#1September 4th, 2005 · 02:50 PM
49 threads / 42 songs
493 posts
United Kingdom
STUDIO TECHNOLOGY : Compression
Quite a few people have asked me the odd question about production/sound recording, which I am quite happy to answer, so I thought it might be helpful to share a bit of the info, particularly for things that used to cause myself some confusion (and probably still do )

For this reason I'd like to start dropping a little basic information around here about studio technology. Many of you will know about this already (or know more) but I think it will be helpful for a few people who are fairly new to recording and looking to expand their knowledge and thus open up more creative opportunities for themselves. Essentially, good studio work is all about conveying your musical vision to the ear of the listener in the best way that you can.

I'm not a producer or a sound engineer by trade, but I am qualified in music technology - although most of what I use is garnered through experience. It's just a shame that my equipment is complete shite - but I'm saving

Anyway, onto the first in a 'Studio Technology' series - Compression. This is a topic of confusion for many people, so this is the best explanation I can come up with...

********************
~COMPRESSION~
********************

Imagine the volume level of whatever you have recorded in decibels. Naturally, this level jumps around a bit as you play.

A COMPRESSOR IS FOR MAKING THE VOLUME LEVELS OF WHATEVER IT IS PROCESSING MORE CONSISTENT It is essentially for 'compressing' the dynamic range, by boosting the lows and reducing the volume 'peaks'. Bear in mind this is not always a good thing. It is easy to overuse compressors and remove all the dynamic expression. Additionally, the above bold statement is only half true - when used in certain ways compressors can actually alter the sound in other, subtle ways.

With a compressor, you can set a 'ceiling' at whatever level you choose to try and limit how much the volume level can obtrude above this 'ceiling'. The 'ceiling' you choose is known as the THRESHOLD and is a level in decibels.

Next comes the ratio. Basically, the compression ratio controls *how much* your recorded signal is allowed to move above the threshold. For instance, a compression ratio of 2:1 means that for every 2db in volume your signal actually goes above the threshold the compressor will only actually let it gain 1db in volume. A compression ratio of 4:1 means that the signal must jump 4db to get one 1db louder in real-estate volume.

After this we have three settings to look at to control how sensitive/responsive the compressor is.

Attack controls how quick the compressor is to cut back on the volume of your signal. This speed is measured in miliseconds. For instance, you may want to set a slightly slower compressor 'attack' (eg 20s) if you want to retain the pick attack of an electric guitar. Similarly to remove this for a smoother sound you would set a lower attack time and the pick attack would get squashed. You can see how a compressor affects things musically as well as keeping volume under control.

Hold and Release (again in ms) control how long the compressor keeps the volume at that level and then lets go. Very low hold and release times may cause a 'throbbing' effect where the volume of the signal seems to pulsate because it is changing too often. Setting these longer can give an instrument more sustain - whereas 'fast' sounds or those with wild transients are best suited to faster release times.

There are more variables - like hard/soft knee compressors. Hard knee simply means that the compressor is more rigid in its volume compensation - whereas soft knee means that the volume level will adjust more smoothly. Hard knee is therefore more relevant to drums etc, whereas strings would benefit better from a soft knee setting.

The last thing is the input/output 'gain' controls. This means that you can adjust the level of your input to bring the quieter parts in line with the volume peaks. It also means that you can increase the volume of the signal leaving the compressor if you need to, to compensate a little for some of the volume reduction going on.

As for a limiter, this is basically a compressor with a more reinforced 'ceiling'. Rather than simply reducing the volume dynamically, the signal finds it very hard or impossible to jump beyond this point. A punchy bass guitar would thus sit well in the mix with a limiter. Imagine a limiter as a compressor with a 'threshold' of 10:1 or greater to get an idea of the volume 'limiting' going on. A true 'brickwall' limiter has a ration of infinity:1 because the volume can never pass the threshold no matter how high the signal level.

