How to make everything the same volume?

[maxamillionk]

Is there an easy and simple way of making the whole MP3 you are editing in Audacity the same volume? So that no part is louder or quieter than another? Is this even possible?

Thanks,
-Max

[steve]

It is possible to even out the volume of a track by using the Compressor effect http://manual.audacityteam.org/man/Compressor

There is also an optional third-party plug-in that does this very well: http://theaudacitytopodcast.com/chriss-dynamic-compressor-plugin-for-audacity/

[maxamillionk]

Thanks. I found that using the built-in compressor with 10:1 ratio, 0.1 secs attack time, and unchecking “Make-up gain for 0dB after compressing” did what I wanted. It’s just that the value for “Threshold” must be played with a little for it to work.

Thanks for your help,
-Max

[Davidoff]

Just wanted to say: this thread saved so much of my time! Thanks so much for posting the solution clearly, how to get the effect that one would expect “Normalize” gives - which it doesn’t.
But your Compression advice works a treat!

[Gale_Andrews]

@Davidoff

how to get the effect that one would expect “Normalize” gives - which it doesn’t.

Normalize is not intended to reduce dynamic range.

Normalize won’t be able to make the overall song loud enough if you have a few isolated loud peaks that are almost at 0 dB. That is why you use compression.


Gale

[platypus]

Hey Guys!

This way still a good option? (The post is ‘old’ 2012-2015).
Loudness is now is a better choice?
And Limiter?

I have some music files (to relax and sleep) and i need set all the files at same volume, some files are louder.

Many thanks! Have a nice weekend.

[Trebor]

Loudness is now is a better choice? And Limiter?
I have some music files (to relax and sleep) and i need set all the files at same volume, some files are louder.

Audacity now has “loudness normalization”
which when applied to each track (individually) can be used to make them similarly loud overall,
but it will not reduce the dynamic range within each track.
To reduce the dynamic range need to use compression and/or limiting.

[platypus]

Loudness is now is a better choice? And Limiter?
I have some music files (to relax and sleep) and i need set all the files at same volume, some files are louder.

Audacity now has “loudness normalization”
which when applied to each track (individually) can be used to make them similarly loud overall,
but it will not reduce the dynamic range within each track.
To reduce the dynamic range need to use compression and/or limiting.

Many Thanks!

[pobbo]

Hello…
I made a Halloween Music playlist with over 200 tracks.
Last year I used Normalization to somewhat even out the loudness, but I did it again this year and some tracks are still noticeably perceivably louder.
What’s the best way to handle 200+ tracks in a way that I can simply select all, and apply?
I’m coming from regular normalization, so if there’s even a slightly better option that is equally as conveniently applied, it would be an improvement.

Thanks so much!

[DVDdoug]

@pobbo

Last year I used Normalization to somewhat even out the loudness, but I did it again this year and some tracks are still noticeably perceivably louder.

Regular normalization sets the peaks at (or near) the 0dBFS maximum but peaks don’t correlate well with perceived loudness. (Another program I have calls it “maximizing” which is a better English word but “normalization” is the proper audio terminology.)

…The majority of commercial music is already normalized/maximized but some songs are still louder than others.

Loudness Normalization tries to adjust the perceived loudness by measuring the short-term average and taking-into account the frequency content (mid-frequencies tend to sound louder).

But it still may not be perfect. Two different people might not agree when two songs are equally loud, especially if a song has starts-out quiet and ends loud, or if one classical music and the other is heavy metal, etc.

And you have to be careful of clipping (distortion) since Loudness Normalization doesn’t check the peaks, the peaks can sometimes be pushed-up into clipping.

Some possible better/easier solutions -

Some software players support ReplayGain. Your files are scanned for “loudness” and "tagged’ with a loudness adjustment which is applied to each song before it starts playing. (I use ReplayGain in Winamp.)

Apple has a similar feature called Sound Check. (I have Sound Check enabled on my antique iPod Classic.)

If you have MP3s, there is MP3Gain which makes “permanent” changes so it doesn’t rely on your player software.

WaveGain works similarly but with WAV files.

All of the above can batch-process files. They also will check for clipping and (with the default settings) when the volume needs to be boosted to reach the target loudness, they will only boost the volume as much as possible before clipping.

