COMPRESS with frequency change?

[KMS]

Does anyone know if there is a compressor that increases sustain as the frequency of the note increases?

If not could a tuning meter be kit bashed with a compressor to make this happen such that every note had the same sustain?


KMS stumbling around online.

[petemoore]

Split the signal to two paths, each going through a compressor, choose circuit and voicing for each sides compressor, blend.
 ...one way...
 There may be a way to limit what the Side Chain sees as far as frequency/voltage input...words about which I'm not certain.

[niftydog]

there are frequency tunable compressors... that is, compressor that only compress a filtered band of frequencies. Given the right type of filter you could acheive this result.

It wouldn't be "frequency controlled" but it would progressively compress higher frequencies.

[KMS]

That's a good idea and it would probably solve my problem.  Thanks.

Isn't that kind of like having two ranges at once?

I just finished  a noise gate (Callate Gate from Tone Pad) and I like the way the pedal works but it cuts the high frequency notes off a lot earlier than the low frequency notes (an unavoidable fact of nature) as the higher notes just don't resonate as long as the low notes.

On 6th string up at fret 19 or so there just is not very much resonance to hold the note, and yes the noise gate has a MOM bypass "I know I'm looking for perfection that might not be found".

I don't have a compressor yet (soon "by Christmas" I will make one) and I'm just wondering if you have tried this and it works "to satisfaction” before I make my compressor (s).

KMS seeking a little more than average.

[KMS]

That’s good, a compressor with filter.   So I could look for circuits that offer a filter control.  Don’t happen to know where this could be found do you?  Thanks for the info.

I was just looking at my tuning meter (cover still on for now)and it has LED’s fore each note and I was think that the LED’s are problem controlled by different resistances and the pot on compressor changes resistance so maybe there was a way to evolve the two together.  


KMS bumping into the funiture.

[Ben N]

Marshall's "Ed the Compressor" has an emphasis control, which may be what you are looking for, although it is subtle.  There is a schematic floating around, and I believe it is a CA3080-based variation on the Dyna/Ross model.  

Ah, here it is.  If you send me your address, I will email it to you.

Ben

[petemoore]

Would not comping the high notes attenuate them?
 I'd think one would want the high notes amped a little more than bass.
 One way I do this is slanting the pickup height adjustment in favor of the small strings. This may or may not be enough to cause the highs to trigger compression in a proportion you want compared to the lows.

[StephenGiles]

Why not voltage control the Compression pot!! Then you could have compression proportional to whatever frequency range you want. Food for thought I'd wager!
Stephen

[gez]

I just finished  a noise gate (Callate Gate from Tone Pad) and I like the way the pedal works but it cuts the high frequency notes off a lot earlier than the low frequency notes (an unavoidable fact of nature) as the higher notes just don't resonate as long as the low notes

I don't know the circuit, but you could probably tweak the envelope follower part of it to be frequency selective (usually fairly easy to do) so that highs get more compression.

[Chico]

I think Petemoore had the right idea with regard to the sidechain.  

I think that sidechain processing is much overlooked and is simple to adapt to just about any one of your favorite compressor designs.

That is, you do not need to find a compressor schematic with a filter.  Just pick the compressor design that you like (typical favorites around here seem to be either the FET types (orange squeezer),  the ROSS/Dyna variants, or the optical variants (Anderton, or Flatline).  

Decide the frequencies that you want to emphasize/de emphasize, and build build a filter accordingly.  Then, insert the filter between the sidechain tap and the rectifier.


Best of luck

Tom

[Mark Hammer]

It's not going to work the way you think.

There is a mistaken impression that somehow the compressor increases sustain, and that by emphasizing compression in one part of the spectrum, that sustain will be disproportionately increased there.  Wrong.

Compression produces the illusion of increased sustain by evening out the volume differences betwen the initial transient peaks and the later parts of the decay cycle.  If the note appears to last as long as it does, it is because there is usually some sort of gain added to the signal to compensate for what the compression part takes away.  In the case of the MXR/Ross unit it is Q2 immediately after the CA3080 which does this.

If one selectively compresses certain parts of the frequency spectrum, then one simply affects the gain of the part of the signal where it occurs.  In the case of mids and highs, they tend to occurring at the start of the note during the attack transient.  As the note goes on, that content dies out very quickly.  Indeed, that is part of why people like distortion so much: it maintains the harmonic content that would normally disappear all too quickly.

De-essers and de-essing functions of full-featured rackmount limters and compressors ARE exactly what you requested.  The sidechain is optimized for responding to a select upper-mid/lower-treble band produced by the 'S' sound.  Compressing it tends to tame it in singers that have a little too much "friction" in their mouths that might overload other parts of the signal chain.  It does NOT increase the treble sustain, but in fact downplays treble relative to bass.

With any sidechain-governed effect, one CAN shape the relative sensitivity across the frequency spectrum simply by tone-shaping the signal fed to the envelope follower.  Some compressors provide more opportunity or flexibility for that than others.  For instance, in the MXR1Ross, the two transistors that form the rectifier/follower are each preceded by .01uf caps.  Reducing the value of those caps will have the effect of attenuating bass content in the signal the follower receives.  This will mean that it will reduce the gain of low notes less than it will for higher notes (in the MXR/Ross case, higher envelope signals mean MORE gain reduction.  This is the opposite of the effect you are seeking.  The same thing would occur if you stuck a much smaller-value cap between the output of the op-amp and 1k5 resistor in the Orange Squeezer.

In the Anderton compressor, since it is built around an op-amp (where high end is simply a matter of caps in the feedback path), both high AND low frequency content going to the envelope follower can be altered.  In that design you could drastically reduce high end going to the follower and leave low end as is.  

