Stereo Tool

You can correct me if I play bad, but in the example the band 2 would be affected by the rest of the bands, even herself?

That would become the values ​​0.2, 0.4, 0.25, 0.1, 0.05?

In the example, assume that all bands have a level of 0 dB (or 1) unless the band 2 having-6dB (ie 0.5). If the accounts do this gives a value of 0.71 approx.

This value is the value at which the sample must be taken or the multiplier should apply?
I think this should be the value where the sample should be as if the multiplier make things worse.
And where would the percentage of coupling this?

_________________
by GAP
"Less is More" (Bob Katz)

True, but this is the case with ANY band regardless which one you choose. You make that band more important than others which I think is bad. Now if you don't use too much coupling it's not THAT bad, but still it seems a flaw in the design to do this.
But even if the track has a big imbalance, fixing it will sound bad if it causes the relation between bands to change all the time.
No, I've seen cases where one band stays between the bands around it, but it still sounds bad because it moves too much between the surrounding bands. That's why I didn't implement something that only reduces the level.
That doesn't seem to make sense? If I want a band to drop by 5 dB if it is 10 dB lower than the surrounding bands, that's 50% - but how would you describe that in dBs? (3 per 6 dB, but then it makes more sense to use a percentage).
No, release times can be different and you would create more problems this way.
I don't know if that's good or not. For the limiters I didn't implement band coupling but IF the multiband acts so fast that transients have an effect I think that they should be shared by all bands.

Somewhere around 0.7 I suppose indeed. Which means that the signal is reduced by 3 dB instead of 6 dB as was originally calculated, reducing the effect of a (single) band that's very different from the ones around it.

Percentage is difficult, since all the values depend on both the percentage and how far each frequency is from that of band 2.


I have no time for programming today and tomorrow but after that I'll try to figure out a better method of drawing a line through the points while - as much as possible - keeping the values of the first and last band intact. The code has already been adjusted such that I can easily try out several options.

Ok, wait for the results, very anxious.
In a PM you've left some details of implementations, you can review them and see if they help your thinking.
Thanks for your patience in answering my thoughts.

_________________
by GAP
"Less is More" (Bob Katz)

It occurs to me that the rate applies to your method, and can serve.
Attenuation would know that without coupling and with coupling should have.
The difference between this two values ​​could be 100%.
Therefore just multiply the desired value of the band 2, in the example, the percentage set for that band (or general)
Band 2 = 0.5 + ((0.71 to 0.5) * 0.25) = 0.55 (for a 25% coupling)

_________________
by GAP
"Less is More" (Bob Katz)

I have an idea...
On sliders where is difficult to tune, will be helpfull to add the exact number by typing it.
I know there is an other way to edit the sts file, but might be usefull.

_________________
Sexy Girls in your city

Beware teeth after all couplings was not the same effect as without multi-band

_________________
Urban Junglist

I suggested this some time ago, without any success. I still find it a must given the number of parameters that can be adjusted. Plus it won't hurt anyone nor compromise the performance of ST.

I'm starting to see more and more that this would indeed be useful, just need to think of a way to implement it in the current GUI though (without requiring a lot of extra space).

Actually that could be easy: If you're in 'number typing' mode the whole slider can be replaced by a text field in which you can type a number.

But that leaves the question: How would you indicate to the GUI that you want to type a number?

Stand alone: http://www.stereotool.com/download/ster ... 04-050.exe
Winamp DSP: http://www.stereotool.com/download/dsp_ ... 04-050.exe
VST: http://www.stereotool.com/download/vst_ ... 04-050.dll

I've changed the way band coupling works a bit, the strength of each band is now calculated using:

percentage(0..1) ^ 2log(ratio main band and band we're looking at).
The good thing is that if we are looking at the main band the ratio is 1, the log of the ratio is 0 and anything ^ 0 is 1 again, so I don't need to perform any tricks to get the main band to '1'. So, this MIGHT be better. But I'm not really sure about that.

I've ran some quick tests (not much time today, I've been busy preparing a presentation for the technical managers of 13 tv and 13 radio stations tomorrow) and it seems to work fine.

Also new: You can give the first and last band more 'freedom', this does make these bands more important to the total level though!