Best way to normalize or compress music volume

[master_blaster]

My setup: Windows 10 Home 64-bit, Audacity 3.0.2 (just updated from version 2)

I hate when my music gets really loud in some places, then really soft in others. I have played around with the Compressor to normalize all my music to a constant and consistent volume. Here are the settings I’ve been using, but I would like to know if there are better ones:

Threshold: -12 dB, Noise Floor: -40 dB
Ratio: 2:1 – usually, although sometimes I will go up to 3:1, or 4:1. Once I did 8:1
Attack time: 0.10 secs
Release time: 1.0 sec
Make-up gain for 0 dB after compressing: Check
Compress based on Peaks: CHECK!

I compress based on peaks because I find this acts as a reverse compressor; instead of squishing the loud parts and leaving the soft parts alone, it amplifies the soft parts and leaves the loud parts alone. When I do NOT compress based on peaks, I find that the loud parts of a song gain this strange muting or pulsing effect that I don’t like, even when I only use 2:1. When I compress based on peaks, I seem to get better results.

However, I’ve recently been getting that unpleasant pulsing effect during the chorus of a certain song (Journey – Who’s Crying Now) even when using my preferred Compressor settings. The unpleasant effect is slight and not obvious. I’m now wondering if it’s just that song or if I have been wrecking my music all this time.

Anyway, I’m looking for other ideas on how to make my music have a consistent volume without distorting the music. Should I use different settings, such as increasing the release time? I’ve tried playing around with Chris’ Dynamic Processor. So far I have used the default settings and it sounds pretty good. Does anyone have suggestions on the best settings for Chris’ Dynamic Processor?

Bonus question: Is it better to run 2 compressions at a low 2:1 ratio, or do one compression at a high ratio (4:1; 6:1)

Bonus question 2: Sometimes I like to change the bass in the Bass and Treble effect. Do you suggest modifying the bass before compression or after?

[DVDdoug]

There are no “one size fits all” presets so you just have to experiment and you’ll probably have to use different settings on different songs. There are 3rd-party compressor plug-ins if you want to try some different ones.

You also can try the Limiter. Limiting is a kind of fast dynamic compression and it’s easier to use without getting obvious side effects. I’d start with the hard-limit or soft-limit settings, and you use make-up gain if you want it louder.

However, I’ve recently been getting that unpleasant pulsing effect

Yeah… In audio terminology it’s called “pumping” (when you can hear the compressor changing the volume moment-to-moment) and you might be able to minimize it by changing the settings. Of course more compression (a higher ratio or lower threshold) will make generally make it worse.

Probably the “best solution” is to manually fade-up and fade-down the volume using the Envelope Tool but of course that’s too time consuming if you want to adjust/process your entire music library.

…I’ve got more but I’ve got to go for now.

[DVDdoug]

I have played around with the Compressor to normalize all my music to a constant and consistent volume.

Normalizing and compression are different but they are often used together for “maximum loudness”.

Normalization makes ONE adjustment to the whole file. It’s no different from changing the volume control every time a song starts. Dynamic compression changes the volume moment-to-moment making loud parts quieter and/or making quiet-parts louder. (Don’t confuse this with MP3 compression which is file compression to make the file smaller, hopefully with no effect on the sound.)

Regular normalization is based on the peaks. I have another application that calls it “maximizing” which is a better English description, but the wrong audio terminology. Basically it makes the file as loud a possible without distortion and without altering the dynamics. But it doesn’t correlate very well with perceived loudness so if you normalize all of your files some will still be louder than others. In fact, most commercial music is already normalized, including many “quiet sounding” songs.

Audacity also has Loudness Normalization which adjust to a target-loudness (so you can match the loudness of several files, etc.). But you have to be careful because you can easily push the peaks into clipping (distortion). Again, this is ONE adjustment to the whole file/song (I think it’s mostly based on the loudest part).

There are other “volume matching” tools such as ReplayGain. These are also linear volume adjustments that don’t compress the dynamics.

all my music to a constant and consistent volume…

…during the chorus of a certain song (Journey – Who’s Crying Now)

Are you SURE you want to compress ALL of your music? Some of us enjoy musical dynamics. Most modern music is highly-compressed ([u]Loudness War[/u]). To me, it makes it boring and I just want to turn it down (or off). When I’m listening to some good-old classic rock where the drums, cymbals, accents “jump out” of the mix it makes me want to turn it up!

MAYBE one of the reasons you like Journey is because it’s less compressed than the more modern stuff?

