Jump to content

Why not use normalize during bouncing?


Recommended Posts

I always tell everyone willing to listen to never, ever use Normalization in the Bounce window. Because that means that you're not bouncing what you're listening to when pressing play, so what's the point?

 

Mix it so it sounds good, then once you're happy with the sound, why would you bounce something else? Bounce what you just mixed. That means that normalize option has to be off.

 

I have tried to lobby apple to take that option away from the Bounce window, who knows if they'll listen. I strongly feel it should at least be "Off" by default.

Link to comment
Share on other sites

First, note we're not talking about the classic destructive normalize feature, but a 32 bit floating point normalize feature. That feature will yield the same result as if you were to sit here and make sure the highest peak in your project reaches exactly 0.0 dBFS, not less, not more. That means the normalized result will be different from the non-normalized one in two ways:

 

1) If your project peaks below 0.0 dBFS, the result will sound louder.

 

2) If your project peaks above 0.0 dBFS, the result will sound softer AND cleaner (since all clipping will go away).

 

I find it easier to get the sound (and level) I want as I'm listening, then bounce exactly what I just listened to, and at the same level.

 

On top of that, who wants their project to peak at exactly 0 dBFS without using a limiter? Either you're doing a mix and you usually want it to peak a little below 0 dBFS, or you're mastering and you usually use a limiter to take care of that.

Link to comment
Share on other sites

Isn't a normalized file meant to sound exactly the same as a non-normalized file? Does normalizing actually degrade the sound quality?

 

First, note we're not talking about the classic destructive normalize feature, but a 32 bit floating point normalize feature. That feature will yield the same result as if you were to sit here and make sure the highest peak in your project reaches exactly 0.0 dBFS, not less, not more. That means the normalized result will be different from the non-normalized one in two ways:

 

1) If your project peaks below 0.0 dBFS, the result will sound louder.

 

2) If your project peaks above 0.0 dBFS, the result will sound softer AND cleaner (since all clipping will go away).

 

I find it easier to get the sound (and level) I want as I'm listening, then bounce exactly what I just listened to, and at the same level.

 

On top of that, who wants their project to peak at exactly 0 dBFS without using a limiter? Either you're doing a mix and you usually want it to peak a little below 0 dBFS, or you're mastering and you usually use a limiter to take care of that.

True... I suppose that when I'm mixing, I try to get the peaks near to 0dBFS, but the reason I'd normalize a mix would be so that when I'm testing it on the car stereo and other systems, I'm getting a mix that's as loud as it can be without delving into the mastering world (if I'm planning to get the track mastered by someone else). But you're right, if my mixes are hitting close to 0dBFS then it shouldn't make much difference anyway. I was just curious as to whether Logic's normalization actually degrades the sound.

Link to comment
Share on other sites

I was just curious as to whether Logic's normalization actually degrades the sound.

 

I wouldn't worry about it unless you actually LIKE the way your mix is clipping the Out 1-2 (if it is peaking above 0 dBFS that is). Some producers like a little output clipping. If you're one of them, the Bounce's normalization feature will ruin it for you.

Link to comment
Share on other sites

I was just curious as to whether Logic's normalization actually degrades the sound.

 

Not noticeably, done one time.

 

 

For hobby projects without mastering:

 

- generally bounce with normalize on, as this will either help make your song loud enough compared to other music, or prevent clipping

 

- the exception is if the levels on the Out 1-2 are going into the red on kick beats, and you specifically like the clipping distortion this is causing

 

 

For professional projects with mastering:

 

- you absolutely should not normalize after mastering

 

- the mastering engineer will probably yell at you if you give him a normalized stereo master (though it doesn't make that much difference with a 24bit file)

 

- there might be some circumstances where you'd normalize a track bounce depending on your workflow

 

- you'd normalize when doing sound design (but probably in an audio editor)

Link to comment
Share on other sites

I don't mind having a Normalize option, especially since I can't hear any signs of reduced sound quality after the normalization process anymore - but mastering guys prefer non-normalized files, because they need some headroom, so if the normalize function would have an option to normalize to eg. -6db (or follow the normalize setting for the Sample Editor), more people would be more happy...
Link to comment
Share on other sites

Keep that master below -6db at all times :wink:

Please explain your reasons for this ridiculous statement.

 

That's actually good advice.

 

You want to leave some headroom for mastering in order to avoid internal overloading if the first process is digital, i.e. the signal will change its peak value (up to 3 dB or more) due to phase changes.

 

Even if the signal goes straight to D/A you could easily have overshoots causing 0 dBFS+ distortion at the D/A if your signal is close to full scale. This is caused by the inherent lowpass filtering during digital to analog conversion but could also be caused by simple analog waveform reconstruction of square waveforms (up to 6 db) - hence the -6 dBFS. At 24 bits a peak of -6 dBFS will not affect your quality, i.e. don't be afraid you'll lose bit quality because your signal isn't maxing out.

