Compression and saturation, in what order?

[tahome]

Music production is a listening art. Theories of DSP coding are another subject entirely.

I did say that, didn't I? And I also said why this is the case in practice.

Read my previous post if you would before you accuse someone of being deaf. We *are* in fact talking about theory here.

If I am really deaf you must be blind!

--th

[Sascha Franck]

Ok, I'll try with freeverb and Logics stereo delay in the most neutral setting.

[bagginz]

I did say that, didn't I? And I also said why this is the case in practice.

O.k fair enough. However I do believe that the original poster was looking for practical advice rather than dsp theory, but anyway...

Also I just did an a/b test and to be fair to you there is much less difference than I remember betweeen verb > delay and delay > verb.

Cheers,
Bagginz

[Krakatau]

I did say that, didn't I? And I also said why this is the case in practice.

O.k fair enough. However I do believe that the original poster was looking for practical advice rather than dsp theory, but anyway...

Also I just did an a/b test and to be fair to you there is much less difference than I remember betweeen verb > delay and delay > verb.

Cheers,
Bagginz

That's what i pointed at, yesterday now one or two pages before

___

@tahome, just wait a while, its late actually and i just need to care a bit of my family....

....tomorrow late morning on that thread, don't worry !

[Krakatau]

@tahome, ...as promised there are of the some links i could recover from kvr's database :

http://www.kvraudio.com/forum/archive/e ... 5/5241.php

http://www.kvraudio.com/forum/viewtopic ... 31&start=0

http://www.kvraudio.com/forum/viewtopic ... highlight=

http://www.kvraudio.com/forum/viewtopic ... sc&start=0

[Krakatau]

if you would ever be curious, i be glad to find again the links to the related threads ...

I am. Didn't I just say that?

--th

[loophead]

tahome, sorry but you are absolutly incorrect. Lets change the example so all can clearly see tha flaw in your logic

Delay and flanger:

Flanger > Delay: with each repeat of the delay we get an echo of the flanged signal and therefor each echo 'sounds' similar.

Delay > Flanger: Each repeat of the delay gets 'flanged' differently ie will hit a different phase of the waveform sweep and thus each echo will 'sound' different.

This of course is clearly heard when the flanger and the delay are no 'locked' together delaying and sweeping at the exact rate/time.

So sorry, but order makes HUGE differences in many, many cases.

[Krakatau]

To this, he already gave an objective answer in this thread

flanger is based on a MODULATED delay that induce a non.linearity
sometimes reverbs themselves are modulated to add character

technically and theorically tahome's right, take a time to read this thread a bit more carefully and you'll probable agree...

[Sascha Franck]

tahome, sorry but you are absolutly incorrect. Lets change the example so all can clearly see tha flaw in your logic

Delay and flanger:

But didn't he say that all this wouldn't be valid for modulating FX?

[Krakatau]

exactly...

[Christian Schüler]

Can't help but post here because my footnails come high, sorry guys.
All this linearity stuff is not so much helpful for the recording engeneers, but rather for plugin developers that can know how to factor internal computation steps. However I think it's also good for a music producer to know theory.


With that said, linear time-invariant systems (LTI) are commutative and distributive:

distributive
1. Effect( Signal A ) + Effect( Signal B ) = Effect( Signal A + Signal B )
2. gain * Effect( Signal ) = Effect( gain * Signal )

commutative
3. EffectA( EffectB( Signal ) ) = EffectB( EffectA( Signal B ) )


In other words, a system is linear if

- the effect is the same for soft signals and loud signals. This means obviously a compressor is not linear. Any device with saturation is not linear either.

- the effect stays the same over time. This means that with any modulation an effect looses linear status. To quote the post with the flanger, a flanger and a chorus are obviously not linear in this sense, and no not commute.

Since you cannot look into plugs, its hard to tell whether a plug is linear. The test is simple: Conditions 1. 2. and 3. must hold!


Linear effects in theory:

- A simple volume gain is linear. Obviously +6dB followed by -6dB is the same as other way 'round, clipping or other profane things like noise and precision not withstanding.

- A simple phase shift is linear. Obviously +90° followed by -90° is the same as other way 'round.

- A pure EQ is therefore linear, since this is just a frequency dependent volume gain and phase shift.

- A pure delay and a pure reverb are therefore linear, since they are a special case of EQs, with very many densely spaced sharp resonant peaks (Q's), regular spacing for delay and irregular for reverb.


In practice, digital effects can get quite close to the linear ideal, the only things preventing it are limited range and precision of the underlying number representation.

[tahome]

Right... er thanks... but... why do you post all this? I've said exactly what you have written here many times in this thread already - you would have saved yourself a lot of time writing this if you had cared to read the earlier posts. Not that I disagree with what you've said, your comment is exactly on the mark.

Cheers,
--th