Related to that… The default target loudness tends to lower the volume of most songs because many quiet-sounding songs are already peak-normalized and lowering the volume of the louder songs is the only way to loudness-match without distortion. Some people don’t like their songs getting quieter but it’s not a problem if you have enough analog gain.

[jamesb1]

I’m having the same issue with pre-mixed playlists I create in Audacity - screenshot attached ~30-40 tracks over 1.5 hours - for a sound healing practice called Innerdance. Normally it hasn’t been an issue but got a 2 day event coming up so have hired some RCF PA speakers x2.

Normally I create the mix, importing tracks I’ve downloaded in 320kps, often they are fully 0-1 on the waveform and have clipping when I load them into Audacity - is that normal? or I’m getting bad quality files? So I adjust them down manually with the envelope tool, plus mix them together between songs, then export the mix as a WAV file and play through media player.

With all the extra ‘head room’ / volume available from the PA speakers versus my normal little sound bar I use, I am finding it a lot clearer when the volume on certain tracks is louder than others - normally this wouldn’t be as noticeable.

Ideally I just want to hit play and forget about having to adjust the volume during the session - i.e. never getting too loud for people, but also for the quiet / gentle sections of delta sounds to not be too quiet.

‘Loudness Normalization’ feature would seem to be what I need, but I ran it over all the tracks and it didn’t seem to do anything?

Should I manually adjust all the tracks with ‘Normalize’ to -1dB, or something lower like -3dB ? The speakers have so much punch everything is at half on the mixer, speakers etc. so plenty of room to turn the volume up if I know it isn’t going to get too loud at some point.

Thanks for your help in advance, much appreciated

[Wrecks0]

I remember a person who has since passed away, who wrote an effect to make his classical music just as loud in the quiet parts, as it is in the loudest parts. It is probably still available somewhere. Maybe even on this forum.

[DVDdoug]

Before I get into my long answer, I recommend trying ReplayGain or MP3Gain as I mentioned above… more than a decade ago, wow!

MP3Gain can batch-process your playlist or your entire music library. (But you might want to back-up the files in case you don’t like it.)

I remember a person who has since passed away, who wrote an effect to make his classical music just as loud in the quiet parts, as it is in the loudest parts. It is probably still available somewhere. Maybe even on this forum.

That would be Chris’s Dynamic Compressor (which Steve gave a link to above).

Just to reiterate, dynamic compression (unrelated to file compression like MP3) adjusts and “evens-out” the volume moment-to-moment. That’s often undesirable in music because it reduces the dynamic contrast and it can make music boring.

Generally, you loudness-normalize a whole song so that one song isn’t louder than another but you keep the dynamic contrast. (Or you can loudness-normalize an album as a whole to approximately match other albums whiling keep the quiet songs quiet and the loud songs loud as originally intended.)

@jamesb1

Normally I create the mix, importing tracks I’ve downloaded in 320kps, often they are fully 0-1 on the waveform and have clipping when I load them into Audacity - is that normal?

It can be normal… MP3s are “more complicated” because the lossy compression changes the wave shape making some peaks higher and some lower. So even if the original wasn’t clipped, the new-highest peaks can go over 0dB. And, MP3 can go over 0dB without clipping but you will clip your DAC if you play it at full digital volume.

Many-many of my MP3s “show red” in Audacity, but that’s because the MP3 peaks go over 0dB but they aren’t really clipped.

Personally, it’s not something I worry about. As far as I know, if you do clip your DAC, the slight clipping isn’t audible and if you’re playing MP3s you’ve already decided to use a lossy “imperfect” format.

And some CDs are clipped during mastering to “win” The Loudness War. If you lower the volume they will no-longer “show red” but lowering the volume doesn’t remove the distortion.

So I adjust them down manually with the envelope tool, plus mix them together between songs, then export the mix as a WAV file and play through media player.

WAV files ARE limited to 0dB so if you are making WAVs from MP3s it IS a good idea to lower the volume a bit. You could use the Limiter to do it automatically. I recommend the older limiter because it actually allows you to set the limit to -1dB or whatever you want (Effects → Legacy -Legacy Limiter).

‘Loudness Normalization’ feature would seem to be what I need, but I ran it over all the tracks and it didn’t seem to do anything?

