@Richard-DeVito

Just to add a little to what @browj2 has said (apologies to John!): "Normalise" is the process by which ALL the content of an audio track is increased, or decreased, so that the loudest part(s) match but do not exceed whatever peak level is set by the user. It makes no change whatsoever to the "relative" levels of one section against another.

Jeff
Forum Moderator

@emmrecs

Hi Jeff,

Good definition but...what do you mean by "audio track" and "section?"

How do you normalise all the content on an audio track? Select all or lasso, then normalise(Br)? Or do one at a time?

Unless there is something that I am missing, Magix Movie Studio 2024 (MMS) only has "Normalize (maximum level)" and "Normalize EBU R128)." For the former, the user cannot select a peak level. Music Editor 3 has this capability.

In MMS, if you put 4 video clips on the timeline (AV on separate tracks) that have widely varying maximum levels, like one that clips and then normalise one clip at a time, look at the results. Undo. Lasso them all and do a normalise and you will get something quite different depending on which clip is right-clicked.

I am able to get the clips that are already at 0 to go way up if the first audio object can go up. In the case on my screen, normalise raises the first audio clip by 5.4 dB. Select all, right-click on the first object, normalise, and they all go to 5.4 dB. However, 2 of the objects would not go up when normalised separately, so they are now clipping badly.

The fourth audio object goes to 22.5 dB when normalised. Select all, right-click on the fourth object, normalise, and all clips go to 22.5 dB.

The next problem is just normalising one at a time and this is for @Richard-DeVito . I have 3 clips of me walking in the forest. There is a lot of background noise from the nearby highways a km away (on each side). In one, there is only ambient noise. In the second, you can hear my footsteps. In the third, I call the dog and there is clipping. If I normalise, the first one has very loud ambient noise, the second puts my footsteps at 0 dB and the ambient noise is fairly loud. In the third case, normalise does nothing and the ambient noise if faint. So, one needs to understand how this all works to get an even sound across similar clips. Calling the dog needs to be reduced by using the volume curve and/or compression. Then normalising or raising the level. Then one has to get the ambient sound equal across the clips, or use noise reduction, then normalise or raise the audio. Just normalising doesn't do the job.

There is a lot more, as you will have seen in the link and the links in the link.

John CB

@browj2

Hi John,

Fair questions but @Richard-DeVito's post says that he has one clip on the timeline so when he tries to Normalise it the complete audio (the "audio track") will be normalised to the peak level, whatever that defaults to in MMS. (I never use this feature in either MMS or VPX preferring to carry out audio editing etc. in Adobe Audition, just my preference.)

By "section" I meant one part of the audio (e.g., the "loudest") when compared with another (e.g., the "quietest"). Whatever you use, whether Normalise or Maximise, the level differences between those two will never change. The loudest will never become any quieter and the quietest will never become any louder, relative to each other!

With my "audio editing" hat on my solution to an audio track whose levels are very widely different would be; I would export it from MMS/VPX "as wav", import it to my Audition and use a Compressor to reduce the level of the loudest parts without affecting the quietest. This would then produce an output where there is not a vast difference between the "loudest" and the "quietest". Then, if needed, I would use Normalise to raise the complete audio track to have it all at a consistent volume level.

Jeff

@emmrecs

Hi Jeff,

Thanks for clarifying that. I missed the "a clip."

So, here is my layman's definition, normalise raises the volume of the entire clip such that the loudest sound goes to 0 dB. This is just like turning up the volume.

@Richard-DeVito, increase the height of the volume track and look at the object volume. There is a handle in the middle. After normalising, click and hold on the handle to see by how much normalise has increased the volume. If it's 0, that means that you have something in the audio clip that is at the top and the volume cannot be increased.

See the image below. Normalising raised the volume by 3.5 dB in this case. You can raise and lower this line yourself, but if I raise it, I'll get clipping - more parts that try go to above 0 and get cut off, distortion.

 the volume is substantially different between beginning and end of the presentation.

That is a different problem. As Jeff has pointed out, he would reduce the volume of the loudest points, then use a compressor.

For reducing the loudest points, in MMS, you can put a volume curve on the object (Ctrl+Shift+V), and put points on the curve by clicking on it, then dragging a point down to reduce the volume. See below. There is a circle on an obvious sound that clips. I have reduced the sound of the other loud points, thus giving me some headroom to raise the overall volume.

Next, is compression. Double-click on the audio part to open the Audio Cleaning interface and select the Compressor tab. Open the drop down beside Function and choose the one that most resembles your audio - music or speech. Start with medium compressed, which uses a ratio of 2. Unfortunately, there are no other parameters available with this, in particular, the threshold above which compression takes place and the make up gain.

Accept. Play back the audio, in particular the quieter parts then the louder parts to see if they are closer together.

What the compressor does is it reduces the volume of the louder parts bringing them closer to the quieter parts. In the image below, I used the defaults except for a ratio of 2:1. That is the part in the circle. So you can see that the maximum volume is reduced by 12 dB, what was -6 dB is now -3 dB, and so on. Everything above the threshold of -24 dB is reduced on a 2:1 ratio.

This gives some headroom and now the overall volume can be increased. The Compressor in MMS does this automatically, whereas most detailed compressors give you a slider, as in the image above, to raise the output gain yourself.

So, try it out as described above.

BTW, here is the result of my reduction of the loud points once mixed down and opened in Music Editor 3. You can see that the maximum points no longer go to the top. They will if I now normalize.

You may want to learn how to use Music Editor 3 or move up to Sound Forge Audio Cleaning Lab.

To properly see the results, you should get YouLean Loudness Meter 2 as per the linked thread.

Thoroughly confused?

John CB

@Richard-DeVito

Hi,

Further to the above, I should have added that you can use the Mixer FX as you have better tools there, and access to even much better tools with plugins.

John CB

@Richard-DeVito

Hello Richard,

I use normalize regularly (Alt+N) regularly. This is what it says in the manual.

Normalize (maximum level): The 'Normalize' function sets the volume of an audio object to the maximum level, without overloading the material. This is done by finding the largest signal peak in the audio material and increasing the level of the object so that this point corresponds exactly to 0 dB (full scale).

Normalize (EBU R128): EBU R128 is a broadcasting standard that normalizes the audio signal to -23 LUFS (loudness units relative to digital full level). In contrast to normalizing to maximum level, this provides sufficient room for short-term loud signals at the top. Cinema productions and classical recordings make use of this, for example.

Kind regards,
Rob