Compression

  • Thread starter Thread starter Orchestrator1
  • Start date Start date
The frequencies haven't changed, but the balance between dominant and harmonic has changed. Am I on the right track?
Yeah, you have the basic idea. There even a bit more to it in that not only has the balance between them changed, but the relative amplitudes of the individual harmonics change as well.

Also, there are other things happening below the threshold than just harmonics of a dominant frequency, there are also other sounds that are independent sounds of their own, but just at lower relative volumes.

But I think you have the main idea basically correct; squeezing the dynamics of an instrument tends to change the tonal balance of that instrument by tending to "even out" (to a degree) the relative volumes of all the different frequencies that combine to make the sound of that instrument.

G.
 
Last edited:
but in a practical sense the harmonic frequencies will end up amplified.

I might add, technically some of those harmonics could be lowered, they don't necessarily have to be amplified.. Via Fourier analysis, any repeating wave can be reproduced by a combination of pure sin waves, ie the fundamental + all harmonics. If you fill out a waveform so it has more area under the curve, SOME of those harmonics have definately been amplified, but other less significant harmonics have probably been diminished.
 
Most of the discussion here is about broadband compression applied to a monophonic instrument, and is perhaps related also to glueing factor in a mixdown. But if, in fact, the OP was referring to a polyphonic solo instrument - like piano, classical guitar, or even something like solo accordion, it's a whole different situation. In that case, something like the effect of compression on overtones is minute compared to the effect of the compression triggered by strong low pitches on the tone color of higher pitches played at the same time.

So, if that's the case, Orchestrator1, you should look into split-band compression also, where the dynamic control can be focused on the lows and low-mids and the higher freq range can be left alone. That can be very transparent re avoiding changes in instrumental timbre.
 
Most of the discussion here is about broadband compression applied to a monophonic instrument
Broadband compression, yes, monophonic instruments, no. The number of notes that can be played at one time has nothing to do with it. Whether one is playing a middle C or a Cmaj chord, compression still will modify the resulting waveform in much the same way.

As far as the more acoustic instruments, yeah there is a bit of difference in degree, but by a higher degree instead of a less one. The more complex the waveform an instrument creates - the more harmonics and dynamics that go into the timbre of the instrument - the greater the potential for a tonal change when applying broadband compression to it.

G.
 
...The number of notes that can be played at one time has nothing to do with it...
No, it has a lot to do with it.

Not in the sense of a "C major chord" as you describe, though. That's not polyphony (except in the midi programming sense), that's a single chord. Multiple continuous voices is polyphony. Try this: take an acoustic piano or classical guitar track that has slightly boomy low end, where the performer is playing a bass line and a separate upper line at the same time. A 2 part Bach piece, for instance. That boomy low end is a common problem with home recordings of acoustic instruments. Now hit it with a broadband comp and, whatever the attack/release settings, you'll hear the compression's undesirable effect on the upper voice every time gain reduction kicks in triggered by the intensity of the lower notes. The upper voice becomes unnatural. Now hit it with a split band comp, very carefully set, and use a high quality comp with high quality filters. Compress the lower range with slow attack and release, and leave the upper range alone. Compare the two results.
 
No, it has a lot to do with it.

Not in the sense of a "C major chord" as you describe, though. That's not polyphony (except in the midi programming sense), that's a single chord. Multiple continuous voices is polyphony. Try this: take an acoustic piano or classical guitar track that has slightly boomy low end, where the performer is playing a bass line and a separate upper line at the same time. A 2 part Bach piece, for instance. That boomy low end is a common problem with home recordings of acoustic instruments. Now hit it with a broadband comp and, whatever the attack/release settings, you'll hear the compression's undesirable effect on the upper voice every time gain reduction kicks in triggered by the intensity of the lower notes. The upper voice becomes unnatural. Now hit it with a split band comp, very carefully set, and use a high quality comp with high quality filters. Compress the lower range with slow attack and release, and leave the upper range alone. Compare the two results.
Well, I agree with you more than I disagree. Let's get the technical detail out of the way first: The ability to play more than one note in a chord IS polyphony. The number of voices is not. Take a look at a MiniMoog or an ARP 2600 synth, for classic examples. Those are multi-voice, monophonic instruments; you can assign multiple voices simultaneously but can only play one note with those multiple voices at a time. The multiple voices only serve to complicate the timbre of the sound created when that note is played.

But more important, I think, is that the situation you are describing is exactly the kind of effect we're talking about; that compression does change the tonal characteristic of the track. The fact that there are two instruments - two voices - contributing to the resulting waveform is academic from the viewpoint of the compressor. It's just processing voltages and changing volume levels based on a limited number of control parameters. The fact that it's a more complex waveform/timbre does potentially increase the audibility of those changes - as I mentioned in the last post - but multi-voicing in and of itself is not a requirement for the tonal change to happen, readily audible or not.

