Subject:NR before or after RMS normalization

Posted by: Gweilo

Date:3/17/2004 2:46:35 PM

I have some dialogue that I am RMS normalizing(with dynamic compression enabled) and then applying NR. Seems like I can hear some hiss riding up and down with the level. Is it better to apply NR before or after RMS normalizing? I figured post normalizing since the noise would likely be higher, thus easier to grab but...maybe the reduced dynamic range mixes the noise in so it's harder to extract from wanted material. Any thoughts/experiences?

Subject:RE: NR before or after RMS normalization

Reply by: SonyNateM

Date:3/17/2004 3:14:24 PM

In my experience it's almost always best to apply Noise Reduction to the raw recording if available. Any filters/processing you apply to a file "grinds in the dirt" a bit more and makes it harder to remove later on. I would certainly suggest applying NR before a process with gain compensation, such as the RMS normalize you are applying. The noise you are attempting to remove will be easier to isolate if it's level isn't bouncing around throughout the file.

Subject:RE: NR before or after RMS normalization

Reply by: Gweilo

Date:3/17/2004 3:39:16 PM

makes sense..thanks.

Subject:RE: NR before or after RMS normalization

Reply by: kbruff

Date:3/17/2004 4:56:11 PM

If you are using - SF7 you can work non-destructively and even apply the noise reduction part by part. Furthermore you can view the Spectral analysis and see just how much you have removed by doing a before and after spectral comparison.

Subject:RE: NR before or after RMS normalization

Reply by: Rednroll

Date:3/17/2004 9:12:14 PM

Another point of doing the noise reduction first is that the noise reduction may actually effect your RMS levels, so then you might find that you have to do another RMS normalization process afterwards to compensate, and it's always best to try and keep processing steps to a minimum to avoid additional distortion added by each process. In my opinion Normalization should always be the last step of processing followed by bit dept dithering if you're at a higher bit depth than 16bit.

Subject:RE: NR before or after RMS normalization

Reply by: Gweilo

Date:3/18/2004 2:44:55 PM

Rednroll,
You mean dither if the bit dept is being converted from say 16 to24? i.e. the conversion causes a need for dithering? Or dither at higher bit rates because the increased accuracy will make dithering benefits audible?

Subject:RE: NR before or after RMS normalization

Reply by: MJhig

Date:3/18/2004 3:00:43 PM

No, and Red will correct me if I'm incorrect as to what he's saying, when processing audio, normalizing should be the last step unless you need to resample or convert to a LOWER sample rate and/or bit depth.

So the chain would go, all processes (effects, editing etc.) then normalize, then re-sample down if need be, then bit convert down if need be.

Re-sampling or bit converting up serves no purpose related to quality. CD audio is 16/44 so if the audio is higher than that it's necessary to resample and/or convert bit depth down.

MJ

Subject:RE: NR before or after RMS normalization

Reply by: Rednroll

Date:3/18/2004 10:14:50 PM

Exactly....