Time and parallel effects

[Mark Hammer]

It is a time-honoured production/engineering strategy that if you want to get a HUGE guitar sound, one way to do it is to use several different amps, and mic them up.  One of the things that may happen in the course of doing that is that the mics may be placed, and even need to be placed, at different distances from the speakers, resulting in ever-so-slightly different arrival times.  The objective is not to produce an audible time difference, but probably to get a kind of harmonic richness and animation.

The topic came up today on another forum, during discussion of the recently released Multidrive from local legend Empress Effects (http://www.empresseffects.com/multidrive.html).  The Multidrive allows for blending of up to three kinds of harmonics-generating effects, each with the ability to apply filters to isolate sections of their signal.

What the Multidrive doesn't do (as far as I know) is allow for that slight time-stagger that can occur in multiple-amp mic-ing.

Since some of you here are far more involved in recording than I'll ever be, I was curious about what sorts of delays actually arise when doing that, and what the practica limits might be be staggering mic'd amps to get a bigger sound.  Seems to me that a mixer with a wee bit of adjustable time-stagger on one or two inputs might be a neat way to accompish the same sort of thing.  Some of the good folks here have contributed ultra-short delay circuits and layouts over the years, either using BBDs or multi-pole allpass filters.  This might be something to tinker with.

[Processaurus]

I've been curious if the pt2396 would be handy for applications like this, it is capable of some short (shorter than the pt2399,  40ms) delay times, that are simple to implement, if you just need a couple choices of short delay times rather than a continuously variable adjustment.

It would be very cool to have a hi quality digital delay circuit to drop into projects that could go from .5ms to say, 40ms.

[CynicalMan]

If you're just looking for a short delay, you could mod maarten's ADT:


Take out both op amp stages, take out the Vr supply, throw on a FET input buffer, and you'd have your delay in 10 parts! Or you could leave the op amp stages and use a 50k pot on the mixer input (with a series resistor) for adjustable delay mix.


Amplitube Free is nice for trying out ideas like this. It allows parallel procession with two stompbox chains, two amps, two cabs, and two sets of rack effects.

[Mark Hammer]

If we're looking at the difference between an amp on one side of the room with a mic 3 ft away, and another one on this side of the room with a mic right on the edge of the speaker, then I doubt we're looking at any sort of time-stagger that would warrant use of 2048 BBD stages.  But that certainly fits the bill in terms of simplicity of design.

[CynicalMan]

You can always use a MN3*06 (128-stage) or a MN3*09 (256-stage) for lower delay times.

[chi_boy]

If you're just looking for a short delay, you could mod maarten's ADT:


Take out both op amp stages, take out the Vr supply, throw on a FET input buffer, and you'd have your delay in 10 parts! Or you could leave the op amp stages and use a 50k pot on the mixer input (with a series resistor) for adjustable delay mix.


Amplitube Free is nice for trying out ideas like this. It allows parallel procession with two stompbox chains, two amps, two cabs, and two sets of rack effects.

I don't mean to jump off topic, but this looks mildly interesting.  Is there any info on the forum about this?  The search button gave me nothing.

[Pablo1234]

Ed VanHalen uses a pitch shifter with just 2-3 cents pitch shifting to simulate a bigger sound, like a static Phase shifter, you could also just do an all pass filter for really short delays

[CynicalMan]

The ADT was a recent winner of the FX-X (which we should totally start up again): http://www.diystompboxes.com/smfforum/index.php?topic=84323.msg710076#msg710076

[Jaicen_solo]

To be honest Mark, if you're close mic'ing an amp, there isn't going to be an audible delay. Sound moves at about 1ms/ft in normal air. If you have a mic 6" from the grill, that's 1/2ms.
I think it's more important to think in terms of phase relationships than delays.
The human ear has a latency of around 10ms, so anything less than that would sound instantaneous.
Most people can't distinguish anything below 25ms (stadium rock threshold) as a seperate event.
As I say, the only effect I can imagine it having is if the phase relationship was introducing some sort of filtering effect.

[Mike Burgundy]

0.5ms would mean 1kHz (and 3kHz, and..) would be perfectly out of phase - that would be audible. There's a combfilter resulting from this, plus the different phase shifts from different filter settings and topologies all add to the "wider" sound - there's a lot extra going on - subtle stuff, but it's there. The filter's phase shifts are frequency dependent too, all adding to the complexity. Based on experience I'd say a range between 0.2ms and 4ms will give you everything you'd want to play with (edit: puts it right smack into 3006 territory!). Maybe add a phase reversal switch too, then tune by ear, pan the two or more parallel lines away from each other for maximum effect. I'd say longer delays work best when panning in stereo, shorter for mono.

[Jaicen_solo]

I agree with pretty much all of that Mike, however I would like to add that although you're correct about the timing putting kHz out of phase, in reality it's not that simple.
There are other factors at work, for example the phase/frequency relationship that each amp will output, the type and number of speakers and cabs etc.
Also, the precise positioning of the listening point (in this case, two close mic's) will affect the phase cancellation. With conventional speakers, they tend to beam out high frequencies, whereas low frequencies are more omni-directional. With an open back cab, you're going to get low frequency cancellation all on its own! So, in reality, your unlikely to have any frequencies which exhibit perfect cancellation, but most of the cancellation will be at the upper end of the spectrum.

However, you're going to get a whole load of strangeness happening at low frequencies, as they'll be reinforced with strange shifting phase variances.
That may be the key to the sound that you're looking for.

I would think you need to be looking at no more than 30ms of delay range really.

[Mark Hammer]

Oh heck, I wasn't even thinking of anything nearly as long as 30msec.

Note that when people talk about mic-ing up two or more different amps for a "huge" sound, the amps are generally always selected for dirt quality.  You never see anyone talking about mic-ing up a silver face Twin and a Visual Sound Workhorse.  They've always got a Marshall and something else equally dirty-sounding.

Which says to me that the "big" sound may well be coming from the animation produced when the harmonic content is not phase-aligned.  In such a circumstance, maybe a half-millisecond  resulting from mic placement DOES make a difference.

[CynicalMan]

OK, using VST amps, I did a comparison of various delay settings. I used the Ignite Amps Anvil into Roxolder's Mercuriall Cab 2.0 as the fixed channel, and a delay into Lepou's Hybrit Head into Mercuriall Cab into an inverter as the adjustable channel. This is what I got, with the various delay settings:

http://sites.google.com/site/distorque/home/projects/sound-clips/delaycomparison.mp3
No delay - ~0.25ms - ~0.5ms - 2ms - 5ms - 7ms - 10ms - 20ms - 50ms

I like the 3rd and 4th clips the most myself. That suggests that comb filters with fundamentals at around 500-2kHz, or 0.5-2ms work well. This might just be another way of getting the V-shape EQ that people like on distortion, but with a comb filter instead of a notch filter.