Finally, there are 'multiband' compressors. These are best for mastering a final mix. Essentially, a multiband compressor is just several (generall 3 or 4) compressors that each compress a separate frequency range. Thus, you can compress the bass, lower mid, upper mid, and treble separately. For instance, you can make your bass and kick sound punchier and more consistent, and roll off those horribly loud cymbal transients that you recorded, without crushing the dynamic range of the vocals and guitar sitting in the rest of your mix. However, if you have mixed properly and used compression on separate elements within your mix then the need for a multiband compressor should be rare.

There are lots of different kinds of hardware compressors, some costing thousands of dollars, and often having varying tonal qualities. However, most of us cannot afford this kind of gear. Personally, I use plugins because my recording work is PC based. The Waves Renaissance compressor is very good. A great 'budget' mastering package is Izotope Ozone, which includes a Parametric Equalizer, Harmonic Exciter, Multiband Dynamics (compression), Stereo Widening, Mastering Reverb, and Loudness Maximiser all in one package. These will perhaps be covered in additional 'Studio technology' threads - depending on the reception of this one.

Anyway, I havent included any settings for compressors here because you should research those yourself depending on what kind of sounds/music and instruments you are dealing with. Experimentation is also important!

Hope this is of use to some of you

Peace and
#2September 4th, 2005 · 06:45 PM
31 threads / 1 songs
434 posts
United States of America
so, like a good hacker I fully respect your suggestion to RTFM on exact setting specifics... however, maybe you should tell us WFM (which fucking manual) to read

as in, on my Substitutions thread I pointed out where to go to get the data, so... where would one go to get the data on this kind of thing? and this is pretending we all know and understand the TIAS approach already but we're looking for something more...

i.e. where is the research that we can read? or where do we go to get/find/do the research?

thx man - I use compression all the time but the only thing I've ever done is just use the factory presets cuz I didn't know what was going on - this is cool, cuz now I can fiddle with it myself

i would still love it if you knew of some good books or net references on exact setting types and stuff
#3September 4th, 2005 · 09:23 PM
49 threads / 42 songs
493 posts
United Kingdom
No problem. Sure I will post a few examples

My idea was to leave people to use google based on the type of music they are making - because their are so many variables. Ideal compressor settings vary even with the compressor itself because every compressor reacts differently.

A punky electric guitar with some hefty palm muting would perhaps do well with an attack of 10ms to keep the chugging pick attack and a release of 50-250ms, as well as a fairly high ratio of 4:1 of maybe even higher to keep it squashed. In comparison, a slow and expressive classical guitar piece with alot of natural dynamics and a smooth sound may want a lower attack of 5ms with a lower ratio of say 2:1 on the guitar part to retain the original dynamics. Similarly, a synth line that tends to clip at higher frequencies you may want to give a longer release time for sustain (eg 500ms) and adjust the ratio accordingly. Vocals would perhaps want a faster attack time (eg 5ms) with a 300-500ms release and a ratio of 4:1 depending on the dynamic range (ie mic control) of the vocalist (soft knee setting if you have the option). For dance (eg Trance) vocals that want to sit in a mix with very little dynamic range try 8:1.

Drums I won't get into too much here because their are so many different kinds of drums/music. For kick/snare, think low (eg 1-5ms) attack times, and fairly fast (eg 200ms) release times with a high (5-10:1) ratio and a hard knee setting.

A good rule of thumb is to start the compressor threshold 6db below the volume peaks and adjust accordingly depending on the level of compression you want.

For some basic compressor settings for 'live' instruments I would recommend looking here for a start (bottom of the page):

http://www.soundonsound.com/sos/1996_articles/apr96/compression.html

For a final mix, you are going to want to master using a ratio of 2:1 as a starting point, again starting about 6db below the peak level. To prevent a chugging effect you will want to set a longer release time (200-300ms) and keep the attack around 20-30ms. If you have alot of drum/cymbal transients then you may want a faster attack time. For music such as death metal you may want a higher ratio (eg 4:1) to squash it further.

Bear in mind that if the compressor threshold is below the kick peak level then the whole mix will be squashed when there is a kick! Good or bad depending on the music.

In short : when recording, EQ first and compress afterwards on a particular instrument/element. When mastering, get the dynamics and levels as right as possible FIRST at the mixdown stage and then look at compressing the mix. Mastering can fix a poor mix to an extent - but I am sure we can all agree that it is better to polish a rough diamond than polish a turd

Anyway, for more advanced compressor information do a search on the above site under the technique section or try google.

