NORMALIZE - PLEASE FIX!

[Jyri T.]

Consider starting a new music or voice over project with a snizzleton (blame Warren Huart for this expression) of clips with audio levels all over the place. It would be very handy to begin with a fast process that normalizes each clip to a preferred level.

But I haven't found an easy and quick way to do so. The in-baked normalizing process would have to be applied separately to each clip.

Hornet has actually several plugins that can normalize audio to a certain LUFS/RMS/whatnot level --- but you have to run the audio through for them to be able to analyze the audio within the clips.

Why can't you do something --- having fancy tools like ARA --- to do this automatically? The user would set the specs and the CPU would do it in all but no time.

Like "set all my clips so that they are at -18 LUFS yet don't raise the peaks above - 6dB".

Should be easy enough, no?

Edited August 4, 2021 by Jyri T.

[rfssongs]

I use Normalize so much I have a shortcut key for it. Great for vocals - helps to even it out without the oppressive feel of a compressor.

Also great for making changes to small portions of a track where you don't want the hassle of an automation.

[Byron Dickens]

 

 

 

 

 

Two-year-old thread

Edited August 4, 2021 by bdickens

[Starship Krupa]

31 minutes ago, bdickens said:

Two-year-old thread

About an issue that's still an issue.

[bitflipper]

I'm surprised that so many people normalize. Honestly, I thought it was a long-discredited practice that only existed in SONAR/Cakewalk as a vestigial feature.

As an admitted non-normalizer, I'm curious as to what normalizing multiple tracks does. Does it just normalize each track individually, one at at time, or attempt to normalize them as a group? And what kind of corruption occurs? Is it a corrupt file header, or is the data trashed? I'm trying to imagine how multiplying each sample by a constant could corrupt a file, given that mathematically it's no different than using a fader.

I agree that it should be nondestructive, though. I wasn't aware that that was the case.

[John Vere]

I used to be normal but now I’m into gain.  Trouble with thinking that peak level is going to make your tracks all the same “ loudness “ is it doesn’t always work that way. It’s certainly a starting point but as said LUFS is way more important. 
 

 

[bitflipper]

2 hours ago, John Vere said:

I used to be normal

This is how all posts should begin, e.g. "I used to be normal but then learned what a DAW was and never looked back".

Your point is valid about how normalizing peaks won't make two tracks sound the same volume-wise. However valid, it's not germane to this thread because the complaints are that the feature doesn't work correctly, not about whether the feature should be used. It's reasonable to expect it to work as advertised, and even more reasonable to be angry if it broke a project.

[John Vere]

I found Cakewalks Normalization feature totally lacking many features I had learned to use in Wave lab. 
First there is no analyzing tool to tell you what the peak level of the track is. 
Second if your working clip by clip the dialogue closes after you render one clip. 
In Wave lab you can use the “ find peak level “ tool. The gain tool also has this. So you’re not guessing. 
And the tools stay open while you work on the track. Huge time saver. 

I used to tool copy into Wave Lab to  work on Tracks but now I have a few work arounds in Cakewalk I use like clip gain.  

Only time I will use normalization is on a live recording when a track is way to low in level to work with. 
Once it’s up around-12 I can use none destructive tools. 

Not that I’m too concerned about the destructive thing. There’s always a back up but I’ve never heard anything go wrong when using normalization in any audio software. 

[rfssongs]

So we have inadvertently raised a thread from the dead.

Interesting comments though.

Edited August 5, 2021 by rfssongs

[Ted K. Ling]

On 8/1/2019 at 7:35 PM, Noel Borthwick said:

Normalization is a leftover from the early digital audio days when most processing was destructive. These days with clip envelopes and 64 bit mixing there is hardly any need to do normalization unless your audio levels a super low. Even so a compressor as a clip effect would do the job.

Not to say that we shouldn't look into that issue.

Well, I use it very often when restoring (or better ... trying to restore) old cassette media, some of which is in dire straits and VERY old ... signals having been recorded terribly low.  I use "normalise to -5dB" to get me started.  AND, I also would like the multi normalisation to work, so that this workflow could be faster.

