Some thoughts on another way to do multi-band compression

[midwayfair]

There's been some talk here and elsewhere about multiband compression, and everyone keeps trying to come up with ways to use two different copies of a compressor and mixing them together. I've been trying to think of a much simpler way.

In the Bearhug, the method of compression is to reduce the gain of a 47uF AC bypass capacitor. If I shrink that capacitor, then the frequency affected by the compression is higher.

So wouldn't it work to drive two different FETs+rectifiers attached to two different AC bypass caps: one very large cap to compress the bass frequencies, and one very small cap to compress just the treble. By controlling the threshold to each FET, you can prevent either band from being compressed.

No cancellations -- because it can all be done on the same transistor. No mixing stages -- heck, you don't even have to add an extra transistor, because you can just decouple Q2 with two separate caps and run the rectifiers parallel. Something like 6 extra parts to an already tiny circuit.

Thoughts?

[stecykmi]

pretty clever. without any sort of testing, this might be an interesting way to make a 2-band multiband compressor.

by using the rectifier as the crossover, you do limit the amount of controls available to each band however (for example, the comp control(s) are not independent). the cap values for the rectifier (that generates the gain reduction envelope) will have to be carefully chosen, too large and it will not be able to compress high-frequencies, too small and the effect may be too subtle.

you may have better luck with using a second JFET in the same AC bypass configuration. this is just a handful more parts and i have a feeling you may be able to get better results.

[midwayfair]

you may have better luck with using a second JFET in the same AC bypass configuration. this is just a handful more parts and i have a feeling you may be able to get better results.

This is what I meant -- two FETs. I can see how I was a little unclear.

Threshold control to each FET would be the comp control for that FET and thus *that band*.

Is there still a danger with the rectifier coupling cap in that situation?

[jkokura]

I can't speak to the issue, but I can think of the application and I wonder how useful this would be? When is a multi band compressor really going to work better than, say, one of the other wonderful designs we have out there right now?

Don't get me wrong, I understand the concept and application of a multiband compressor, and if one gets designed here I'll likely build or play with it, but rather - I never use a multiband comp on my guitar signal when I'm recording and compressing (though I can), when a regular single band comp works just fine?

I guess I'm playing devil's advocate a bit here and asking - is there something broken that needs fixing?

Jacob

[ThePastRecedes]

Kinda irrelevant but I had an idea to use your cardinal tremolo and in place of the lfo use an envelope to control the vectors. I am still yet to bread board it yet. Only an idea in my head currently.

[midwayfair]

Kinda irrelevant but I had an idea to use your cardinal tremolo and in place of the lfo use an envelope to control the vectors. I am still yet to bread board it yet. Only an idea in my head currently.

This has the frequency cancelation problem. In fact, the cardinal is kinda made to take advantage of the cancelations (hence the phasing).

Jacob:

Here ya go. http://www.ovnilab.com/articles/dualband.shtml

I'm not going to pretend it's a holy grail or anything, but it's similar to blended compression.

Also ... Because we build lots of things we don't need.

[stecykmi]

you may have better luck with using a second JFET in the same AC bypass configuration. this is just a handful more parts and i have a feeling you may be able to get better results.

This is what I meant -- two FETs. I can see how I was a little unclear.

Threshold control to each FET would be the comp control for that FET and thus *that band*.

Is there still a danger with the rectifier coupling cap in that situation?

i think it should work, but you would obviously have to filter the rectifiers to the appropriate bands, ie cut the high freq from the low freq rectifier.

bass players would probably find this design useful! a fully adjustable crossover might be beyond the scope of this project, but a guitar/bass switch can probably be implemented with just a couple of cap substitutions.

[midwayfair]

you may have better luck with using a second JFET in the same AC bypass configuration. this is just a handful more parts and i have a feeling you may be able to get better results.

This is what I meant -- two FETs. I can see how I was a little unclear.

Threshold control to each FET would be the comp control for that FET and thus *that band*.

Is there still a danger with the rectifier coupling cap in that situation?

i think it should work, but you would obviously have to filter the rectifiers to the appropriate bands, ie cut the high freq from the low freq rectifier.

Can you explain why? It seems to me that the treble is either allowed to compress, or it's bypassed and boosted and the bass band can't affect those frequencies.

The actual frequencies that the rectifier converts to DC shouldn't affect the frequencies .. Um, I think.

The crossover freq is determined by the size of the treble band's bypass cap, so it's easily adjusted.

I need to make a schematic or block diagram, but my desktop is dead at the moment.

[stecykmi]

i think a block diagram would be useful.

my reasoning is since we'll probably want different gain reduction responses from the treble and bass bands (otherwise why bother, that's the whole point of multiband comp), the envelope generators should only be triggered by their appropriate frequencies. for instance, typically a faster response is used for treble frequencies so the treble rectifier shouldn't receive any bass signal that might "contaminate" the response.

i'm not sure we are totally on the same page with the implementation here, haha.

[jkokura]

In general I recommend multi-band compression for tappers, slappers who want a bright sound, people who play a lot of harmonics and ghost notes, extended-range instrumentalists, and others who find compression takes too much away from the detail of their sound while playing across a wide frequency range. For most other players -especially ones who prefer smooth squish, deep boom, or a more "glued together" sound quality- I recommend traditional single-band compression.

See this makes sense to me, and it seems like it would be best applied for a Bass player, especially one who uses 5 or 6 strings. I think that's why many of the comps on that list are for Bass.

Jacob

[kothoma]

i think it should work, but you would obviously have to filter the rectifiers to the appropriate bands, ie cut the high freq from the low freq rectifier.

my reasoning is since we'll probably want different gain reduction responses from the treble and bass bands (otherwise why bother, that's the whole point of multiband comp), the envelope generators should only be triggered by their appropriate frequencies.

That's what I'd think too.