Parallel Compression in a stompbox

[John M]

Firstly apologies if this is a stupid question or has been discussed before, but I have searched and can't find anything on it (maybe you guys use a different term). While waiting for some parts to go on my merry way of experimenting, I thought I would ask to see if there is anything obvious I am missing. As a bit of background, I have yet to find a stompbox compressor that works for me. Having spent years finding the guitar and amp I want, I find that the compressors that I have tried destroy all I worked so hard to find. What I am looking for is something that can be left on to fatten my sound while retaining dynamics, especially when playing at lower volumes.

When recording my guitar I have had good results either using mild limiting or none in parallel with some mild compression to fatten out the sound. So I thought (and this is maybe where it gets dangerous), what if I tried to do that in a stompbox? So here are my questions:

Is it as simple as going into a compressor (switchable on and off) in parallel with your clean signal (switchable on and off) which has an optional switchable clean boost, with both going into a limiter (switchable on and off) , thus allowing you to have (1) your clean signal, (2) a boosted clean signal, (3) your clean or boosted clean signal with some parallel compression and (4) any of the previous 3 with some limiting.

It just seems so simple that if this is all it took, I should see products out there that did it, so am I being stupid trying to put a DAW into a stompbox?

The compressor I shall be trying when the parts arrive is the flatline compressor. Is there a transparent DIY limiter anyone would recommend?

Your expertise would be welcome.

John

[petemoore]

  There are many ways that parallel compresison could be approached.
  evening out the treble / bass balance of the triggering [might only require pre-treatment processing of the signal sent to the envelope].
  Or having an external or internal...anything processing wise you like, such as reverb or echo, only influencing the HF's, I'd use jacks to 'insert' the effect, having the bass and treble parallel and each it's own envelope filter would allow treatment of the signal before the envelope filter [frequency filtering comes to mind, LP filter for the lows, or perhaps a bandpass filter to compensate or overcompensate for or against [so to speak...would allow different triggering alteration and the resultant the gain-sweep.
  Perhaps what you want is what you described, perhaps...sounded good too !
  There are 'too many' options, but usually the simplest approach that solves the all or most of the shortcomings is preferred...because it gets complicated.
  1 comp here has treble orientation filters added, and the envelope can be independently doctored from the
  Other comp for bass, also with an independantly doctorable detector.
  Each has eq set or adjustable for the relevancy of warranting the adjustor knob.
  Probably doing pretty good, large, but featured.
   

[Mark Hammer]

To my mind, you clearly want a combination of some "average" level and a "reserve capacity" to introduce audible peaks and transients.  In some respects, that might be achieved through clever use of a 570 compander chip.

Consider...

The compression half of a 570 (or 571) will deaccentuate the dynamics of the signal, reducing the contrast between the quietest and loudest portions.  When the other half is configured as an expandor, it accentuates the dynamics such that the difference between the quietest and loudest portions is exaggerated.

Now, let us imagine we split the signal at the input (with suitable buffering, of course), and feed it to the two sections of a compandor chip; one configured for compression, and the other configured for expansion.  Both of these sections now get fed to a mxier stage such that we can blend them.  When blended "just right", any quiet segments that would be nearly inaudible from the expandor, would still be quite audible, coming from the compressor.  And, since they would be exaggerated, any peaks where you dig in with pick attack, that would be reined in by the compressor, would be audible from the expandor half.

At least I think that's what you're looking for.

That sounds eminently do-able, except it may want some careful planning out of levels and filtering.  So, it may require sending a hotter level to the expandor than compressor half, but filtering out the bass somewhat from the expandor output (so that the expanded signal is primarily pick attack), and maybe trimming back the expandor output a bit so that appropriate balance is easier to achieve.

Des that sound in the ball park, or am I thinking too much in terms of a Telecaster here?

[John M]

Thank you for your responses gentlemen, good thoughts. I will consider these in my experiments.

Mark you are right it is a telecaster (well actually a Fret King Country Squire Yardbird). That is why retaining decent dynamics is important.

