overdub pedal!?!?

[Sir_Ian]

So I was thinking.....and luckily I didn't hurt myself or anything....

Well, I love the sound of overdub (who doesn't). Billy Corgan is said to have overdubbed his guitar 40 times for "Soma." So is there any pedal out there that would give kind of a live overdub sound. I've never seen nor heard of anything really. I'm not even a guitar player...but work was boring and I was just thinking about this.

I imagine to get the sound, you'd have to split the signal into separate channels and delay each one just a tiny bit and mash them back together.  I know there are chorus and delay pedals out there (i'm not a complete idiot), but neither of them is really the same. A chorus is just one delay and it varies(which isn't the same sound you get when its overdubbed), and a delay just doesn't sound like what I'm talking about. What I'm thinking is more like your one "dry" signal and a couple of "wet" signals that are just very slightly delayed from the original. (each wet signal would probably have to be derived from a separate delay line.)

Even then....this still might not sound the same as overdubbing in a studio, because in a studio, you can overdub with different guitars, amps, effects, compression levels, etc. But anyways...has anyone else ever thought about this?

On another note.....I bought the "art of electronics" after seeing many people recommend it, and it just came in the mail today. So if I seem like a dope, its probably because I am one. But hopefully I'll get smarter.

[newfish]

Maybe I'm being simplistic here, but wouldn't a very short delay / reverb (with minimal repeats) with wet / dry outs do something similar?

Particularly if you put the wet / dry outs through different EQ...

[willie johnson]

Boss (and many others) make something called The Loop Station. These pedals allow you to record a guitar part by stepping on the pedal when you begin playing and stepping on it again when your done. Then, when you step on the pedal agian, it plays back the part you originally played and you can play (and record) another guitar part. After that when you step on the pedal again, it plays back both parts while you play another, etc. You can continue to do this as many times as you want. And, it will store the parts if you want to. This is a very cool pedal. I often use it in live situations to play harmony lead parts ala The Allman Brothers (you can have one part of the harmony already in the pedal, then just step on the pedal and play along with it.) If you are the only guitar player in the band, you can use the pedal to record the rhythm part as you play it and then when you play your lead part, the pedal will back you up on rhythm so that the bottom doesn't fall out while you are playing the lead part.

[Boogdish]

this might garner some ideas for you:
http://en.wikipedia.org/wiki/Automatic_double_tracking

[Mark Hammer]

I might suggest that you use a pair of choruses in tandem with a splitter/mixer, and ever-so-slightly bump the delay range over for one or both of them.  ADT generally adopts a longer delay range than chorussing (and certainly than flanging).  Simply bumping the delay range over for a single chorus won't do it, however, since the periodicity will be front and center (i.e., the pitch goes up and down and up and down and....).  The illusion of multiple players is much more effectively achieved if periodicity is reduced.  The Dimension-C chorus is one way, but I think a better way is to use two unsynchronized chorus pedals that will never (or rarely) produce the same pattern twice.

Probably the cheapest route these days is to grab a couple of FAB Chorus pedals somewhere for $20@, and tack on an extra small-value cap tot he clock-cap in one of them to bump its delay range over.  This will be a small-value cap (under 470pf) on the board nearest the MN3102 or its equivalent.  If that cap was, say 220pf, add a 33pf-47pf cap in parallel to lengthen the delay time.  I realize the FAB units use SMD components, but they should be recognizable.  You'll want to keep the depth settings quite modest.

Feed each chorus with the same signal (or a slightly modified one if you like), and blend their outputs together.  Will it be instant Abbey Road?  I don't know, but it will be decidedly thicker and less annoying than any single chorus, and for a reasonable cost too.

[MikeH]

Mark brings up what I think is the key point: parallel effects.  A 4 way splitter/mixer will allow you to run 4  different delay/chorus combinations with ever-so slightly different settings or eqs or delay times, and I imagine that would literally, sound like a chorus of guitars.

[Ben N]

I would combine what Mark and Newfish suggest: A very short stereo delay (wet+dry, or wet+wet with a third dry line), with one or both sides running through some kind of unsynchronized modulation and eq so the overall tone is different and there is a bit of detuning going on in ways unpredictable (hopefully) to the ear. The either play them in stereo or mix them back together. Or just do what Mark said.