Using a bandwidh-limited compressor will give you more surgical control over the signal, and, sure, when used properly will keep unwanted artifacting to a minimum. No argument there. The fact remains, though, that just by simply using any compression whatsoever, one is altering the volumes of various frequencies. That is what a compressor does, broadband or narrow band. By it's very definition, that specifically describes altering the tone of the signal.

It's no different in that high concept than a frequency equalizer. Sure, an EQ and a compressor do their jobs quite differently, but at the base of it all, all they are both doing is changing the amplitudes of various frequencies within the waveform. To say that a broadband compressor changes the tonal balance of an instrument but a narrow-band one does not - or that the exact nature of the source of the waveform changes the answer much - is tantamount to saying that a tone control or a graphic EQ changes the sound of something but parametric EQ does not, or that whether they change the sound or not depends upon what the source of the sound is.

G.
 
...The ability to play more than one note in a chord IS polyphony. The number of voices is not...
Incorrect, except as I said, in the sense of programming artificial instruments playing multiple pitches simultaneously rather than actual musical polyphony as the term is used in the realm of Western European art music.

I'm guessing from your responses that you haven't worked much with recording polyphonic acoustic instruments like solo piano or classical guitar. If so, anywhere I can hear your work?
 
Incorrect, except as I said, in the sense of programming artificial instrument
An analog synth (not MIDI) is no more "artificial" than a Steinway grand piano. They are both instruments for creating sound. And as the analog synth shows, it is entirely possible to have multiple voices played on a non-polyphonic-capable keyboard.

Maybe we're just arguing shifting definitions of specific words here, in which case we're both just spinning our wheels. I grew up in an environment where polyphonic meant being able to use more than one finger to play the instrument. But then again I grew up in the 1970s, not the 1670s.

Educate me: on a modern piano, when you strike a single key, the hammer is striking 2 to 3 separate strings. By the terminology I grew up on, that is technically 2 to 3 different simultaneous voices. Does that constitute a polyphonic sound by the definition you're used to using? If so, then what do you need to bring the extra instrument into the equation for? And if not, then that is proof right there that poly-voice is not the same as polyphonic.

But the definition of polyphonic is not important to the subject at hand. We're arguing words there. But regardless of the words used, it nevertheless remains that any change in amplitude by frequency is the very definition of a tonal change. It's a simple as that.

Sure, it can be argued that the human audibility of such changes varies depending upon the source and by the amount and type of compression applied. But is that what we are using as our definition of "tonal change"? If so, who's ears are used to make that definition? What percentage of change in the spectral fingerprint is not enough to be a "tonal change", but another 1% added on is? Give us that specific value and then maybe we'll have common ground to work with.

G.
 
Alright, I did some research, including talking to a musician friend of mine who eats and beathes and sleeps music theory, and I understand now that there are several different uses or definitions for the term "polyphonic", and that the one I wrongly assumed you meant when you first brought it up was different than what you actually meant. My mistake and apology for assuming and for not being aware of the other definitions/uses.

The original point stating that any use of compression causes at least some frquency-dependent change of amplitude to all or part of a signal remains.

G.
 
A 2 part Bach piece, for instance. That boomy low end is a common problem with home recordings of acoustic instruments. Now hit it with a broadband comp and, whatever the attack/release settings, you'll hear the compression's undesirable effect on the upper voice every time gain reduction kicks in triggered by the intensity of the lower notes. The upper voice becomes unnatural. Now hit it with a split band comp, very carefully set, and use a high quality comp with high quality filters. Compress the lower range with slow attack and release, and leave the upper range alone. Compare the two results.

That would be an entirely different way to use a comp from what I was describing. I was not referring to the use of a comp to "fix" a track. MB compression can be great for that, you are absolutely right.

Also, just because it brings out the harmonics, doesn't mean that it's a good thing - as SG said above........
 
That would be an entirely different way to use a comp from what I was describing...
True. My original points were that the OP asked about compression's effect on the tone quality of an instrument without naming the instrument, and the general discussion here was making some assumptions than might not apply if in fact he was asking about acoustic instruments playing polyphonic repertoire.
 
My Apologies to all, I should have told you the instrument was an accordion.
I purchased a Pro VLA twin chn tube leveling amp, and and after a lot of trial and error not too sure about the end result. I normally play in musette which is two sets of reeds, one tuned 15 cents positive and the other tuned 15 neg so lots of harmonics. I am perhaps getting the feeling its not much more than a pre amp.
Rex.
 
My Apologies to all, I should have told you the instrument was an accordion.
I purchased a Pro VLA twin chn tube leveling amp, and and after a lot of trial and error not too sure about the end result. I normally play in musette which is two sets of reeds, one tuned 15 cents positive and the other tuned 15 neg so lots of harmonics. I am perhaps getting the feeling its not much more than a pre amp.
Rex.
The question is what are you trying to achieve with the VLA? (I own and use a VLA myself, BTW.) For what purpose are you using the compression? Are you doing it for dynamics control or are you purposely trying to change the sound (i.e. trying to get "that toob sound"?)

