Normalising volume; what it is and how do I do it?

[Petter156]

Hi.

For a long time now I´ve been wondering what 'normalising' the volume of an audio file actually means and how people go about doing it.

My impression is that it means adjusting the volume level of a track to match that of other tracks/files. So if I have say an MP3 with noticeably lower volume I could "pump it up" a bit so that I wouldn´t constantly have to adjust the volume slider on my PC.

Is this correct?

And if so, what program should I use to do this.

Thanks.

[greenmachine]

I think this is what you're looking for. The Faq and the included help file should answer most of your questions.

[IdiotSavant]

Hi.

For a long time now I´ve been wondering what 'normalising' the volume of an audio file actually means and how people go about doing it.

My impression is that it means adjusting the volume level of a track to match that of other tracks/files. So if I have say an MP3 with noticeably lower volume I could "pump it up" a bit so that I wouldn´t constantly have to adjust the volume slider on my PC.

Is this correct?

And if so, what program should I use to do this.

Thanks.

←

'Normalising' an audio file is basically changing the volume.

For example, say you've got a bunch of ripped MP3s that you're going to listen to as a compilation.. not all of them will be at the same volume, so you'll either have to mess around with the volume a lot or simply put up with the volume levels being erratic. However, you could 'normalise' them and alter the volumes so that they all sound the same volume on any given volume setting.

Alternatively you might have an album of songs where the volume seems at a low level, so you could normalise the album in order make it sound louder at any given volume setting (as loud as you would expect it to sound in the first place.)

I hope that helps a bit... I'm not really well versed on it (so I can't explain it too well) but I do it a fair bit; try to think of 'normalising' as adjusting the recording level, but after the track or tracks have been recorded.

My favorite program for normalizing is MP3Gain which - as the name suggests - only works on MP3s... but it does work a treat.

[bug80]

However, you could 'normalise' them and alter the volumes so that they all sound the same volume on any given volume setting.

←

Be carefull, you talking about two different things.

Normalizing the volume means, that the amplitude of the audio files are made the same, for example 0 dB. This is very straightforward. Algorithms that normalize the volume search for the absolute peak value of a file and after that multiply the whole file with a certain factor to make the peak 0 dB (in this example).

What you mean is normalizing the loudness which is not the same. Whether or not two files have the same loudness to our ears depends on a couple of things, for example the frequency content. Algorithms that normalize the loudness, like MP3gain and ReplayGain generally integrate the audio over small blocks and perform frequency weighting (dB(B ) filtering for example). After that, statistical analysis is performed, which results in the final "loudness" of the file.

edit: dB(B ) without a space becomes dB(

Edited August 12, 2005 by bug80

[Petter156]

Thanks, informative indeed.

I´ll try out the MP3Gain program.

What about audio editors (Audacity?), do they do normalising?

[IdiotSavant]

Be carefull, you talking about two different things.

Normalizing the volume means, that the amplitude of the audio files are made the same, for example 0 dB. This is very straightforward. Algorithms that normalize the volume search for the absolute peak value of a file and after that multiply the whole file with a certain factor to make the peak 0 dB (in this example).

What you mean is normalizing the loudness which is not the same. Whether or not two files have the same loudness to our ears depends on a couple of things, for example the frequency content. Algorithms that normalize the loudness, like MP3gain and ReplayGain generally integrate the audio over small blocks and perform frequency weighting (dB(B ) filtering for example). After that, statistical analysis is performed, which results in the final "loudness" of the file.

edit: dB(B ) without a space becomes dB(

←

Thanks man. I did warn that I wasn't well versed in it though But hey... at least my inaccuracy prompted someone who knew what they were talking about to speak up

[greenmachine]

Thanks, informative indeed.

I´ll try out the MP3Gain program.

What about audio editors (Audacity?), do they do normalising?

←

Yes, the vast majority does. But they can't normalize compressed material like mp3 without an additional decoding-encoding step, which should be avoided. Use mp3gain for normalizing mp3s, wavgain or an audio editor like audacity for altering the amplitude of uncompressed wav files.

[Petter156]

Thanks, I guess I´m going with MP3 gain then... most the music on my HDD is in MP3 - since Ogg support around the world is still a bit limited.