It should do something depending on the loudness setting you choose. The “problem is”, you may not know where to set it and it doesn’t check for clipping so you can end-up pushing the peaks into clipping. ReplayGain and related DO check the peaks and by default so for tracks that need to be boosted, they will only adjust-up as much as possible without clipping.

Should I manually adjust all the tracks with ‘Normalize’ to -1dB, or something lower like -3dB ?

NO, because regular normalization adjusts for a target peak level which doesn’t correlate well with perceived loudness.

FYI - Limiting is a “fast kind” of dynamic compression. Automatic Volume Control or Automatic Gain control is a "slow kind’ of dynamic compression. Clipping is a “bad kind” of limiting.

[jamesb1]

Thanks a lot DVDdoug for your prompt reply

Regarding ReplayGain via a media player, will that work on the Audacity mix? I guess that is more for individual files one by one? If I export a single 1.5hr mix I assume it won’t do anything?

@DVDdoug

WAV files ARE limited to 0dB so if you are making WAVs from MP3s it IS a good idea to lower the volume a bit. You could use the Limiter to do it automatically. I recommend the older limiter because it actually allows you to set the limit to -1dB or whatever you want (Effects → Legacy -Legacy Limiter).

Ok thanks. Is there any point exporting in WAV, I know it is ‘lossless’ that’s why I do, but yes all inputs are MP3. If I export in MP3 would it be any different? Anyway the initial issue with the volume I’ve noticed is playing straight out of Audacity pre-export, which I guess exporting might add more changes to how it sounds?

I will try and use the limiter on the mix instead of normalise and see what happens. I did normalise the ‘loudest’ tracks last night to -1dB, even the ones I thought I had ‘brought down’ enough and could see it adjusting them further down, I haven’t listened to it yet through the speakers so maybe those tracks will be too quiet

‘Loudness Normalization’ feature would seem to be what I need, but I ran it over all the tracks and it didn’t seem to do anything?

It should do something depending on the loudness setting you choose. The “problem is”, you may not know where to set it and it doesn’t check for clipping so you can end-up pushing the peaks into clipping. ReplayGain and related DO check the peaks and by default so for tracks that need to be boosted, they will only adjust-up as much as possible without clipping.

I asked on another forum and somebody said “If loudness normalization doesnt work for you then it means you are working with tracks with very different dynamic range.” - which is probably true as we have quite loud/powerful tracks dropping down into periods of very relaxing delta type music.

Would you agree or maybe I haven’t done it right? I just selected the entire 1.5hr mix, including loud/quiet parts + voice/spoken quote sections etc. and applied ‘Loudness Normalization’ which yes when I looked back at the mix haven’t seemed to do much to the waveforms, but maybe I need to play it through the speakers and see if there is a difference?

Thanks for your help

[DVDdoug]

@jamesb1

If I export a single 1.5hr mix I assume it won’t do anything?

Right, if you have one long file it will adjust the whole file by the same amount.

The individual tracks have to be level-matched first. It’s too late unless you want to start over.

There are leveling tools or automatic gain control effects but they usually don’t work very well. You can usually tell that something is messing with the volume. Compression and limiting can be used to help level-out the sound but you you need “extreme” settings and you might damage the sound.

Ok thanks. Is there any point exporting in WAV, I know it is ‘lossless’ that’s why I do, but yes all inputs are MP3.

For the best quality, use WAV and regular (peak) normalize so the peaks don’t exceed 0dB.

When you open an MP3 in Audacity (or any normal audio editor) it is decompressed. If you re-export as MP3 you are going through another generation of lossy compression and some “damage” accumulates. You may not hear any quality loss but it’s something that you should be aware of, and you should try to minimize the number of times it’s compressed or re-compressed.

I will try and use the limiter on the mix instead of normalise and see what happens.

Used in moderation it will have a very small effect on the sound (or loudness) but it can prevent clipping. If you simply want to prevent clipping, just use the Normalize or Amplify effect.

I asked on another forum and somebody said “If loudness normalization doesnt work for you then it means you are working with tracks with very different dynamic range.” - which is probably true as we have quite loud/powerful tracks dropping down into periods of very relaxing delta type music.