[SteveB]

No pedal can match that sound on the fly, in real time. Now, as Willie said, you can prerecord parts & play on top of them. I do this quite often.

It seems that something could be designed to sound similar to manual double tracking. How about a delay with the delay time controlled by a variable oscillator? I know many delay pedals feature modulation, but they are just bending the pitch of the delayed signal, not altering the delay time and pitch. I seem to remember playing though a rack mounted delay in the '80's that did speed up & slow down the repeats.

Okay, say you get that kind of delay modulation happening, you will still end up a bit of mechanical sounding because of the even rate of the modulation. But, what about throwing in random or sample-hold variables to control the oscillator? The delay speed & pitch change would be more random & hopefully more like real double tracking. You can hear what it would sound like by having someone play while you twiddle the delay time slightly & randomly on the delay pedal. Of course, it has to be a delay pedal that changes pitch as it speeds up & slows down, not the digital stuttering type.

Steve

[Mark Hammer]

Anything that modulates the delay time will alter pitch.  Indeed, that is what produces the infamous "pitch wobble" that some folks detest about chorus pedals.  That being said, on my old blue MXR Digital Delay, it is possible to apply modulation to any of the delay ranges right up to the 320msec range.  Adding just a tiny bit of modulation to something delayed between roughly 20 and 40msec (as the basic fixed delay) produces that pleasing Pat Metheny type doubling.

Your inclination towards "random" modulation is precisely why I proposed a couple of unsynchronized chorus pedals.  Not true randomness, I agree, but the goal of both your approach and my own is aperiodicity, and two or more unsynchronized LFOs will do that. 

I might point out that there are some chorus effects on SS Marshall amps that use two LFOs to provide a sort of random modulation, but that's semi-random modulation of one delay.  The pitch wobble may be unpredictable but it will still be audible.  Use of several parallel delays being modulated differentially, however, will create the impression of multiple players with no particular pattern of sharpness or flatness overall. 

The Boss Dimension C (now available in clone form from Behringer and positively reviewed by folks here) uses two paralleled BBDs, but they are counterswept in synchrony.  That produces a very pleasing chorus that does not appear to be either sharp or flat (because one of the BBDs always is), but does not produce the impression of double-tracking.  Partly because they are synced, but also because the delay range is very short and identical.  Having two staggered delay ranges and asynchronous sweeps will help a lot.

[Nasse]

one fabulous four made a song about a street named after their fav studio´s fav effect (cloned version only sorry)
http://www.youtube.com/watch?v=gAoJJDuCEgI

[SteveB]

Anything that modulates the delay time will alter pitch.  Indeed, that is what produces the infamous "pitch wobble" that some folks detest about chorus pedals.  That being said, on my old blue MXR Digital Delay, it is possible to apply modulation to any of the delay ranges right up to the 320msec range.  Adding just a tiny bit of modulation to something delayed between roughly 20 and 40msec (as the basic fixed delay) produces that pleasing Pat Metheny type doubling.

Your inclination towards "random" modulation is precisely why I proposed a couple of unsynchronized chorus pedals.  Not true randomness, I agree, but the goal of both your approach and my own is aperiodicity, and two or more unsynchronized LFOs will do that. 

I might point out that there are some chorus effects on SS Marshall amps that use two LFOs to provide a sort of random modulation, but that's semi-random modulation of one delay.  The pitch wobble may be unpredictable but it will still be audible.  Use of several parallel delays being modulated differentially, however, will create the impression of multiple players with no particular pattern of sharpness or flatness overall. 

The Boss Dimension C (now available in clone form from Behringer and positively reviewed by folks here) uses two paralleled BBDs, but they are counterswept in synchrony.  That produces a very pleasing chorus that does not appear to be either sharp or flat (because one of the BBDs always is), but does not produce the impression of double-tracking.  Partly because they are synced, but also because the delay range is very short and identical.  Having two staggered delay ranges and asynchronous sweeps will help a lot.

Controlled randomness. There is a random setting on the Line 6 Liqua-Flange & that's where I was coming from.

Steve

[Sir_Ian]

Thanks for the input guys.

Sounds like the best way to make a "custom" pedal that would do this would be to add onto the General Guitar Gadgets "Parallelyzer."

