normalize?

P

paresh

Member
Hi - I posted & started a thread quite a while bk on whether you should normalize. the concsnsus, as I remember, was don't ever normalize - bec it raises the noise floor or bec it's a render(?).

Anyway, some of my tracke & sometimes imported midi files are too quiet even though I record right at the red zone. I've been using compression but recently tried normalize & it sounds better than compression to me. And I don't notice the noise floor any worse. I use C.walk GT Pro which is supposed to have a 3 db pad on the input...sometimes I push the red a little & don't really notice a difference either, even tho I've been told it can cut off the wave form a bit.

If anyone has comments, I'd appreciate it. Thanks.
 
the biggest problem with normalization is the way it accomplishes it's task. Basically it searches the entire waveform for the loudest peak and turns up the volume of the entire track with reference to this peak....this way you don't clip anywhere. The problem with doing it this way is that peak transients can vary. Say for example most of your audio averages down around -18dBFS or so. And there is one peak that happens to get up to around -2dBFS. Well if you normalize it to just below 0dBFS then the entire audio will increase only about +2dB so that your highest peak doesn't clip. This puts your average audio around -16dBFS. Not really enough to warrant using normalization.

The truth is, normalization is just an automatic "smart" volume control. It just turns up the volume. Something must be up if you say you're recording your files close to the redzone and they still aren't loud enough (and you do mean the AUDIO is close to the redzone, correct??....not the MIDI tracks? there's a big difference).
 
Normalize vs. Amplify

Well, I never used to normalize...probably because it was a feature I always ignored. My friend is taking an entire classical music program at the University of Windsor (Canada, ON.). One of his courses is "Recording Technology" or something of the like. He told me that his intructor taught them on Pro Tools and Adobe Audition, and basically told them that every track should be normalized after recording, before mix down. So, in the end, those tracks that you want to be lower in volume in the mix (ex: very subtle piano, etc.) can be "turned down" in the mix, instead of turning up all the tracks that need to be louder in the mix.

I used to use a different approach before he told me about normalizing. I wouldn't normalize anything, I would just leave those quiet tracks quiet without normalizing+turning down, and it sounded a lot weaker this way. I would then find myself "turning up" anything that needed to be louder using the on-screen mixer. That, to me, seems like it would raise the sound : noise ratio a lot more than alternative 1. After recording was complete and levels were mixed, I would mix down (still without having normalized). At the final stage I would just amplify the whole track by enough decibels while trying not to make anything clip. So, my old method is obviously a terrible, noisy, hissy, clippy way of doing things...and I think if you don't have a powerful enough pre-amp or whatever gear you're using isn't giving enough juice to the input during recording, then alternative 1 would be the cleaner way to do things.
But that's just me.
 
If there is a track that came in quiet but the noise floor is fine, you're going to turn it up in mixing anyway. Not much point in turning it up just to turn it back down in the mix. If you are working nondestructively and at 32 bit float, there is probably no difference, but it's just an extra step.

It's not a good idea to have tracks peaking at 0dB before you mix. You could get overs in the mix unless you turn down the tracks first. So if you really want to normalize, go to -6 instead of 0.
 
RecordingMaster said:
I would just leave those quiet tracks quiet without normalizing+turning down, and it sounded a lot weaker this way. I would then find myself "turning up" anything that needed to be louder using the on-screen mixer. That, to me, seems like it would raise the sound : noise ratio a lot more than alternative 1. After recording was complete and levels were mixed, I would mix down (still without having normalized). At the final stage I would just amplify the whole track by enough decibels while trying not to make anything clip. So, my old method is obviously a terrible, noisy, hissy, clippy way of doing things...and I think if you don't have a powerful enough pre-amp or whatever gear you're using isn't giving enough juice to the input during recording, then alternative 1 would be the cleaner way to do things.
But that's just me.
Click to expand...

Does not make sense if you are working properly. A gain change is a gain change. If you were truncating files between gain changes by saving to 16 bit (and to a lesser extent 24), that could be a Bad Thing.

