Compressor settings for live music w/ a few sharp peaks

[lambdalove]

Audacity 2.3.2
macOS Catalina 10.15.2

Hello! I’m arranging live music tracks (files) I’ve acquired into compilations (like old fashioned “mix tapes”), and I’m using Audacity to apply basic edits such as fading in and fading out. The tracks are from various sources and are various quality, but they are all recordings of rock bands.

I’ve found the Normalize effect to be very useful for making all the tracks have a similar volume level. However, sometimes I come across a track that has just a few sharp peaks in volume - this is a good example. Here the peak at 3:00 is a snare drum hit that’s significantly louder than anything else:

The track as you see it here has already been normalized to the default setting of -1.0 dB, but it sounds significantly quieter than other tracks in my compilations (I assume it’s because there are a few peaks in volume that are very steep and high compared to everything else in the recording). So the listener is probably going to have to adjust the volume of their player as they listen, and I’d really like to avoid that.

I think the Compressor effect is what I might use in this situation:

But I’m rather new to editing audio and I don’t know what the best settings for the effect are, given my situation. Would it be possible for someone to recommend the best settings to use for editing live music when there are a handful of tall peaks in volume?

Thank you!

[kozikowski]

Most of the Audacity tools work on sound tips and peaks, so setting loudness (not directly related to peaks) can be a challenge.

You can try RMS Normalize. That one does directly affect loudness and follow that with Effect > Limiter with Soft Limit selected to tamp down extreme peaks like your example.

https://forum.audacityteam.org/t/rms-normalize/45334/1

There are legacy broadcast processors that do those two jobs and did for a long time.

There is an all-in-one processor called Chris’s Compressor. He wrote it so he could listen to opera in the car without having to turn the volume up and down for the dramatic volume changes.

https://theaudacitytopodcast.com/chriss-dynamic-compressor-plugin-for-audacity/

Koz

[Trebor]

… I’m rather new to editing audio and I don’t know what the best settings for the effect are, given my situation. Would it be possible for someone to recommend the best settings to use for editing live music when there are a handful of tall peaks in volume?

Those “tall [drum] peaks” are a fraction of a second in duration.
That calls for a limiter rather than a compressor, (as the latter is too slow to respond) …

https://manual.audacityteam.org/man/limiter.html

[kozikowski]

You might start with these settings. These are the settings for audiobook processing which has very strict requirements for volume and peaks. They’re in abbreviated form, but in general, they fill in the control panel settings top to bottom.

Effect > RMS Normalize: Target RMS Level -20dB > OK.
Effect > Limiter: Soft Limit, 0, 0, -3.5dB, 10, No > OK.

Try applying those two tools to each song and then you should be able to create a mix tape where all the individual songs are the same (or very close to the same) volume.

If you need slightly higher volume, you can try RMS Normalize to -17dB instead of-20dB and Limiter to -0.5dB instead of -3.5dB.

This is probably no surprise, but as you pump the volume up, some songs are going to get noisy and any distortion is going to come up with the sound. The noise and distortion was always there but now is more obvious in comparison. You might also mix in mono to keep stereo/mono shifting from causing abrupt changes in the sound.

Tracks > Mix > Mix Stereo down to Mono.

If all the work is stereo (two blue waves), you don’t have to worry as much about this.

Koz

[lambdalove]

Thank you very much for your replies and advice! I will try out the suggestions and see how they work for me.

[lambdalove]

Here’s an update on how it’s going:

I first realized that there’s a significant discrepancy in the volume level between tracks of my “mix tape” when I was playing it back and I found myself having to adjust the volume of my player. As I mentioned earlier in this thread, there is the track with the snare drum hit that’s louder than almost everything else in the track (see screenshots in my post above). This track ended up sounding quiet after normalization.

But I realized that wasn’t the only issue: the next track on the mix tape was sounding unusually loud, too!

Here is what the track looked like originally:

After applying RMS Normalize as Koz suggested (Target RMS Level -20dB), it looks like this:

Unsurprisingly, applying the Limiter after that didn’t have any affect.

I applied RMS Normalize and the Limiter to every track on the mix tape, and as far as I can tell it sounds pretty good. I can listen to the whole thing without having to adjust the volume of my player. I can’t hear any impact on the sound beyond the volume levels sounding more consistent.

I’m a complete novice at this, but I do find it interesting that tracks can have such different waveforms. Clearly the track in the screenshots here doesn’t have much variation. Listening to track, I wouldn’t have guessed that the waveform would look like that.

If the waveform is very condensed and uniform like in the screenshots above, should that impact my strategy when it comes to editing with the normalizer and limiter? Would you likely end up tweaking the settings in some way that you otherwise wouldn’t?

[kozikowski]

I can’t hear any impact on the sound beyond the volume levels sounding more consistent

That was the goal, yes. Listen to an audiobook chapter after chapter without having to adjust the volume, no matter how the work was originally read.

I do find it interesting that tracks can have such different waveforms.

Past the obvious differences in performing techniques and venue/studio, music can experience a darker condition called loudness wars. Music marketing and distribution people figured out music is more popular and sells better if it’s louder.

Most transmission and recording systems have absolute limits (radio stations can experience profound technical and legal problems if you try to make them louder), so music can’t actually get louder by making the blue waves bigger, but it can sound louder by making the sound dense. You can crush all the expression and variations out of a song…



So no matter what the performer does, it all comes out the same volume.



That can sound louder without actually violating any waveform conditions.

This can lead to complaints such as: “No matter how I play my guitar at home, it doesn’t sound anything like the music I bought.”

Right. Because the music you bought has had all the life squeezed out of it on the alter of increased sales.

You can get some of those effects in Audacity, but some record labels have a “secret sauce” and we can only guess at it.

But more to the point of your job, there’s no good, easy way to make radically different songs match. We can’t take the compression out of a song, so your only option may be beat up the ordinary music to match the crushed tunes.

In the era of actual mix tapes, the music was generally all taken from vinyl records, so everything more or less matched and it was a matter of pushing Record and dropping the needle. Maybe not any more.

You might also try that Chris Compressor thing. That one is applied to the whole “Tape” at once rather than each song. It will go through and push the loudnesses around to make everything match. I change the first setting from 0.5 to 0.77 to get a denser and better match to broadcast.

As far as I know, the only shortcoming is the two ends of the show. You have to add several seconds of content to both ends so Chris will have something to chew on. It hates running off the end of a show. After Chris is done, you can cut off the extra work.

Good luck.

Koz