What happens if I put a low-pass in the amplifier driving an envelope?

[midwayfair]

Let's say I have a simple amplifier followed by a rectifying diode, then the familiar resistor and capacitor pair for the decay:

Amp > diode >
                       |
                    decay
                       |
                    Ground

For whatever reason in this hypothetical circuit I need to keep as much signal as possible, and I'm using a smaller capacitor and larger decay resistor, so an attack resistor would have to be quite big to work properly. But I can very easily add a single capacitor for a low-pass filter to the amplifier driving this (or in front of the amplifier). Will that change the attack? If it does, is there any way for me to calculate the attack time based on the frequency?

I keep thinking I've seen an "attack" control that was basically just a low-pass filter capacitor before a rectifier rather than as a resistor after the rectifier, but I can't for the life of me recall it.

[Mark Hammer]

A lowpass filter ahead of the rectifier simply restricts the bandwidth of what you're rectifying.  Remember that the rectifier gives us a DC voltage, whose rise/fall time can be varied by how quickly a storage cap is charged and discharged.  The sort of LP filter you're thinking of is working on AC, not DC.

All of that simply makes me the first person to post in response.  Doubtless, Paul/PRR or someone else will post here, in followup, and show me where I'm dead wrong, making me scurry back into my little hole. 

[midwayfair]

A lowpass filter ahead of the rectifier simply restricts the bandwidth of what you're rectifying.  Remember that the rectifier gives us a DC voltage, whose rise/fall time can be varied by how quickly a storage cap is charged and discharged.  The sort of LP filter you're thinking of is working on AC, not DC.

All of that simply makes me the first person to post in response.  Doubtless, Paul/PRR or someone else will post here, in followup, and show me where I'm dead wrong, making me scurry back into my little hole. 

If I could remember the circuit I'd seen it, I would have just asked them! Your answer makes the most sense to me, though. Obviously many circuits have a band pass or high-pass to tailor what triggers the envelope.

Although ...

I guess if the highs are filtered out of the envelope, they will not trigger it, forcing the the "slower" low frequencies to be the ones triggering the envelope? Hmmmmmm.

[StephenGiles]

The duty of such a low pass filter may be to minimise high frequency rubbish during the decay portion of the note.

[anotherjim]

I think you would get a weaker attack, since the guitar tone is so much brighter in the attack phase, a low pass will take a lot of that energy away. Better to keep it all in and slow the attack after the envelope rectifier.

Ripple is worse with low frequencies. High frequency content will keep a smoother DC envelope.

Similar to what Stephen said, when the signal is meant to be weak and especially if there's a lot of gain before the rectifier, HF noise could artificially maintain the envelope when you want it to shut up. In that case, some careful pre-filtering is called for, but you have to watch the effect on the high notes. A single pole filter is going to have to start to cut off at quite low f to have an appreciable impact.

You might well need to filter where you have some kind of tracking thing going on - like a squarer feeding further processing - and then you need to restore an envelope at the output. You wouldn't want the tracking to drop out before the envelope, and tracking works best if the signal is low pass filtered to limit miss-tracking on harmonics. So then you'd probably use the low passed signal to feed the envelope.

[bool]

Try to center it at cca 6kHz first, and gradually lower it closer to the anticipated fundamental freq. range. This will make the control signal to appear more compressed (or having weaker transients).

Not sure what you're after, but filtering (band-passing actually) is better suited for gates usually.

[Mark Hammer]

The duty of such a low pass filter may be to minimise high frequency rubbish during the decay portion of the note.

At first blush, that makes sense...until one remembers that the ripple tends to be generally in the range well below fundamentals; say in the under 150-200hz zone.  I'm sure the filtering helps, but it does not cure.  Any LP filtering that extended down far enough in the spectrum would also impose severe limits on attack time.  In principle, a steep 100hz filter would imply that attack times faster than 10ms would be largely out of the question.

[StephenGiles]

The duty of such a low pass filter may be to minimise high frequency rubbish during the decay portion of the note.

At first blush, that makes sense...until one remembers that the ripple tends to be generally in the range well below fundamentals; say in the under 150-200hz zone.  I'm sure the filtering helps, but it does not cure.  Any LP filtering that extended down far enough in the spectrum would also impose severe limits on attack time.  In principle, a steep 100hz filter would imply that attack times faster than 10ms would be largely out of the question.

I just listened to an EH Riddle Qballs demo again, and it seems to have no ripple at all. How do they do that I wonder??

[Mark Hammer]

I thought you had already concluded that it doesn't directly follow the envelope of the input signal, but uses a derivative of the start/stop triggered envelope fund in a variety of EHX products?

[StephenGiles]

I thought you had already concluded that it doesn't directly follow the envelope of the input signal, but uses a derivative of the start/stop triggered envelope fund in a variety of EHX products?

I'm not so sure now, because the controls are slightly different.