Discuss & Share your Audio Production, Mastering, Post Production Techniques here.

Production Tips - Round 3: Use Compression

Production Tips - Round 3: Use Compression

Postby dilvie » Fri May 19, 2006 1:27 pm

It looks like some of you know that applying compression to the main channel can bring more coherence to your mix and maximize the output volume, but a lot of you seem to be missing out on a good opportunity to better control the dynamic range occupied by individual channels.

Using compression let's you give each sound it's own volume range -- keeping it loud enough to be heard over background-elements, while keeping the maximum volume in check at the same time.

Compression can also be used to draw out reverb tails on percussion and bass, or adjust the attack portion of a sound.

Experiment with it on your individual channels. I think you'll find it's worth the effort.

- Eric
Posts: 86
Joined: Tue May 16, 2006 4:40 am

Postby Sonicade » Sun May 21, 2006 10:27 am

Good insight on how to apply compression. Thanks for posting. :)
User avatar
Compo Admin
Compo Admin
Posts: 773
Joined: Fri Mar 17, 2006 6:42 pm
Location: California, USA

Postby groovyone » Sun May 21, 2006 3:13 pm

Compression is probably the most tricky of all sound engineering skills to master.

I've noticed a lot of you tend to squash your sound with a fast attack.

Here is the secret of compression:

Say you have your kick, and you want to bring the kick up a bit so you can hear the resonance. The kick bite sits at 0db as it's been normalized.

Threshold: - At what level does the sound start compressing.

All sound above this threshold will have compression applied. This controls at which point volume squashing will occur.

Ratio: - How much compression is applied.

All sound above the threshold will be applied a compression ratio. If you have a ratio of 1:2 all sound above the threshold will be compressed 2x.

Attack: How fast will it compress
- Depending on the sound, the attack will make or break the sound. Take for example a kick drum or snare, if you have a fast attack, you'll kill the initial transient.

The transient is the movement between a 0 point (silence) to the maximum volume. This is your bite. If you kill the bite your kick or snare will become lifeless. What you want to try to do is give enough time for the transient to still have bite but then pull the ring of the snare up the snare and then release it at the right time.

Release: How long to slowly release compression.

Compression occurs above the threshold after the intial attack time. Once the sound drops back below the threshold, automatic compression ends and the release is applied. The release will shape the trail of the sound.

Gain: - What you will need to make up to bring the sound back to normalisation.

Once compression is applied you lose volume, you've squashed the sound down. Gain will replace the loss in volume by bringing the sound back up to the normalised amount. Use your VU meters to watch this and adjust.

The trick to compression on individual sound is first you want it to shape the sound. Think of compression working with tempo. Play back an individual channel - say a kick or a snare, and get it so that the compression is dancing with the sound - ie it comes on but releases in time for the next sound. If it's still compressing by the time the next sound is playing then you'll lose the transient again.

Another trick of compression is compressing instruments together.
ie Kick and Bass.

Buss the Kick channel and the bass channel to a separate channel. Then apply compression over the combination. This will 'gel' the sound together. You'll also want to EQ each channel so that they work together and don't just chew up the same frequencies too much.

Sound is frequencies. a Kick may be between 80-300Hz and a bass could be between 100-400Hz. Which instrument do you want to have sub? Is it the bass, or the kick (this is up to you). If you choose to have the kick to have sub Roll off EQ on the bass so that it drops down below say 200Hz and also roll off some of the kick above 200hz. This doesn't have to be a brick wall EQ, it can be slight, but enough so that the two sounds fit better as a Jigsaw.

You may ask, but won't you lose the sub of the bass? Well, yes, but when combined with the sub of the kick that power is put back into the bass - sound adds together, never acts alone.

I'll explain EQ if it's not covered already by Dilvie in the EQ thread.

Mastering Compression - Compressing the whole song.
Why do we compress the master bus? Mixing the sounds together gets your EQ balance nicely, but there are still a lot of individual transients in the music, you may notice the "bulk" of the volume is under -6db and a lot of the transients are between -6 to 0db. Well, compressing will bring the entire mix up a bit. You don't want to over do it here, subtlty is the key.

Think of these two scenarios:

1. Heavy Compression at %80
Threshold: %80
Ratio: 1:5
Attack : 1ms
Release: 80ms

2. Subtle Compression at %40
Threshold: %80
Ratio: 1:1.5
Attack : 1ms
Release: 80ms

Example 1 will squash down all the transients above 80% very heavily 5x whatever the volume is above the threshold.

This may be what you're looking for but your mix will start to sound flat and lack dynamics. It's good for things like rock I guess.

Example 2 will start compression much lower but only subtly compress. It will bring all the sound up from %20 of max voume at 1.5x the amount. This method will give you back volume and bring up the reverb, the lower volume sounds. The song will still have dynamic range.

Don't over compress, your song will sound loud, but it will lack "rhythmic dynamics"
Posts: 63
Joined: Mon Apr 17, 2006 4:49 am
Location: US of A

Postby dilvie » Sun May 21, 2006 4:59 pm

Great tips, Groovyone.

Couple of examples:

Say you've got a ride cymbal that's competing too much with other instrumentation -- it's there primarily for the transient rhythm, and you don't really need to hear so much body. Of course, if you have control over the volume envelope, you can keep the attack at 100% and apply a steep slope to the right level for the tail fade, but if the sound is coming from a VSTi, or a multi-sample set, that may not be so trivial. You can do it with compression, though.

Apply at least a 3:1 ratio, set an attack between 1-4 ms (too long, and you'll de-emphasize the attack portion), and set a fairly long release (50ms+).

You might also want to shape it with some EQ -- drop the frequencies that are competing, but keep the frequencies you like in the attack portion. A drop of 6db generally sounds about half as loud. Use that as a guide.

I'm a big house music buff, and I often want to have some action on the off-beat, but I don't necessarily want that to be overly percussive. I often use compression to intentionally squash the attack portion of an open hihat:

I'll apply a strong ratio (5:1 - 8:1), no attack (so it compresses immediately), and use my ears to set an appropriate release (which will vary depending on the source sound -- generally 30-60ms).

Yes, I did say that I intentionally squash transients. Don't do that to all your percussion sounds, or you won't have any percussive rhythm! I do this to select sounds to achieve a particular effect -- it's okay to break some rules as long as you're using your ears, and you know what you're doing.

It's also worth noting that compression is just a glorified gainer if the sound never crosses the threshold. Apropriate threshold levels depend entirely on the incoming signal strength. Use your ears and the compressor visual feedback as a guide.

- Eric
Posts: 86
Joined: Tue May 16, 2006 4:40 am

Postby Harmony » Fri May 26, 2006 4:11 am

A very useful writeup, thank you both!!
User avatar
Posts: 68
Joined: Thu Apr 06, 2006 10:43 pm
Location: Newcastle, Australia

Return to Production / Mastering

Who is online

Users browsing this forum: No registered users and 1 guest