Who’s Crying Now has some pretty obvious dynamics. It starts-out quiet and has a couple of other quiet parts. There is an [u]online database[/u] of “dynamic range” for various recordings. (I don’t think they use the best algorithm for measuring dynamics but there’s no perfect-simple way of defining/measuring musical dynamics.)

I compress based on peaks because I find this acts as a reverse compressor; instead of squishing the loud parts and leaving the soft parts alone, it amplifies the soft parts and leaves the loud parts alone.

No… It’s still dynamic compression. The opposite of compression is expansion which makes the loud parts louder and/or the quiet parts quieter. Expansion is rarely used in audio production, except for noise gates that make the quiet parts quieter or make them completely silent. Usually if you want more dynamics it’s part of the original performance.

Most compressors “push down” the loud parts and then often make-up gain is used to “push everything up” to make it louder overall.

without distorting the music.

Normalization won’t cause distortion unless you “try” to normalize the the peaks over 0dB.

With compression you don’t usually get “regular distortion” but it can have side effects/artifacts. As with any effect you have to listen to make sure you are making an improvement, and especially to make sure you’re not making it sound worse. Limiting can sometimes sound like clipping, or hard-limiting can sometimes actually become clipping. (Most of the Audacity limiter settings use look-ahead so it doesn’t usually alter/distort the wave shape.)

Bonus question: Is it better to run 2 compressions at a low 2:1 ratio, or do one compression at a high ratio (4:1; 6:1)

It’s really up to you but the more processing you do the more likely you are to damage the sound. Sometimes pros do that or they do parallel compression but that kind of thing is usually a special case on one instrument/track before everything is mixed.

Bonus question 2: Sometimes I like to change the bass in the Bass and Treble effect. Do you suggest modifying the bass before compression or after?

In most cases you should equalize (bass & treble, etc.) first to get the sound/frequency balance you want because these adjustments can push the peaks into clipping. Then you might want to normalize before compression. But compression can alter the frequency balance so it can be an iterative process.

It’s not included in Audacity but there is something called multiband compression that compresses different frequency bands separately. I think it’s mostly used by “loudness war” mastering engineers to squeeze the maximum loudness out of a song.

[master_blaster]

There are no “one size fits all” presets so you just have to experiment and you’ll probably have to use different settings on different songs.

Ok, that’s what I’ve been doing so far. I’ve been compressing on a song-by-song basis.

There are 3rd-party compressor plug-ins if you want to try some different ones.

Any good ones? I tried Chris’ Dynamic Compressor once or twice and it seemed ok, but I think I prefer the Audacity compressor, especially because of the “Compress based on peaks” option.

You also can try the Limiter. Limiting is a kind of fast dynamic compression and it’s easier to use without getting obvious side effects. I’d start with the hard-limit or soft-limit settings, and you use make-up gain if you want it louder.

I tried using the Limiter for the first time. I’m not sure if I was using it correctly, but it was not nearly as effective as the compressor. I was doing the maximum hard-limit on a song with wildly different volumes and then making up the gain, but it was still way too dynamic.

Probably the “best solution” is to manually fade-up and fade-down the volume using the Envelope Tool but of course that’s too time consuming if you want to adjust/process your entire music library.

I’ve never tried the Envelope tool. Is it a steep learning curve compared to the compressor? It took me a while before I felt like I really understood how to use the compressor properly.

Audacity also has Loudness Normalization which adjust to a target-loudness (so you can match the loudness of several files, etc.). But you have to be careful because you can easily push the peaks into clipping (distortion). Again, this is ONE adjustment to the whole file/song (I think it’s mostly based on the loudest part).

Would Loudness Normalization be as effective as the compressor in making a song have a uniform volume?

There are other “volume matching” tools such as ReplayGain. These are also linear volume adjustments that don’t compress the dynamics.

There seem to be multiple ReplayGain software tools. Is there one you recommend?

all my music to a constant and consistent volume…

…during the chorus of a certain song (Journey – Who’s Crying Now)

Are you SURE you want to compress ALL of your music? Some of us enjoy musical dynamics.

No, I don’t want to compress all my music, but I would like all my songs to be a relatively consistent volume. I really hate when I have to turn up the volume of a song because it starts off with a nice soft melody, only to have my eardrums assaulted by the chorus.

Most modern music is highly-compressed

I’m not just listening to modern commercial music. I’m listening to fan-made song remakes or indie stuff on Bandcamp and Youtube which are not processed by a professional recording studio. I also like movie soundtracks, but their volume can all over the place. I recently bought about 150 songs on itunes and the volume on some of them was so low you could barely hear the song, while others were super loud.