BTW, this restoration work is a sideline.

[Ted K. Ling]

1 hour ago, John Vere said:

First there is no analyzing tool to tell you what the peak level of the track is. 

Switch on the Peak Level indicator on the track.  You probably mean "Clip" ... just thinking out loud.

[Terry Kelley]

I normalize to raise the level. If I can get the level I want without it, it won't bother but there are times where with the fader is up and the prochannel is maxed and throw in the Sonitus compressor cranked to the right, it's still not enough then I will normalize, but only to -6db tops to give me some headroom. Why didn't I record at higher levels in the first place? Damn good question but I didn't and I need to fix it since I can't do it again.

I don't see any reason to be scared about it, it's just gain which is all over Cakewalk in every step. Cakewalk isn't analog. That said, I use it sparingly post-effects since many have a processing noise floor that can become obvious real quick. But Noel's point about the 'dynamic range' available makes it generally unnecessary in most cases.

And what is wrong bring up an old thread? There seems to be people that just can't handle it. And if they start a new thread, someone will post that a thread already exists. Put down the badge and enjoy the forum.

Edited August 5, 2021 by Terry Kelley

[John Vere]

36 minutes ago, Ted K. Ling said:

Switch on the Peak Level indicator on the track.  You probably mean "Clip" ... just thinking out loud.

You would have to play the whole track to get a reading. An analyzer does this in 3 seconds.  My solution now is the You Lean loudness meter and you drag and drop the clip or track for instant analysis 

By clip I’m referring to working with short clips. Not clipping the signal 

Edited August 5, 2021 by John Vere

[Ted K. Ling]

2 hours ago, John Vere said:

 My solution now is the You Lean loudness meter and you drag and drop the clip or track for instant analysis 

Hi John, yes ... this is a way.  As I only got Youlean a few days ago (after your plug, thanks) I will give this a go, but I think you'd have to export the track/clip to be able to use Youlean "offline", so there is still a time trade-off.   If I have a bunch of songs, I put them all on separate tracks, all starting at zero, then run them in real time with volume = 0 to get peak levels on all.  Exporting each one to run thru' Youlean could be faster or slower depending on how many tracks you need to check.  YMMV.  Anyway, I'm grateful to you for introducing me to Youlean and I paid for it.

Edited August 5, 2021 by Ted K. Ling
More info ...

[John Vere]

If you have the Pro version as you say ,   you just drag and drop the audio track  directly from the Track Pane into the GUI. You can drag and drop just about anything that has audio including movies. 

I put the YL meter on my Master bus and then if I open the YL meter GUI , I can drag tracks or clips to that. I also just drag the effect from track bin to track bin as I need it. No point opening more than one instance. The drag and drop feature was the main reason for my purchase. I don't think any of the others do this. Also the ability to set custom targets is another feature unlocked by Pro. 

I do use the Cakewalk Level meter on real short clips as that is faster just to play a 2-12 second clip and see where it's at. 

To do midi tracks you do have to freeze them first. 

To analyze a bunch of songs I just open my Wave Export folder where my masters are in the Cakewalk Browser. Then I just drag and drop from there. You can also use the stand alone version for this. 

Edited August 5, 2021 by John Vere

[Ted K. Ling]

Thanks for the info.  I'll try it.  Lots to learn ?

Later: Just coming back to say "thank you" again ... tried YL in situ in CbB pulling WAV files from my library and testing LUFS in the plug-in is easy.  I dunno why I was under the notion to have to export out first.  Real time saver.  I suppose it's RTFM time.

Edited August 6, 2021 by Ted K. Ling
More info ...

[Gswitz]

When you record a full band all night, you might make gain adjustments during the night. When you mix each song, normalizing the tracks gets the gain level into the fx fairly predictable from song to song. This way you can save time and mix faster. Each song will need adjusting, but it is something i will use.

[William W. Saunders, Jr.]

I only normalize a final mix (to -1 dB) before exporting, and then send it out through my mastering software in the FX bin.  I started doing this after reading a post or article by Craig Anderton which (as best as I can recall) suggested doing this.  I've found that normalizing tracks before mixing is counterproductive.