Crossover + Compression

[Eddododo]

Here's my swing at a multiband comp- the goal was to have low-end compression to fatten up the bass with the highs passing unaffected.

You'll see the obvious state variable filter for a tunable crossover point, and a fan favorite in the Hollis Flatline for the compression section. I chose the flatline for its compress-to-unity quality.

Note that its a bipolar supply.. I dont have the biggest reason for it other than my amusement, and i like the idea of minimizing extra caps.
Also- I'll address the switching around the SVF caps. It's an on-on-on switch that shorts the 10meg resistors. the resistors are something I use a lot to alleviate cap swtiching pop.
The pot gives a lot of range, but switching caps will give three different 'behaviors;' the initial low band compression to fatten bass guitars, a 'most of the range' compression that simply lets the highs pass unaffected, and then a vintagey mode, where the crossover happens in the highest mids, and then the high band will be lowered in volume. anyway...

A couple of questions..
the red circle- this is a wet/dry blend for the low band- note that the dry is being fed from the op amp in the SVF with the wet obviously being fed from the main compressor amp . Will i run into problems with this setup? I didn't want a whole amp to isolate the dry from the wet.

The blue circle- I need a little context for how the LED will be affected by the ground reference instead of a v/2 reference. If there's much of an issue to overcome I may just do a more formal envelope follower with some envelope control.

Any other issues with how I'm putting this together?

[Sooner Boomer]

The problem I see right away is phase.  You're running the highs through a unity-gain buffer amp, then a filter to pick the highs off, then into a summing amp to blend everything back together.  At the very least you're going to get a phase delay error because of all the extra stuff the lows go through.  But I could be wrong.  It might sound great.

[Keppy]

Cool idea! I've done something similar with distortion, but haven't tried it on a compressor.

the red circle- this is a wet/dry blend for the low band- note that the dry is being fed from the op amp in the SVF with the wet obviously being fed from the main compressor amp . Will i run into problems with this setup?

No, but you will in the next stage. The inverting stage following the red circle will put the polarity of your lows opposite the highs. Fortunately, it looks like you can just convert it to a non-inverting buffer.

The blue circle- I need a little context for how the LED will be affected by the ground reference instead of a v/2 reference. If there's much of an issue to overcome I may just do a more formal envelope follower with some envelope control.

That section isn't referenced directly to ground or v/2, just to the outputs of the opamps. I don't see a problem. You do have twice the voltage potentially available to drive it, but it doesn't look like you're hitting it hard enough for that to matter.

Any other issues with how I'm putting this together?

Consecutive inverting stages usually have coupling caps so that any tiny voltage offsets of one stage don't get amplified by the next. Without tons of gain you probably won't have a problem, but if you find DC voltage on your opamp outputs when debugging that's probably why.

You have an extra opamp stage in the compressor section. The buffer isn't needed coming directly off the output of the stage in red.

I don't get the parallel caps at input and output, especially after you mentioned minimizing extra caps. It's not a problem, just something I noticed that seems odd. I also suspect you won't need all three compressor pots and all three volume controls, but it's hard to know WHICH you don't need until you have it put together

The problem I see right away is phase.  You're running the highs through a unity-gain buffer amp, then a filter to pick the highs off, then into a summing amp to blend everything back together.  At the very least you're going to get a phase delay error because of all the extra stuff the lows go through.  But I could be wrong.  It might sound great.

The main phase issues are inherent to the filter, which is a known-good design. You'll hear it like you hear any crossover, but nothing worse than that. The only component in the compressor itself that affects the signal phase is the one cap in the diode section, which doesn't seem like a big deal.

tl;dr – I think you're good to go outside of that one inverting buffer before the summing amp. Good luck!

[Eddododo]

No, but you will in the next stage. The inverting stage following the red circle will put the polarity of your lows opposite the highs. Fortunately, it looks like you can just convert it to a non-inverting buffer.

Well actually this is intentional- the output of the HP and LP of the filter are out of phase with one another. This corrects it

That section isn't referenced directly to ground or v/2, just to the outputs of the opamps. I don't see a problem. You do have twice the voltage potentially available to drive it, but it doesn't look like you're hitting it hard enough for that to matter.

Yea, my only real concern was it behaving differently than a v/2 as in the original flatline, but I suppose some actual testing must be done haha

You have an extra opamp stage in the compressor section. The buffer isn't needed coming directly off the output of the stage in red.

Im using it to try to sum the dry and compressed signal.. seemed necessary especially as without it theres a blend pot (arbitrarily 25 k) going straight into the 100k vol pots.. perhaps there is a more elegant solution.

I don't get the parallel caps at input and output, especially after you mentioned minimizing extra caps. It's not a problem, just something I noticed that seems odd.

its TOTALLY unnecessary, but its just a little nod to 'hifi'... it would be a large value electro paralleled with a regular ol' film. Pretty overkill with my messy circuits, but its just to preserve high response better in the presence of electrolytics.. one of those things that I don't know if i notice, and no one seems to do it but it only costs me an extra cap here and there.

As far as all the knobs... well... it's easier for me to overdesign on paper than to breadboard test for hours with a 2-year-old chasing me around lol. I end up setting and forgetting knobs with  more ease than making a damned decision at the design phase. 

thanks for the input!

[Keppy]

the output of the HP and LP of the filter are out of phase with one another.

I thought the HP and LP outputs were in phase with each other, opposite the BP output. I'm no expert, though.

You have an extra opamp stage in the compressor section. The buffer isn't needed coming directly off the output of the stage in red.

Im using it to try to sum the dry and compressed signal.. seemed necessary especially as without it theres a blend pot (arbitrarily 25 k) going straight into the 100k vol pots.. perhaps there is a more elegant solution.

Not that one, the one at the bottom of the drawing. The blend pot is exactly what I would do in the situation, and plenty elegant.

[Eddododo]

oh duh. what you meant actually occurred to me and then I forgot by the end of my quote job.