Compressor plugin theory question

[Metaltastic]

So all this talk of different compressor plugins recently got me thinking - is it really that hard to make a plugin that just functions as a transparent automatic volume control? Obviously the plugins that are designed to impart coloration are another story, but I guess I just don't see why any compressor plugin couldn't do its job (controlling dynamics), without imparting any other qualities on the tone (because of the invariable nature of digital audio and DSP regardless of level). To my ears I feel Reaper's stock compressor (ReaComp) does just this, which is why I'm happy with it, but maybe someone can shed some light on the subject?

Also, on a side note, what's the difference between a compressor and a leveling amplifier?

 

[punkrockacademyfightsong]

I've noticed that Gcomp does this job really good. I agree that Reacomp can do this job really well too, but does take more tweaking than Gcomp.

 

[26]

is it really that hard to make a plugin that just functions as a transparent automatic volume control?

Assumedly not exactly what you meant, but stumbled over this one lately. Looked somehow worth trying to me.

http://www.scrollworks.com/products/slammer/

 

[ScootBaloo]

So all this talk of different compressor plugins recently got me thinking - is it really that hard to make a plugin that just functions as a transparent automatic volume control? Obviously the plugins that are designed to impart coloration are another story, but I guess I just don't see why any compressor plugin couldn't do its job (controlling dynamics), without imparting any other qualities on the tone (because of the invariable nature of digital audio and DSP regardless of level). To my ears I feel Reaper's stock compressor (ReaComp) does just this, which is why I'm happy with it, but maybe someone can shed some light on the subject?

Also, on a side note, what's the difference between a compressor and a leveling amplifier?

I think that it would not be that hard to make a plug in that functions as a
automatic volume control. The questions is, how transparent would it be ?
When you apply amplitude modulation to a signal ( similar concept to gain reduction )
you change the spectrum of the signal, or to put in other words, you add
distortion to the signal. There is no such thing as a "transparent" automatic
gain control.

 

[mick thompson]

I think that it would not be that hard to make a plug in that functions as a
automatic volume control. The questions is, how transparent would it be ?
When you apply amplitude modulation to a signal ( similar concept to gain reduction )
you change the spectrum of the signal, or to put in other words, you add
distortion to the signal. There is no such thing as a "transparent" automatic
gain control.

i would imagine in a digital environment this wouldn't happen (how transparent would it be) - after all when going analogue different circuit components would add specific tonalities to spectrum but when it comes to digital processing your using code to tell it effect the gain levels and do not eq the spectrum. code cannot have a tonality unless its programmed in to have one

 

[Erkan]

Assumedly not exactly what you meant, but stumbled over this one lately. Looked somehow worth trying to me.

http://www.scrollworks.com/products/slammer/

The waveform images on the webpage look like they've been GClipped to fuck But GClip doesn't induce any latency, and it doesn't "scale" the peaks, it simply cuts them off... so I guess Peak Slammer would sound a lot different to GClip and any other clipper since it has latency and is a lot smarter than any clipper. I might give that free trial a go when it's time for mixing and stuff, thanks for the link.

 

[TheDriller]

The way i understand it is that compression always alters the amplitude envelope of the sound, and as such can never be truely transparent.
even if you design a plugin that has no "colour" of its own, the very fact that it is compressing the signal will change the "shape" of the audio, and thus change the sound.

I think this image illustrates what i'm saying,

if you want transparent level control, then just use volume automation.

 

[Erlend_]

A comp plugin that writes automation data.....

 

[greyskull]

there is this....
http://www.automaticmixing.com/

 

[ScootBaloo]

i would imagine in a digital environment this wouldn't happen (how transparent would it be) - after all when going analogue different circuit components would add specific tonalities to spectrum but when it comes to digital processing your using code to tell it effect the gain levels and do not eq the spectrum. code cannot have a tonality unless its programmed in to have one

The distortion that I was referring to was not due to
imperfections in the analogue components but a theoretical
distortion that always occurs when amplitude modulation is
applied to a signal ( in both the analog and digital domains ).

Above you argue that the spectrum isn't altered when gain
reduction is applied but it is altered and it depends on the
frequency and amount of the gain reduction.

A simplified model of amplitude modulation might be

out( t ) = in( t ) * ( k1 + k2*sin( k3*t ) )
= k1*in( t ) + k2*sin( k3*t )*in( t )

where :

in( t ) : The signal going into the compressor.
out( t ) : The signal coming out of the compressor.

