Kotelnikov in mid/side mode?

[indigo76]

Hi there,

I'm warming up to TDR Kotelnikov (free) as a mix bus compressor. It can be set to sum (mid) or diff (side) mode, but it does not allow to control both mid and side in one plugin so I have to use two instances. I would like to process both the mid and side signal at the same time on the mix bus.

So, how do I go about using Kotelnikov in this way in Ableton Live? Do I set up a parallel processing chain as an audio effect rack or is it ok to just insert two instances in series (setting one to sum and the other to diff)? The manual is a bit sparse in that regard.

Thanks!

[vitocorleone123]

Hi there,

I'm warming up to TDR Kotelnikov (free) as a mix bus compressor. It can be set to sum (mid) or diff (side) mode, but it does not allow to control both mid and side in one plugin so I have to use two instances. I would like to process both the mid and side signal at the same time on the mix bus.

So, how do I go about using Kotelnikov in this way in Ableton Live? Do I set up a parallel processing chain as an audio effect rack or is it ok to just insert two instances in series (setting one to sum and the other to diff)? The manual is a bit sparse in that regard.

Thanks!

As far as I know, you put them in a series in any DAW: one set to Sum, the other to Diff. It is a UX flaw to have to use 2 separate instances, but it works just fine, and the value of the plugins is unquestionable.

[indigo76]

Thanks! Using two instances one after the other seems to work well.

More specifically, my objective is to copy a setting Dan Worrall used in one of his videos: For the "mid" compressor, a slow attack / fast release, gentle compression with an RMS focus and for the "side" compressor a more aggressive setting, like a fast attack / fast(ish) release and higher ratio as well as a focus on the peaks. Sounds great and really "pushes" the stereo image

[Guillaume Piolat]

Hi there,

I'm warming up to TDR Kotelnikov (free) as a mix bus compressor. It can be set to sum (mid) or diff (side) mode, but it does not allow to control both mid and side in one plugin so I have to use two instances. I would like to process both the mid and side signal at the same time on the mix bus.

Thanks!

Since Kotelnikov nulls when not compressing (though the use of differential oversampling), which is somewhat rare, there will be no adverse consequence to putting both instance in series, one on mid and one on side. (EDIT: well except you'll pay double the latency)

[legendCNCD]

Hi there,

I'm warming up to TDR Kotelnikov (free) as a mix bus compressor. It can be set to sum (mid) or diff (side) mode, but it does not allow to control both mid and side in one plugin so I have to use two instances. I would like to process both the mid and side signal at the same time on the mix bus.

Thanks!

Since Kotelnikov nulls when not compressing (though the use of differential oversampling), which is somewhat rare, there will be no adverse consequence to putting both instance in series, one on mid and one on side.

Thats... thats great to know!

[indigo76]

Thats... thats great to know!

Indeed, makes it all the more useful! This capability should be highlighted more. Blows my mind, really -- where other devs highlight every little feature, this is is just a side note somehow, if at all...

[osiris]

This is something I was reading about yesterday but with OTT compression. Never heard of it before. I think I'll try it. It said put in 2 OTTs with one downward compression then up on the other.

[Guillaume Piolat]

Indeed, makes it all the more useful! This capability should be highlighted more. Blows my mind, really -- where other devs highlight every little feature, this is is just a side note somehow, if at all...

Yeah TDR mentionned "Differential Oversampling" once.
I can only imagine how it works but my guess is that it works like this:

In a typical compressor gain stage that works with oversampling:

Code: Select all

output = downsample(upsample(gain) x upsample(input))

With differential oversampling:

Code: Select all

output = input + downsample(upsample(gain - 1) x upsample(input))

What happens then is that nothing is changed if gain is one.
The oversampling artifacts (say, phase distortion or ringing) are scaled linearly with the gain reducing, which distorts less for small gain reductions.
So in short, only the Delta signal is oversampled and downsampled.

This is clever but you kinda loose possibility of output stage THD.

[chagzuki]

Kotelnikov still tends to be my favourite mix compressor, but I wish it had mid/side with control over channel linking, so that one could control the subtlety. Somehow the overall sound tends to win out over other compressors for me. I spent a long time with Unisum which has such options, but I still find myself going back to Kotelnikov.

[vitocorleone123]

Kotelnikov still tends to be my favourite mix compressor, but I wish it had mid/side with control over channel linking, so that one could control the subtlety. Somehow the overall sound tends to win out over other compressors for me. I spent a long time with Unisum which has such options, but I still find myself going back to Kotelnikov.

Unison is the KGE for mix engineers. Everyone else can get really close with KGE and forego all that money for Unisum - but if you have the skill and knowledge, it seems like Unisum can go the extra 10%.

I skipped Unisum and KGE is always on my mixbus.

[chagzuki]

Unison is the KGE for mix engineers. Everyone else can get really close with KGE and forego all that money for Unisum - but if you have the skill and knowledge, it seems like Unisum can go the extra 10%.

Unisum is hugely configurable and very deep, and I spent ages seeing how closely I could mimic KGE's behaviour in the hope that it would become a one-stop solution, but there's something about it that doesn't quite gel with me. Though I should add that I also tend to use spectral compression prior to Kotelnikov, and that adds extra non-linearity. Unisum has the multiband detector which allows for complex compression behaviour, but I appear to prefer to derive the complexity from spectral compression, and then have Kotelnikov impart the pristine tightness which it excels at. Even with saturation turned on Kotelnikov sounds tight. No rubbery flubber.