 

For hobby projects without mastering:

 

- generally bounce with normalize on, as this will either help make your song loud enough compared to other music, or prevent clipping

It won't really make any difference in terms of loudness. We evalute loudness primarily based on RMS (average) and not peaks. Since normalizing only addresses peaks it usually has very little impact in terms of loudness.

Link to comment
Share on other sites

We evalute loudness primarily based on RMS (average) and not peaks. Since normalizing only addresses peaks it usually has very little impact in terms of loudness.

 

Exactly, this is why you can find a very special normalization tool called an "RMS Normalizer" in

various editing programs. This will bring up the RMS or average level and needs to be used very carefully because it has the potential to ruin a mix. The relationship between the different levels can be adjusted where the relationship between levels is fixed in peak normalization.

Link to comment
Share on other sites

This way, you are giving the mastering engineer some headroom to work with.

This is is utter nonsense - An myth that gets passed around internet forums.

If a mastering engineer needs more "headroom to work with" he will attenuate the signal accordingly.

 

The actual reason for avoiding 0dBfs is a precaution against intersample peaks that may not be indicated by your digital peak meters. No other reason.

 

If you wish to learn more about intersample peaks, this is the standard reference {dead link removed by admin}.

Link to comment
Share on other sites

This way, you are giving the mastering engineer some headroom to work with.

This is is utter nonsense - An myth that gets passed around internet forums.

If a mastering engineer needs more "headroom to work with" he will attenuate the signal accordingly.

 

The actual reason for avoiding 0dBfs is a precaution against intersample peaks that may not be indicated by your digital peak meters. No other reason.

 

If you wish to learn more about intersample peaks, this is the standard reference

 

Read Lagerfeldt's post. He explains it very well. The myth is that there is no purpose to it.

 

And yes, intersample peaks is yet another good reason.

Link to comment
Share on other sites

This way, you are giving the mastering engineer some headroom to work with.

This is is utter nonsense - An myth that gets passed around internet forums.

If a mastering engineer needs more "headroom to work with" he will attenuate the signal accordingly.

 

The actual reason for avoiding 0dBfs is a precaution against intersample peaks that may not be indicated by your digital peak meters. No other reason.

 

If you wish to learn more about intersample peaks, this is the standard reference

 

Daft, I understand what you're thinking of but you are mixing things up a bit or forgetting the practical reality of things.

 

Read my post (again?) perhaps.

 

You say "the mastering engineer wlll attenuate the signal accordingly". Yes, I could do just that. And of course that happens all the time too. Just doing a 20 Hz lowcut as the first step during mastering can (and often will) lead to a massive jump in volume as the phase shift sets in.

 

But what I've received is a 24 bit fixed signal. Not a floating point. So attenuating that signal to avoid overloading during digital processing would be doing unnecessary processing. I'd rather have that 24 bit file peaking lower to begin with.

 

Also, let's assume I get a very hot mix close to full scale. It's been mixed with intentional clipping on individual tracks. This could lead to quite a square waveform of the entire mix without the mixing engineer ever clipping or limiting the actual master. That kind of mix could lead to 0 dBFS+ signals at the D/A, exceeding the specs of my D/A, leading to distortion. Again, I'd have to attenuate digitally, altering the signal in an unnecessary fashion. Remember that super hot mixes or masters can lead to more than 6 dB above the ceiling when played back.

 

You can read more about it here:

 

Level Control in Digital Mastering

http://www.tcelectronic.com/media/nielsen_lund_1999_level_co.pdf

 

0 dBFS+ Levels in Digital Mastering

http://www.tcelectronic.com/media/nielsen_lund_2000_0dbfs_le.pdf

 

Overload in Signal Conversion

http://www.tcelectronic.com/media/nielsen_lund_2003_overload.pdf

 

Stop Counting Samples

http://www.tcelectronic.com/media/lund_2006_stop_counting_samples_aes121.pdf

 

Programmed for Distortion

Listen to the artifacts produced when hot CDs are sample rate converted or reproduced in a CD player.

http://www.tcelectronic.com/media/Programmed_for_Distortion.zip

Link to comment
Share on other sites

For hobby projects without mastering:

 

- generally bounce with normalize on, as this will either help make your song loud enough compared to other music, or prevent clipping

It won't really make any difference in terms of loudness. We evalute loudness primarily based on RMS (average) and not peaks. Since normalizing only addresses peaks it usually has very little impact in terms of loudness.

 

This always happens to me when I try to keep things simple for new users : )

 

Yes, "loudness" is qualitative, but peaking at 0dB is certainly louder than the same material peaking at -12dB.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...