The VLA does use fairly typical 12AX7 tubes in the preamp stage of each channel, but the compression itself itself is actually based upon an optical compressor circuit, not tube. The act of compression itself on that device will impart some - but not much - tube-ish coloration unless you overdrive the preamps themselves, and even then it won't necessarily sound all that positive - depending on just what you're looking for.

G.
 
...the instrument was an accordion.
I purchased a Pro VLA twin chn tube leveling amp, and and after a lot of trial and error not too sure about the end result...
If you want dynamics control via a hardware unit that's inexpensive, easy to use and doesn't squash the highs like most broadband comps tend to, try the FMR RNC (Really Nice Compressor) in "supernice" mode. Very good on acoustic instruments.
 
Hi SouthSide Glen
I am experimenting with it to try and get a better sound because I am never satisfied with what I have. I use an MPA Gold preamp and am very happy with it but as I said not sure about the PRO VLA compressor.
Rex
 
Gentlemen, this has been Xtremely fascinating reading. Compression is a huge issue 4 me on the 'Need to understand it better end'. Instead of participating in the mathematical equations of how Compression works, although I'm sure I'll need to get that down someday. Can I ask this question from a Lyman's viewpoint?

Using Computer software i.e. SX3 & Ozone with built in processors, btw this statement
SouthSide Glen said:
I'm using the term distortion in it's pure sense: any difference in the output from the input - any change in the signal - is a distortion of the original signal. The only question is to whether any given difference is intended or incidental. The coloration caused by circuit design is a form of incidental distortion; pushing that circuit out of equilibrium is a form of intended distortion.
truly hit home and although seemingly what we all know to be true, is the foundation of what we need to remember when we try to add or subtract from what our system recorded. Thanks SSG

Anyway, if I get what I feel is a nice clean recording of your basic instrumentations i.e. Drums - Bass - Keys - R. Guitar - Solo Guitar - Vocals.

How can you tell if you need compression or not when there are so many variables in how you monitor your Mix? I read somewhere that compression should be applied in small measure to the entire mix rather than individual tracks. But I can't get my head around that concept because what if I think an instrument(s) sound good the way it already is?

Lastly, some of the very best recordings I've ever gotten where with a sequencing software going straight out to record during playback. The spread was excellent. Every instrument was distinguishable from the next and now that I have analog recording capabilities, I often get a muddy feel from the rhythm instruments if/when I apply compression and/or reverb across the board.

Sorry if I interrupted the flow of this conversation ..
 
Hi SouthSide Glen
I am experimenting with it to try and get a better sound because I am never satisfied with what I have. I use an MPA Gold preamp and am very happy with it but as I said not sure about the PRO VLA compressor.
Rex
Again, when you say you are trying to get a "better sound", what is your reason for choosing a compressor as your tool of choice to accomplish that?

What is the problem you're trying to solve? What doesn't sound good enough with the tracks you're capturing now?

There's no reason IMHO why your current ART gear should keep you from getting a good sound, but if you're looking for it to create a good sound for you, I'd say that you might want to look elsewhere first, such as the start of the chain. Mainly, what mic are you using, how are you miking the accordian, and what are the acoustic properties of the room in which you're recording. Getting that right is priority #1, everything else after that, while important, is secondary.

Also, are you feeding it more tradional Eastern European polka/waltz type stuff or more in a Cajun zydeco or British skipple vein?
Gentlemen, this has been Xtremely fascinating reading. Compression is a huge issue 4 me on the 'Need to understand it better end'. Instead of participating in the mathematical equations of how Compression works, although I'm sure I'll need to get that down someday. Can I ask this question from a Lyman's viewpoint?
...
How can you tell if you need compression or not when there are so many variables in how you monitor your Mix? I read somewhere that compression should be applied in small measure to the entire mix rather than individual tracks. But I can't get my head around that concept because what if I think an instrument(s) sound good the way it already is?

Lastly, some of the very best recordings I've ever gotten where with a sequencing software going straight out to record during playback. The spread was excellent. Every instrument was distinguishable from the next and now that I have analog recording capabilities, I often get a muddy feel from the rhythm instruments if/when I apply compression and/or reverb across the board.

Sorry if I interrupted the flow of this conversation ..
No problem, we can multi-task; the flow has already swirled and eddied a couple of times anyway ;) :D.

Blu, maybe you might want to start with a basic tutorial on compression. If you haven't already done so, skipple on over to www.independentrecording.net and click on the "Compression Uncompressed" logo in the lower right corner. That should do a fair job of covering the basics of why, when and how to use basic compression fairly well.

G.
 
Back
Top