT a few days ago.
sounds like an interesting project, love to hear your finished product.

Regarding the noise,: We share similarities. I'm a pro musician turned mixing and mastering engineer ( or whatever else finds its way into my life) now really interested in paying it forward and helping or finding new talent. I also love acoustic music. Nice to have you here!

I'm personally thinking this is a trial and error but my first guess would be to apply the reduction at the end of the mix or mastering. You may be eqing some noise out or adding more of something during the mix so I'd personally wait and treat to NR as part of the mastering process, BUT! without hearing the entire track(s) thats also hard to say. If only some of the tracks have the noise, why add that into the final mix, get rid of that asap.
Maybe the noise is natural reverb that simply just needs to be mixed in a tiny bit and also time aligned yes?

I use Sequoia which has an awesome restoration suite. One of the tools is Spectral Cleaning. Sounds like that would be something you could use too.

What is the noise, audience, street, AC, etc noise?

Hope that helps

 

[niclaus]

As audiokid said, it's hard to give advice on this kind of problem without hearing first...
Try and post samples so we can better help you...
But what audiokid said make sense...

N.

 

[thatjeffguy]

Exactly what kind of noise are we talking about? Treating hiss or hum is entirely different than eliminating audience coughing or fire engine sirens.

Typically, I treat noise as close to the source as possible. This means treat the noise on each individual track prior to any other treatment you apply to the track such as EQ, compression, reverb etc. The reason is simply that when you apply EQ etc. you will be exacerbating and multiplying your problem.

Don't apply at the end of the mix (sorry audiokid, we don't agree here). Why? Because at that point you have a potentially broadened bandwidth of noise. Since noise reduction works by sampling the noise spectrum and applying reduction to the specific frequencies and amplitudes found in the sample, you will be applying a broad spectrum reduction to your mix which will inevitably effect the quality of your finished product.

Nip the noise in the bud on each track. This will apply the narrowest band of reduction to only the tracks that need that particular spectrum. You don't want the spectrum identified by a sample from Track 1 to be applied to Track 2 do you?

Also, I have to say that Izotope has saved my butt many many times. Their "Spectral Repair" module is nothing short of a miracle worker!

Best of luck,

Jeff

 

[audiokid]

agree Jeff but you never read my entire post, I only suggested this "after" if the entire mix was finished and you had no choice, then you have no choice. Until the OP clarifies what and where the noise is, we cannot know where to apply a solution.

 

[gmartina]

Don't apply at the end of the mix (sorry audiokid, we don't agree here). Why? Because at that point you have a potentially broadened bandwidth of noise. Since noise reduction works by sampling the noise spectrum and applying reduction to the specific frequencies and amplitudes found in the sample, you will be applying a broad spectrum reduction to your mix which will inevitably effect the quality of your finished product.

Jeff

thumb

definitely ...
analyze noise for each track that you think needs to be removed...

 

[Kapt.Krunch]

thumbdefinitely ...
analyze noise for each track that you think needs to be removed...

This means treat the noise on each individual track prior to any other treatment you apply to the track such as EQ, compression, reverb etc.

I'd think that would be best. Each track may have slightly different noise to remove. Getting each track's noise level down, independently, will produce much less accumulated noise in the final mix, which would likely be a different range than any individual tracks.

Also, go easy on how hard you clamp down. It may be better to make more than one pass on any particular track, using lighter settings....especially if there are long reverb tails. Getting too aggressive to try to get it all in one pass may cut those off pretty drastically.

Kapt.Krunch