And one thing - big thing - a compressor is a 'processor', not an 'effect'. A 'processor' such as an EQ unit or compressor affects the entire signal, and is not supposed to be mixed in 'wet' with the 'dry' signal in the same way that you would with an 'effect' such as a reverb unit. Always make sure that your compressor is 'inline' ie using the 'input' perspective of the mixer rather than aux/effects sends or whatever.
#4September 6th, 2005 · 05:45 AM
115 threads / 18 songs
1,415 posts
United States of America
i think we've found another possible sticky?
#5September 6th, 2005 · 06:45 AM
31 threads / 1 songs
434 posts
United States of America
totally sticky worthy
#6September 6th, 2005 · 07:32 AM
8 threads / 4 songs
246 posts
United Kingdom
thats brilliant Dave ... ive spent a large amount of my time trying to get to grips with the mystery of compressors, and you've just gone and nailed it for me!

one final question - if you have a song which has a very quiet section and a very loud section, is there anyway you can compress it so that the quiet section doesnt vanish completely?!  for example my song "my turn" has a very quiet verse but a loud bridge, so the begining is barely audiable.  is there anything i can do about this (apart from the obvious 'play louder')?
#7September 6th, 2005 · 08:08 AM
31 threads / 1 songs
434 posts
United States of America
yeah, well you can use either your initial host application or a wave editor to just manually turn up the volume on those parts... that's probably what I'd do... I've got the same problem because of the volume at which I sing low notes compared to the volumen at which i sing mid to high range notes... i just manually went through and split the wave file at the appropriate junctures and then increased the volume of the soft parts - i'll admit that the way I went about it ended up sounding choppy but when combined with the rest of the parts you can't/don't notice it and also if you've got totally quiet verses, there (theoretically) should be at least a split second of absolute silence before your chorus, so just be sure to chop it there and you're straight
#8September 6th, 2005 · 11:24 AM
49 threads / 42 songs
493 posts
United Kingdom
randomdave wrote…
thats brilliant Dave ... ive spent a large amount of my time trying to get to grips with the mystery of compressors

I still am hehe...

As for balancing the levels of different passages, there are multiple options I can think of mate...

As entheon said, manually using a wave editor to increase the gain at the quieter periods. In general, though, artificially adding gain like this is never great (professionally you would be recording at -something db anyway, so an engineer sliding the fader up would not be a problem). Ideally, silly as it may sound, it best to get the dynamics closer to the final sound you want (ie if it was played live) when you record it, than do it later.

However, the standard way of doing this at the mastering stage is using a loudness maximiser. Essentially, this is used at the very final link in the audio chain (unless you are dithering, although several loudness maximisers will do both jobs).

I will get into all this stuff at another thread, but essentially imagine that you are mastering to a peak level of 0db (fairly loud). This gives the in-your-face sound of a professional CD, and also means that your tracks sound consistent on the distribution medium (much better than 'normalising' them using the shitty option in your CD burning software).

A loudness maximiser (such as the fantastic Waves L2 UltraMaximiser) is similar to a compressor in that it boosts the level of the whole mix and then clips the peaks. However, you will find that their is no 'ratio' control - the peak limitation is exactly that, a 'limiter', and thus the peaks are controlled more agressively.

However, more advanced loudness maximisers such as the one above feature ARC which (like an automatic compressor) means that the plugin will automatically adjust its release time to better suit a full mix.

So basically in my masters in Wavelab (not all of the below are always used):

Steinberg DeNoiser ->
Waves 10-band Parametric EQ or Renaissance EQ ->
Waves Maxxbass ->
iZotope Ozone 3 Harmonic Exciter ->
Ozone 3 Multiband Stereo Widening ->
Waves Renaissance Compressor ->
Waves Renaissance Reverb ->
Waves S1 Shuffler (Stereo Imaging) ->

And finally in the 'Dithering' section goes 'Waves L2 Ultramaximiser'

NB: The UltraMaximiser also allows you to 'dither'. This is essential for retaining quality and preventing audio artifacts when converting from 24-bit recording to 16-bit audio for CD.

