Quote:
OK, but what is if you have 7 or 9 bans and want to couple e.g. 3 to 4. That´s only possible with the Matrix.
True. There is a global control (the Band Coupling slider in the Multiband main panel) that controls all couplings... but what would you want to achieve with this? (Why would you want a coupling between bands 3 and 4 that's different from -say- the one between 4 and 5?)
Quote:
And I think it´s not a band coupling issue to prevent loud highs from distortion when loud bass waves do occur. This is the task of ABDP, and this section does this job very well.[/quote[
True, but I've seen it done in other processors. It can smooth things for the clipper allowing -maybe- higher clipping levels without getting distortion.
Quote:
In Basics the band coupling does add the gain reduction signal of the source band to the gain reduction signal produced by the target band’s compressor sidechain. This sum determined the gain of the target band’s VCA.
The only downside of this is, when this action loweres the drive to the target band’s sidechain below the compression threshold for that band, it causes the sidechain’s gain reduction signal to decrease towards 0 dB at a speed determined by the target band’s release time setting.
Now I'm confused. What currently happens is that if band 4 wants 6 dB of reduction and band 5 wants 12 dB of reduction, and coupling (on band 4) from band 4 is 100% and from band 5 is 25%, with Normalize turned on, then the actual values used are 80% and 20% and band 4 will get 6 dB * 80% + 12 dB * 20% = 7.2 dB reduction. So, the reduction can also become less - if band 5 would have "asked for" 0 dB reduction and band 4 still for 6 dB, you would get 4.8 dB reduction in band 4.
Quote:
To get rid of this, the behavior has to be changed so that the gain reduction in the target band immediately becomes the higher of the target band sidechain’s gain control signal or the source band’s gain reduction signal as multiplied by the relevant Band Coupling control.
You mean always use the maximum, so in the examples above that would lead to 7.2 dB and 6 dB instead of 7.2 dB and 4.8 dB? That might be a good idea...
Quote:
So the attack and release time constants of the target band’s compressor sidechain no longer affect the band coupling.
I don't understand this one...
Just a small explanation of what band coupling - in my mind - was supposed to do.
Say, you have a preset with 9 bands. Now if there's a 'hole' in the spectrum, it can happen that one band is not reduced at all, while all the others are reduced a lot. Which would
sound the same as boosting one frequency a lot - say by 6 or 12 dB. I probably don't have to tell you that this sounds pretty bad (you can easily try it out, just increase the Band Mix of one of the center bands by 6-12 dB).
The whole idea of band coupling as it currently works (the slider, or the matrix) is to avoid having 1 band stick out too much. If the surrounding bands are down a lot, they should pull this center band with them.
I can indeed imagine (as I wrote above) that the opposite might be less needed: If one band is very far down momentarily that's probably ok, especially since there will be a lot of audio in that band at that moment anyway (otherwise it wouldn't be down that far).