When I’m listening to some good-old classic rock where the drums, cymbals, accents “jump out” of the mix it makes me want to turn it up!

Agreed. I think classic rock (and almost every kind of rock and pop) doesn’t need compression as these songs usually have a consistent loudness.

MAYBE one of the reasons you like Journey is because it’s less compressed than the more modern stuff?

Who’s Crying Now has some pretty obvious dynamics. It starts-out quiet and has a couple of other quiet parts. There is an [u]online database[/u] of “dynamic range” for various recordings. (I don’t think they use the best algorithm for measuring dynamics but there’s no perfect-simple way of defining/measuring musical dynamics.)

I don’t like Journey more than other bands, I was just using that song as an example of one I found tricky to compress to my liking. Sometimes I really love the soft part of a song so I crank up the volume. But then the chorus comes crashing in and I have to lower it again. Who’s Crying Now is a song where I really like the quiet parts such as the beginning so I wanted to make those parts louder.

PS: I couldn’t get your link to work

I compress based on peaks because I find this acts as a reverse compressor; instead of squishing the loud parts and leaving the soft parts alone, it amplifies the soft parts and leaves the loud parts alone.

No… It’s still dynamic compression. The opposite of compression is expansion which makes the loud parts louder and/or the quiet parts quieter. Expansion is rarely used in audio production, except for noise gates that make the quiet parts quieter or make them completely silent. Usually if you want more dynamics it’s part of the original performance.

Most compressors “push down” the loud parts and then often make-up gain is used to “push everything up” to make it louder overall.

I was just pointing out that I preferred the “Compress based on peaks” option because it makes the quiet parts louder instead of pushing down the loud parts. I get less pumping with this option.

Bonus question: Is it better to run 2 compressions at a low 2:1 ratio, or do one compression at a high ratio (4:1; 6:1)

It’s really up to you but the more processing you do the more likely you are to damage the sound. Sometimes pros do that or they do parallel compression but that kind of thing is usually a special case on one instrument/track before everything is mixed.

If I understand your response, you’re saying that running a single high compression is better than running two lower compressions?

Bonus question 2: Sometimes I like to change the bass in the Bass and Treble effect. Do you suggest modifying the bass before compression or after?

In most cases you should equalize (bass & treble, etc.) first to get the sound/frequency balance you want because these adjustments can push the peaks into clipping. Then you might want to normalize before compression. But compression can alter the frequency balance so it can be an iterative process.

I’ve noticed that when I increase the bass and then compress, the bass gets unpleasantly amplified. But I wanted to know the community’s opinion on this.

It’s not included in Audacity but there is something called multiband compression that compresses different frequency bands separately. I think it’s mostly used by “loudness war” mastering engineers to squeeze the maximum loudness out of a song.

I’m not going for maximum loudness. I would describe my goal as “relatively equal loudness” with minor dynamic range.

PS: Thank you for your suggestions and your detailed reply!

[kozikowski]

You might want to revisit Chris’s Compressor. 1.2.6 is current last I looked. I use it with stiffer compression than is normal.

The first control value is Compress ratio and I change that from the default 0.5 to a stiffer 0.77. When I do that, the shows come out very close to the local radio station volume control. Perfectly listenable, but you don’t have to constantly reach over and adjust the volume.

One serious advantage Chris has is the ability to start from anywhere. There is no pre-conditioning to get the work within range. Just point it to your work and stand back.

Chris wrote it so he could listen to opera in the car, so it knows what music and expression are. He’s also listening in a noisy environment which can’t deal with operatic volume changes.

It does have one programming problem. It doesn’t like running off the ends of sound files. It’s look-ahead but doesn’t know when to stop looking.

I never noticed this because I cut down shows from the delivered and processed files to shorter for listening. You should put “something” at both ends of files for Chris to “chew on” and then cut it off later when he gets done.

I don’t use it for music. I used to record a talk/variety show off air and listen when I wanted to in the car. I found I could get much better sound quality with a download…but. The internet show didn’t go through broadcast compressors and it turned out that one host liked to mumble quietly in his beer and the other had a thermonuclear laugh what could decimate small towns. It was unlistenable in the car. Chris solved all that.

Sounds like your problem, doesn’t it?

Koz

[DVDdoug]

Any good ones?