I imagine it would have a straight "dry" mix, and say four loops out and back. And before the four "wet" mixes get mashed back with the dry signal, either have 4 delays, or you could custom build four delays in the box. Use four bucket brigade chips. I have yet to use bucket brigades and build a delay, but what I'm saying would be simpler than a chorus or an echo. Because each chip would would just do one delay, you wouldn't have to mix it back with the dry signal (because the parrallelyzer is already doing this), and you wouldn't need LFOs, because it would be just a set delay, and lastly you wouldn't need feedback, because you don't want echoes.

Maybe I'm being confusing, so let me give a better example. First you have your dry mix. Then you have paths, 1 through 4. You could change the delay in each parallel path with a pot to a certain set delay. So say...set path 1 to a 50ms delay, path 2 to a 130ms delay, path 3 to 260ms delay, and path 4 to a 300ms delay. And those would all bet blended back with the dry. This would be the closest (I think) to real time overdubbing. Its kind of what is described in the wikipedia article linked to by Boogdish on "Automatic Double Tracking."

I think this would sound more like overdubbing than using choruses in parallel, because as discussed, choruses give a pitch shift as they speed up and slow down. And I guess I'm thinking that when a good guitarist is playing, each of the tracks is pretty much in sync with itself. By which I mean that a guitarist is staying in time with the beat, with quarter notes being quarter notes, whole notes being whole notes. Then when each "in sync" track gets mixed, they are just each starting a little bit off from each other. Each part ISN'T acting like a chorus and speeding up and slowing down. And maybe each track when recorded has a little different dynamics, harder attack, more distortion, etc, and that would be simulated by using different pedals in the loop outs of each parallelyzer loop.

This seems kind of complicated, but I was thinking about someday building a the parallelyzer for the heck of it. If I do, maybe I'd try and incorporate this into it. Feedback?

[Mark Hammer]

Fixed delays will sound like a bad reverb.  You want to have some modulation in there, just not enough to produce obvious pitch changes.  Indeed, when people overdub, there is no constant lag between the recorded track and the overdub, since they may try and anticipate what they have to play.  That's not to say they may play it earlier than the recorded track, but they sure as hell won't play it consistently 12msec late.

Do you need more than two modulated delayed versions?  Possibly not.  Three, I can see, but 4 will likely step over the line to diminishing returns.  That's why I suggest 2 for starters.

Also note that when two signals being clocked at different rates get mixed, there is a risk of heterodyning.  Three just increases the risk.

[Sir_Ian]

Fixed delays will sound like a bad reverb.  You want to have some modulation in there, just not enough to produce obvious pitch changes.  Indeed, when people overdub, there is no constant lag between the recorded track and the overdub, since they may try and anticipate what they have to play.  That's not to say they may play it earlier than the recorded track, but they sure as hell won't play it consistently 12msec late.

Do you need more than two modulated delayed versions?  Possibly not.  Three, I can see, but 4 will likely step over the line to diminishing returns.  That's why I suggest 2 for starters.

Also note that when two signals being clocked at different rates get mixed, there is a risk of heterodyning.  Three just increases the risk.

Ok, I see why you would want some modulation, that makes sense. As for heterodyning...forgive my ignorance, but I thought you only get that when you multiply signals (like through a diode mixer) and not when you just add them. Am I right? Because, I used to use a mixing console for our high school musicals and it just adds all the signals together...no heterodyning right?!? And the parallelyzer is analogous to a mixing console.

And so maybe my plan would have to add LFOs to each path, but I would want a small sweep.

[Minion]

Maybe Play through 40 amps at once ... that might work , but I would hate to be yer Rodie

[Mark Hammer]

Ok, I see why you would want some modulation, that makes sense. As for heterodyning...forgive my ignorance, but I thought you only get that when you multiply signals (like through a diode mixer) and not when you just add them. Am I right? Because, I used to use a mixing console for our high school musicals and it just adds all the signals together...no heterodyning right?!? And the parallelyzer is analogous to a mixing console.