It may be that the low level caused you to monitor the mix too quietly, and you couldn't hear problems until the final gain change. In that case, change your monitor gain to compensate for the low pre-mixdown level.
 
i think you should post up a sample of the un normalized track and a sample of the normalized track - n see what toher people think

seeing/hearing is believing - and a lot of times, theory can get in the way of practical . . .

and like ez willis said - if it soiunds right it is right!!!
 
paresh said:
I've been using compression but recently tried normalize & it sounds better than compression to me. .
Click to expand...

Well, compression and normalization are two completely different things. If normalizing the track sounds better to you, then you obviously just needed the track a little louder in the mix. This could also be accomplished by turning up the fader.....but no matter. Whatever works for you.
 
RecordingMaster said:
... and basically told them that every track should be normalized after recording, before mix down. So, in the end, those tracks that you want to be lower in volume in the mix (ex: very subtle piano, etc.) can be "turned down" in the mix, instead of turning up all the tracks that need to be louder in the mix.
Click to expand...
Interesting, that this should be taught as default step in the process. Depending on how many tracks etc, wouldn't you might just as well have to turn them all down to come in below zero on the mix -even if you didn't add this extra normalize step?

But then just as curious, now you may have a mix of tracks, say some percussion staying about where they were, and some low peak tracks (like strings, or organ, background/choral), that are now hovering up around max, above nominal.
That plus there is some good info out there that basically says that just as in analog, digital mix and gain structures -and how well plugs' react to hot levels, benefit from staying nearer the nominal' range.
I recall some decision a while back that even went somewhat in the opposite direction, that given maxed out tracks it might behoove us to actually attenuate them prior to the digi mix stage.
Go figure.
;)
Wayne
 
I'm sorry, this is the dumbest thing I've heard in a long time. Normalizing and turning up the volume are the same thing. Compression is different.

If you simply use proper gain staging when you record, there is no reason to ever do this. Normalizing doesn't do anything but raise the volume.
 
compress, then normalise

If normalising simply raises the volume then try compressing/expanding first. This will even out the overall volume, reducing the spikes and expanding (gain volume on) the softer parts. When the spikes have been smoothed and the soft parts are hotter, you have a relatively even volume throughout the whole track. Then, normalise it to get an overall louder volume throughout the complete track.
 
I did not know anything about this stuff and I always recorded all tracks loud enough, but never too close to 0 dB.
Than I turned down one track by one leaving main tracks up (solo; vocal etc).
:confused:

Few days ago I heard that was OK :o
 
sikter said:
Few days ago I heard that was OK :o
Click to expand...
Not to worry. It still is. :)
The bottom line is to try to do what's needed and for reasons that make sense. Think first to what are you trying to accomplish.
Compressing, limiting, to reduce the dynamic range. Why? (Hopefully) because it needs it and makes it sound better, or it needs to be louder compared to something else.
Normalizing? Why? There's already a volume knob at every stage in the chain.

Really, if we called it Abnormalizing would we still be so inclined to think it was the normal thing to do?
 
the reason you don't normalize is... its a destructive process.

when I appy compression/limiting etc... to a track or to the mix, I'm adding as a plug-in. if 3 months from now something better comes along, I take no loss. if normalizing was non-destructive, I would use it to get all of my tracks to at least -3db.

i tend to track hot enough but on the occassion, i will bump up the gain on a recorded track until the peak of the track hits -3db. of course, i'm assuming that all tracking software has both gain and volume setting like cakewalk sonar.

one of the mixing books i have read suggests that you get all of the tracks equally hot and then lower the volume of tracks to get the mix where you want it.

using gain + volume in sonar and mixing in the afformentioned manner is working well for me. it is very easy to get a rough mix this way and then fine tune it. i use to start with everything at -inf and then bring up the kick, bass, snare, hats, etc...
 
You must log in or register to reply here.
Back
Top