It is the second term in the above equation ( k2*sin( k3*t )*in( t ) )
that is the cause of the spectrum change.

 

[Metaltastic]

Interesting, thanks for the input Scott! Naturally I'll have to take your word for it, since that formula is completely over my head, but I guess that is what I was asking for Naturally if it's desirable coloration I certainly wouldn't mind it in a comp. plug, my question mainly was just trying to get an idea about how capable even the most basic of compressors (e.g. ReaComp) could be, but I guess I'll just have to demo some others and see how they all stack up!

 

[Metaltastic]

The waveform images on the webpage look like they've been GClipped to fuck But GClip doesn't induce any latency, and it doesn't "scale" the peaks, it simply cuts them off... so I guess Peak Slammer would sound a lot different to GClip and any other clipper since it has latency and is a lot smarter than any clipper. I might give that free trial a go when it's time for mixing and stuff, thanks for the link.

Yeah, +1 to that, I kinda have my doubts if the only tweakability is a threshold slider, but set subtly it could be useful maybe for the last stage of dynamics control! (catching the biggest peaks)

 

[26]

Also found this.

http://www.ethanwiner.com/peakslammer.html

 

[mick thompson]

The distortion that I was referring to was not due to
imperfections in the analogue components but a theoretical
distortion that always occurs when amplitude modulation is
applied to a signal ( in both the analog and digital domains ).

Above you argue that the spectrum isn't altered when gain
reduction is applied but it is altered and it depends on the
frequency and amount of the gain reduction.

A simplified model of amplitude modulation might be

out( t ) = in( t ) * ( k1 + k2*sin( k3*t ) )
= k1*in( t ) + k2*sin( k3*t )*in( t )

where :

in( t ) : The signal going into the compressor.
out( t ) : The signal coming out of the compressor.

It is the second term in the above equation ( k2*sin( k3*t )*in( t ) )
that is the cause of the spectrum change.

hmm obviously you know much more about this then me - i was speaking more on analogue terms for colorization on the channel. i presumed when you digitalized the track that gain reduction etc would be colorless due to its digital nature not being effected by circuit components. i suppose im never going to fully know / understand that equation above unless i learn how to develop a plugin to do it. then i could monitor its effect on the channel in relation to colorization of the track and see its change in the physical world in relation to the data input

thanks of the post - very interesting - i just wish i didn't sleep in math class

 

[TheDriller]

i presumed when you digitalized the track that gain reduction etc would be colorless due to its digital nature not being effected by circuit components.

see, the very fact that you are compressing the signal at all (digital or otherwise) is going to change the shape of the signal and "colour" it.
its not just to do with analogue componants introducing distortion, the compression process itself is a distortion. (see the diagram i posted above)

essentially, you can't have your cake and eat it if you want to change the signal, its going to change.
there is no such thing as transparent processing, because if it was transparent, it wouldn't be processing at all.

 

[Metaltastic]

I guess then the only question is, do different compressor plugins have vastly different means of either minimizing this distortion or making it more pleasing to the ears? (at least enough to cause sizeable audible differences). But of course, there's one way to find out! (I guess I asked mostly because I didn't want to go to the effort of demo'ing different compressors and trying to objectively compare the differences - but since the math is inevitably always gonna be completely over my head, I guess that's my only option )

 

[mick thompson]

so what in your opinions would be the cleanest compressor? that would cause the least amount of colorization? or is it an unfair question to ask because effecting gain levels changes sound tonalities regardless to a very different waveform? (covering the entire human frequency and perhaps some barley audible sub freq's)

 

[TheDriller]

because effecting gain levels changes sound tonalities regardless to a very different waveform?

exactly my point

though i suppose you could make a compressor that has absolute minimum of extra colour.

 

[Ermz]

so what in your opinions would be the cleanest compressor? that would cause the least amount of colorization? or is it an unfair question to ask because effecting gain levels changes sound tonalities regardless to a very different waveform? (covering the entire human frequency and perhaps some barley audible sub freq's)

Generally optical compressors tended to be seen as the most transparent, or 'gentle'. This is likely in large part to their optical cells providing multi-stage release curves, which are quite gradual.

In more recent times, the Dynamic Spectrum Mapper really blew that world wide open.

A general +1 to what the others have said about gain changes distorting a waveform. The way that a compressor remains transparent to us is usually due strictly to its design philosophy and the psycho acoustic effects of the way it shapes the envelope.