The clock may not always be audible, but make no mistake that there is a clock signal in there and the audio signal is chopped up into several thousand little pieces with a (thankfully) inaudible stairstep character.  When the chopping patterns of two clocked signals converge, they CAN produce sideband products just like a ring modulator: the sum and difference.  Obviously the sum of two clocks up in the tens of thousands will be above the hearing range.  But the difference between a 12khz and 11khz clock signal will be well within hearing range.  This is one of the reasons why some people will report irritating noise when using a digital effect with a switched-power-supply external adaptor; both are using clocks and generating artifacts when they collide.

The heterodyning is not guaranteed to happen but it can, so just a heads up.

And so maybe my plan would have to add LFOs to each path, but I would want a small sweep.

That's exactly what you want.  Just enough so the delay time is not constant.  So, for instance, where a "normal" chorus effect might involve sweeping between 4 and 12msec (a 3:1 ratio), the modulation you want might involve sweeping from 15 to 20msec (a 1.3:1 ratio).

[ninjaaron]

I don't know about the monkey circuit (except that I've used it, and I do believe the speaker sim is on when the pedal is off. I'm A/B'd it and there was a big difference), but I do know that ROG has a great cab sim circuit.

[Sir_Ian]

The clock may not always be audible, but make no mistake that there is a clock signal in there and the audio signal is chopped up into several thousand little pieces with a (thankfully) inaudible stairstep character.  When the chopping patterns of two clocked signals converge, they CAN produce sideband products just like a ring modulator: the sum and difference.  Obviously the sum of two clocks up in the tens of thousands will be above the hearing range.  But the difference between a 12khz and 11khz clock signal will be well within hearing range.  This is one of the reasons why some people will report irritating noise when using a digital effect with a switched-power-supply external adaptor; both are using clocks and generating artifacts when they collide.

If I understand what your saying right then that means that I wouldn't have to worry about sidebands and bad frequencies if I didn't use clocked frequencies. Hence why a normal mixing board or just the  straight parallelyzer doesn't have this problem. However, once I start running delays with clocks in them...I could "possibly" run into problems.

Say I ran 5 paths in parallel with 4 of them delayed. What would you say the chances of heterodyning were? Do you get this same problem if you run effects in series and parallel? And lastly is it the LFO frequency themselves that are the cause of them problem or the delay chip? That last question isn't super clear....what I mean is....would I face a chance of heterodyning if I ran 100% analog pedals such as phasers in parallel, or do I only run into the problem with a digital chip?

Thanks for the help and ideas. Much appreciated.

[Mark Hammer]

A clock is ine EVERY delay device, whether analog or digital.  Do not confuse it with the LFO.  The clock is on continuously and the LFO simply modulates it.  The good thing about digital effects is that the clock is often a very high frequency such that sum and difference are often well out of hearing range.

[liquids]

Ive tried some of this in the past...as a counterpoint to what has been said

I think the truest answer, live, is mixing (and micing) two+ amps with different tones.   The more variables the better - different speakers, different preamps, different EQ curves, different gain levels, etc.  Lots of guys do this.  You really never 'copy' a mic track and paste it down again.  Guys I know that use the studio layer the same part over with different guitars, different amps, mixing and thickening up the timbre to get a big sound, combined with the inevitable miliseconds of different in timing.   Likewise, though, too much of that, too many such tracks creates mud, so it's done tastefully.

Similarly, I find modulation and 'chorus' to be muddying to the sound live, rather than on tape, and doesn't create the desired effect of bigger unless in stereo, and even that is debatable (see Eric Johnson's clean tone). A few cents of constant, non-modulating pitch shift (possibly more complicated than its worth) is something I've heard Van Halen uses live, possibly Metheny as well, as an alternative to chorus, for a 'bigger' sound.  Modulation and the original "layered tracks" sounds heard on a record are very different to me.  I do admit that I generally hate 'chorus' 

That being said, most of us use one amp at a time on any gig.  Given that limitation, I think it could still be be cool to get an one box that combined functions of a splitter/mixer and maybe even add a subtle 'delay.'  Something that would take your signal before it goes to your amp/in your FX loop (placement might be subjective), splits the signal into 2 or more (keeping phase continuity of course), and runs one side as 'straight' tone and then another with EQ shaping to your liking, and then also offers a few MS of 'delay' to the one as an option,  , then returns the two (or more) signals to a mix/blend of the EQd signal with the 'straight' one at varying degrees for thickness.  It could work well... 

Anyone try this kind of thing?