The Waves plugin package is at least 2000, so if you dont have access to this then just find free alternatives and apply the same principles. Izotope Ozone is much cheaper and is a very good 'all-in-one' mastering solution for the money.

The bottom line to answer your question is - *any* form of compression or limiting sitting over your mix is reducing the dynamic range of the whole mix. The art is deciding how best to do it! You will find that if there is too big a difference in volume between quiet and loud passages, if you try to compensate for this during mastering simply using compression/limiting then you will severely overcompress your mix and crush all dynamic life out of it. Thus, I suggest you edit some more gain into quieter sections by hand, and use a loudness maximiser - and just try and record the dynamics closer to a 'live' performance next time.

Hope this is of use
#9September 6th, 2005 · 02:12 PM
13 threads / 1 songs
408 posts
United States of America
Well, Dave, you are the man. This might be completely off the subject (and I'm hoping you'll do some more tutorials) but I was hoping you could listen to some Suzanne Vega for me and give me some pointers as to what she did, if one can hear what's going on behind the scenes, I'm sure you could.
There is something about her style that I like and I can't fully say it's her voice or music.
I. E. In her song Headshots, what is she doing with vocals during the lines "In what we see...That's history". I know she's panning, but that's all I can put my finger on. And her voice is very crisp. Know what I mean? Or if you know how I can acheive an effect like hers, I'd be surely greatful.
Thanks.
#10September 8th, 2005 · 07:03 AM
49 threads / 42 songs
493 posts
United Kingdom
I'm not sure of the song, cant get hold of it to download.

Common effects/processors for vocals would be (not necessarily all)

EQ
De-Esser (High-frequency limiter to remove percussive vocal sounds)
Compression
Harmonic Exciter (be careful with this one, but can sound really good)
Reverb

A common trick for 'depth' is to use double-tracking ie record the same vocal part twice and pan them both slightly apart of stereo 'centre'. A useful effect is known as ADT (automatic double tracking) whereby the vocal part is automatically doubled with a controllable delay (generally the effect is most subtle with a delay of 1-5ms). You can then also control the width between the two vocal parts.

The 'crispness' you describe is probably from careful EQ work, compression and a harmonic exciter for adding 'sparkle'.
#11September 8th, 2005 · 07:16 AM
13 threads / 1 songs
408 posts
United States of America
A harmonic exciter. interesting, thanks, I'll look into it.
#12September 9th, 2005 · 11:47 AM
8 threads / 4 songs
246 posts
United Kingdom
thanks dave - you're a top bloke.  i wait for further articles with baited breath!
#13September 2nd, 2007 · 08:31 PM
49 threads / 42 songs
493 posts
United Kingdom
*bumpy mcbump*
#14September 5th, 2007 · 04:43 PM
17 threads / 3 songs
185 posts
United States of America
Hmm  I  get confused with compression . I was told to compress at the preamp/mixer (insert) stage very lightly
so you can get the signal levels (hope I remember this right) peaks or spikes or whatever you want to call it,  some what leveled out here and adjusted so you can gain up some.. Like you said earlier this will depend on the input signal as to what setting you would use
ie my vocals and ac guitar ..I set about  -10 db and about a 1.8or 2.0 :1 ratio  attack and release also depends on tempo and style of music ... at this setting most of the time the compressor db indicator shows no compression going on but just in certain spots and only at the most maybe a 2 or 3 db reduction from the  loudest peak

I  can compress more later on in the software realm if I need to  but I tend to try to use the least amount of compression that I can get by with (personal taste for me) cause I like more open sound and dynamics

my biggest thing is percussion settings and bass guitar settings I don't know what to use if I squash it,  seems like some of the upper harmonics go away in the bass and snap of the snare and kick

I know it takes more experimenting but do you know of a good staring point for lets say bass
and just  straight ahead rock drums?

help?
greg
Sorry, you do not have access to post...
Wanna post? Join Today!

Server Time: August 13th, 2022 · 5:58 PM
© 2002-2012 BandAMP. All Rights Reserved.