What would the audible effect be of doing this?  It would produce compression of lower notes, but much less compression of higher notes.  That means that bass notes would "pull back" when you pick them harder, but higher notes would be pretty much as loud as you pick them.  Would it increase the sustain of higher notes?  No.  It would simply restore normal dynamics, and the note would ring as long as would normally ring with that degree of boost and amplifier volume and finger vibrato.

To achieve what you want, several things would need to be in place.  First, as others have suggested, the audio path to the envelope follower would have to be split into higher and lower (and maybe even 3 bands) with each path having its own individual envelope follower.  The envelope follower for upper-range content would need to have a longer decay time than that of the lower range.

Just as important as this, however, the audio signal would have to be split into different gain recovery paths.  So, for example, the bass/mids might have a gain recovery stage set for x3 while the treble might have a gain recovery stage set for x8.  The multiple gain recovery stages, meanwhile would have their outputs mixed down to mono.  The long decay of the treble envelope follower would mean that the gain at that end gets reduced for a longer time, but since the gain recovery of that same band is set higher it should keep the level of that content somewhat higher for a long time.

It's obviously a much more complicated design than simply changing a cap here and there.

[niftydog]

There are obviously many opinions on how to do this, but for my bass rig, to emphasise the high frequencies, I squish the hell out of the low frequencies. However, I do agree that getting sustain (as opposed to gain) in the high register is going to be tricky without resorting to the feedback type of sustain!

Still, I would have thought a good compressor should handle this without the need to go with any frequency specific processing. The lower frequencies have more energy, thus they should be squashed more. The high frequencies result in less triggering of the compressor and thus they should be relatively untouched.

Any kind of dynamic control of a compressor is likely to result in "breathing". This is what happens when compression is very high; it results in a noticeable rise in the noise floor during quiet passages and a sudden "whooosh" at the end of every song!

[KMS]

Ok I read all of the above and there are lot things to think about here.

First of all, I was not considering the frequency of a chord.  That would kill my theory quick.  What frequency is an Edim? The noise gate will start attacking when the “signal strength” reaches a low threshold.  

I guess I’m learning as I go. I played around with music for years and now I’m finally getting serious about FX and there is a lot to learn.

A preamp that adjusts the output to increase as the input decreases would be more along the lines of what I’m looking for, I think.


KMS

[niftydog]

A preamp that adjusts the output to increase as the input decreases would be more along the lines of what I’m looking for, I think.

that is essentially what a compressor does, but it ACTUALLY decreases the output as the input increases. But the result is the same.

[KMS]

Thanks,

I think the problem has been solved.

Somewhere in the chain….A splitter then two or three compressors filtered/voiced and max setting then mixed back together and next a pre-amp primarily set on high or with a foot volume control for output then somewhere later in the chain, the noise gate.  That would be exactly what I was thinking about.  That would keep the signal high enough on the high frequency notes (with some adjustment) to stay away from the decay on the noise gate and when I finger mute the strings, silence will be heard.

I finished all the mods on the noise gate and took the attenuation back off, it decreases my signal and everyone should know by now that I don’t like that.

The release and attack mods smoothed the decay out good enough that I can hear the decay with enough time to finger mute before the choppy clipping occurs as the gate finishes the mute.

You can find the circuit here http://www.tonepad.com/project.asp?id=18

I gave up on trying smoothing it out with a small cap at the decay threshold as one of the build reports suggested.  I think I found the right spot (ground to emitter at Q2) but it added distortion.

I found another spot too (forgot where somewhere at Q3) that totally smoothed it out with no distortion but the gate would not quite close all the way.

Does anyone know what part of the above circuit is the decay threshold? I think it is one of the leads on Q2 but I don’t really know.

It works real nice, and I’m glad I chose it for my first project because now I have a direction for the rest of my chain.  I tried my Cry Baby 535 with boost (noisiest thing I got) inline with my Boss Turbo OD also kind of noisy and the decay on the noise gate was  not  an issue, but I have the boost up on the 535, which is not what I want because I loose the boost when the 535 is off.  I guess I could modify that and use the 535 boost as a fulltime preamp.

The end of my chain (for now) is a modified (to about 2 seconds) rack mount Electra EP 200 analog delay with reverb (separate stomp box with umbilical cord for both delay and reverb). Old but true analog.   I had to put it after the noise gate because the gate decayed/clipped the end of the delay sound.  Surprisingly, the EP 200 does not make any noise by itself and the reverb is even quiet.  I had never used it with clean sound for any length of time and I was always disappointed with it because it multiplied the noise of my Boss to the max so it has set on the shelf unplugged for a long time. Not a problem now.  

I wish I had the money to buy everything factory, but then I wouldn’t know what to buy and how this should all work and why it works.  

KMS good night.

[Mark Hammer]

Somewhere in the chain….A splitter then two or three compressors filtered/voiced and max setting then mixed back together and next a pre-amp primarily set on high or with a foot volume control for output then somewhere later in the chain, the noise gate.  That would be exactly what I was thinking about.  That would keep the signal high enough on the high frequency notes (with some adjustment) to stay away from the decay on the noise gate and when I finger mute the strings, silence will be heard.

Just bear in mind that compressors boost everything and only turn down the gain when the input signal is loud enough to justify it.  That means that any hiss or noise coming in the input will be boosted unless it is loud enough in the first place.  What THAT means is that you usually want your compressor circuitry to be either as low noise as possible, or as simple as possible (to avoid sources of noise), in addition to your input signal being as noise free as possible.  What you've described above is certainly feasible, but my guess is that identifying a truly low-noise way to do it, without taking months and months of trial and error, will be hard to do, especially if you are just finding your way around now.

IF you want to learn more, it will be an EXCELLENT situation to do so.  If you want to get playing again, maybe not.