John

[Mark Hammer]

Jack Orman used to have a circuit called the Comprex (my copy is dated 1999), which used an NE570/571 to drive a clipping stage.  The user could select either the compressor half or the expandor half.

Seems to me the basic circuit could be used without clipping diodes (and extra gain in the mixing/overdrive stage), as I described earlier.  Rather than selecting compression OR expansion, you could use a blend pot.

I'd send you the schem, but it may be something Jack prefers to retain control over.  Pop him a note at AMZ.  If he doesn't mind me sending it to you, I'll be happy to send it along.  I think it may well be modifiable to do what you want.

[Processaurus]

This is a  mix  knob add-on for the Ross compressor, it does the parallel compression to that popular stompbox compressor.  No limiting afterward, that would be another circuit.  It turned out sounding nice, though, and made the Ross's slow attack sound more natural, and less *pop*-ey. 




That's my favorite stompbox compressor right now, though rack compressors are going to be superior, as far as noise performance, and having proper attack/release/threshold/ratio/gain controls.  I've been using an inexpensive FMR Really Nice Compressor and love it on clean guitar.  They should build it into a pedal!  Maybe there's room on top to stick in a stompswitch...

THATcorp has some interesting app notes for compressors, using their specialized IC's, that look superior to the traditional 70's stompbox circuits.

http://www.thatcorp.com/Design_Notes.shtml

[John M]

A question for Mr Processaurus (or anyone who would care to answer). If using this blend circuit with the Flatline, where would be the ideal place in the Flatline circuit to take the dry signal from?

Thank you
John

[John M]

Flatline schematic for the above question.

www.hollis.co.uk/john/flatline.jpg

John

[Processaurus]

Hi, the circuit would need some modifications to work with the flatline design, because it flips the phase twice to correct for the quirk of the Ross Comp's output being out of phase.  It flips the dry out of phase, then uses a classic panning circuit to mix the out of phase dry with the (also out of phase) wet, compressed sound, then takes that mix and flips the phase again, leaving the final output in phase with the input.

Hollis's doesn't  invert the phase, so you'd want to alter my circuit around so that the first TL072 is a non-inverting amplifier with a gain that approximately matches the dry signal to the full volume of the flatline.

[Nasse]

Not sure if dod fx80b does something like that

[JDoyle]

made the Ross's slow attack sound more natural, and less *pop*-ey.

Processaurus -

Are you sure you mean the attack? In a Ross the attack occurs in around 1 msec, which is awfully fast.

Regards,

Jay Doyle

[Processaurus]

made the Ross's slow attack sound more natural, and less *pop*-ey.

Processaurus -

Are you sure you mean the attack? In a Ross the attack occurs in around 1 msec, which is awfully fast.

Regards,

Jay Doyle

That's interesting, is that 1ms figure from an analysis of the circuit, or some audio?  To my ear it has a funny, slow, 10ms-20ms attack.  When I first built one I recorded some guitar through it and looked at the waveform on the computer, I could see a big spike at the beginning of the notes, where a couple cycles made it through at high gain before the envelope kicked the compression on and squashed down the signal.

[JDoyle]

Well - I haven't looked at mine on a scope in quite a while to be honest. That said, the only restraint on attack time is the re of the transistors as they dump the current from the cap and before the waveform has completed a cycle both re's will be in parallel reducing it further. Assuming worst case it shouldn't take more than 2 msec to discharge that cap, maybe up to 3 with low beta transistors in the FWR, but definitely not 10-20...

At any rate - I've built a version with bypassable attack and decay controls and if I do set the attack long I can hear what you refer to but when I bypass the A/D circuit (which makes the overall circuit a stock Ross) I don't have any issues with long-ish attack letting initial transisents bleed through...

Interesting...

[John M]

This has been on the back burner for a while, but started to think about it again. Would the Buff 'n Blend be more suited to my needs or is it not likely to work well?

John

[Processaurus]

This has been on the back burner for a while, but started to think about it again. Would the Buff 'n Blend be more suited to my needs or is it not likely to work well?

John

Much better, use the Buff'n'Blend or variation.  The circuit I'd posted is just kind of a specialized, Ross specific version of that idea.