No, I haven’t tired them. Like I said, I LIKE dynamics and I don’t do original music production so I don’t have that much use for a compressor. Compression is mostly used during music production (mixing & mastering). And it’s used by radio stations for “constant loudness” and to be “louder than every other station”.* …Chris’s compressor is “different” from most and was designed for what you’re trying to do. Still you should decide for yourself.

I tried using the Limiter for the first time. I’m not sure if I was using it correctly, but it was not nearly as effective as the compressor.

I’m not sure what you mean by “effective”. You can get an MORE effect by setting a lower limit (and then use make-up gain or normalize to get the loudness back). But if you over-do it you probably won’t like the sound.

I’ve never tried the Envelope tool. Is it a steep learning curve compared to the compressor?

In some ways it’s easier but it’s manual so it’s time consuming and you have to listen and use trial-and-error. And if you make lots of volume adjustments it’s VERY time consuming and tedious. The “trick” is to leave the envelope end-points unchanged and fade-up or fade-down the volume so there are no sudden-unnatural changes.

Would Loudness Normalization be as effective as the compressor in making a song have a uniform volume?

They are different tools for different purposes. Loudness normalization makes ONE adjustment to the whole song so the quiet-parts remain relatively quiet and the loud-parts remain relatively loud. It does NOT compress the dynamics at all. But overall, quiet-songs will be louder and/or loud-songs will be quieter (if everything is normalized to the same loudness) so you’ll get less song-to-song variation.

Compression will even-out the variations in a song as well as even-out the variations between songs.

Compression can be followed by loudness normalization for better song-to-song matching. Radio stations use both, or they use a kind of compression that accomplishes both.

There seem to be multiple ReplayGain software tools. Is there one you recommend?

It depends… The “real” ReplayGain is a software-player feature. I use it in Winamp. ReplayGain doesn’t “touch” the actual audio. It adds a “tag” to the file that tells the player to turn the volume up or down by a certain amount for that particular song. If your player doesn’t support ReplayGain, the tag will be ignored and nothing will happen.

ReplayGain has an Album option that makes ONE adjustment to the whole album so quiet songs remain relatively quiet and loud songs remain relatively loud as was intended by the artist/producer.

Apple has a feature similar to ReplayGain called Sound Check. I use Sound Check on my iPod which is connected to my car stereo.

WaveGain and MP3Gain make “permanent” changes to the files so they work with any player. (If you have uncompressed files and you’re going to make MP3s, it’s best to use WaveGain before converting to MP3. MP3s can only be volume-adjusted in 1.5dB steps without decompressing so MP3Gain isn’t as precise.)

Note that since many quiet-sounding songs are already maximized/normalized and they can’t be boosted (or can’t be boosted enough) without clipping. For that reason, ReplayGain uses a default target loudness that tends to lower the volume of most files. Some people don’t like that but it’s not a problem if you have enough analog gain to make-up for the volume loss. The target volume is a compromise and a few files will still to too-quiet.

I recently bought about 150 songs on itunes and the volume on some of them was so low you could barely hear the song, while others were super loud.

I’d start by normalizing the quiet songs. It may not help but it’s the best place to start and it’s just a simple volume adjustment so it doesn’t affect the quality/character of the music. If there are “overly loud” songs you can simply adjust-down the volume (a negative value in the Amplify effect) or Loudness Normalize.

Agreed. I think classic rock (and almost every kind of rock and pop) doesn’t need compression as these songs usually have a consistent loudness.

It’s more complicated than “constant loudness”. You can have a song with constant loudness but also with “micro-dynamics” like drum-hits and accented notes that stand-out in the mix. A slower compressor (longer attack and release times) will leave those alone and act more like an “automatic volume control”. But slowing it down might cause pumping.

BTW - There is a Leveler effect (automatic volume control). It’s an option for the Distortion effect, but I wouldn’t consider it “distortion”. I haven’t actually tried it, but those kinds of effects tend to mess-up music (pumping-related problems) and are usually better for narration.

PS: I couldn’t get your link to work

Right… I’ve fixed it. It’s Album list - Dynamic Range DB (direct to the Journey page.) I don’t know what a “good number” is for you, but a higher number is (supposed to be) less compression.

If I understand your response, you’re saying that running a single high compression is better than running two lower compressions?

I’m saying do whatever sounds best to you! But you probably don’t need double-compression on every song.

I’ve noticed that when I increase the bass and then compress, the bass gets unpleasantly amplified. But I wanted to know the community’s opinion on this.

Whatever works for you! Some people complain about compression making “harsh-sounding” highs…



\

• The popular streaming services use loudness normalization but no dynamic compression.