What they are saying is that it may not work as expected, or to your satisfaction, or you may not agree that the tracks are volume are matched.

For example, how do you volume-match a track that has fairly-constantly loudness with a track that starts-out quiet and ends loud? I believe the automatic tools will match the loudest part of both, but you might not say that they are equally-loud after the adjustment.

Or in general, two different people might not agree when two tracks are equally loud.

I find that ReplayGain or Sound Check work pretty well. I have an iPod (with Sound Check enabled) plugged into my car stereo, playing tracks randomly, and I don’t often find myself reaching to re-adjust the volume when a new song starts.

[jamesb1]

Oh and I used -9LUFs on Loudness Normalisation, as a quick google a Mastering website referenced that as what is used on CD’s and I didn’t want to make my mix too quiet, but maybe that was to win the loudness wars as you mentioned

But maybe I should try with -14LUFs which it seemed to say is Spotify settings for upload which is where most of the tracks came from or the default -23LUFs? and then just turn up the speakers if I think it is too low?

[jamesb1]

“The individual tracks have to be level-matched first” - can you send me a good link with info for doing that? I assume you mean using Mp3gain or similar upfront before importing the tracks into Audacity? Is there anything I can do in Audacity? The mix is still in there but yes I don’t really have time to import them all again and remix

ReplayGain I’m sure looks good but doesn’t really work for mixes right? thats for individual tracks one by one, I’m chopping changing etc.

[DVDdoug]

@jamesb1

Oh and I used -9LUFs on Loudness Normalisation, as a quick google a Mastering website referenced that as what is used on CD’s

There’s no actual “standard” for CDs. The loudest guy wins the loudness war!

Compression and limiting does reduce the dynamic contrast but without limiting, there would be clipping. The reduction in dynamics is the why there was a backlash against loudness wars.

If you use a -9dB target without limiting, you’ll get clipping unless the tracks are already loud and limited.

But maybe I should try with -14LUFs which it seemed to say is Spotify settings for upload which is where most of the tracks came from or the default -23LUFs?

The lower you go, the more room you have to adjust without clipping. You should always check for clipping with loudness normalization. -14dB is probably good but -23dB is OK too as long as you have enough analog gain in your playback system.

As I mentioned, ReplayGain checks for clipping by default so some tracks may not be adjusted-up loud enough to match the other tracks.

IMO - ReplayGain (or similar) is the best solution!

Replay Gain (usually) uses a different algorithm and it’s calibrated to a dB SLP (for “good reasons”). The default is +93dB SPL and I’ve forgotten what that approximately works-out to in LUFS but probably close to the Spotify standard.

[jamesb1]

@DVDdoug

IMO - ReplayGain (or similar) is the best solution!

Replay Gain (usually) uses a different algorithm and it’s calibrated to a dB SLP (for “good reasons”). The default is +93dB SPL and I’ve forgotten what that approximately works-out to in LUFS but probably close to the Spotify standard.

Thanks, can I use ReplayGain or Similar in Audacity though the way I want to make a mix, or I need to adjust the tracks first then load into Audacity and mix?

I don’t really want to just have a playlist of 1 song after the other in a media player with ReplayGain to level them, for that I could just use Spotify like I have before but was aiming for more immersive soundscapes now with multiple layered sounds etc

Thanks again for your help Doug, much appreciated

[jamesb1]

@DVDdoug

WAV files ARE limited to 0dB so if you are making WAVs from MP3s it IS a good idea to lower the volume a bit. You could use the Limiter to do it automatically. I recommend the older limiter because it actually allows you to set the limit to -1dB or whatever you want (Effects → Legacy -Legacy Limiter).

So what do you think is my best option to make this playlist easy to move through without lots of manual volume intervention? Session is tomorrow so I have some time for small adjustments but can’t redo it all

1. Use limiter to -1dB on all the louder tracks? with hard limit setting, others/gains set to 0dB and 10ms hold?

2. Or loudness normalise to something like 14 or 16LUFS on the louder tracks, leaving the quieter ones alone? 16 LUFS sounded fine on a louder track, 23 LUFS was definitely a lot quieter without turning the volume up

3. something else?

Per my screenshot, mix has ~40+ songs mixed

Thank you