Double tracking-effect on live guitar?

The very best realistic ADT I have found is the WAVES plug-in "Abbey Road Reel ADT." It's terrific/sounds like the real thing, because it is modeled on the real one at Abbey Road..

The problem most of us have in AXE FXII -- in trying to "double" an original part a second time for the ADT effect, the second track cannot be a constant proportion to original signal (that is TOO perfect)... and choruses pitch delays etc. don't vary enough "randomly" to make it seem realistic using the usual sine waves or set times, etc.

You need modulation -- but you need it to be truly random within parameters that don't make things go stupid out of tune. That seems to be the secret.

So, you might want to set a 100% delay on one side (pan each side HARD L/R) and then have the delay time "ranges" from something like 1ms to 24ms minimum to maximum (please do experiment...try 1 to 10, 10 to 30, 5 to 20ms see what you like best tonally, there is some comb filtering in some cases) ...but create/make the delay TIME range respond to some modifier that is moving randomly (not constant ratio), such as pitch or envelope.... Unfortunately "random" replacement for sine wave LFO form does not work on this as far I can have found, but I may have missed something. But you can get quite a stereo spread this way that responds to how you play. The other tips above like Ian's can work in concert with this.
 
austinbuddy said:
The very best realistic ADT I have found is the WAVES plug-in "Abbey Road Reel ADT." It's terrific/sounds like the real thing, because it is modeled on the real one at Abbey Road..

The problem most of us have in AXE FXII -- in trying to "double" an original part a second time for the ADT effect, the second track cannot be a constant proportion to original signal (that is TOO perfect)... and choruses pitch delays etc. don't vary enough "randomly" to make it seem realistic using the usual sine waves or set times, etc.

You need modulation -- but you need it to be truly random within parameters that don't make things go stupid out of tune. That seems to be the secret.

So, you might want to set a 100% delay on one side (pan each side HARD L/R) and then have the delay time "ranges" from something like 1ms to 24ms minimum to maximum (please do experiment...try 1 to 10, 10 to 30, 5 to 20ms see what you like best tonally, there is some comb filtering in some cases) ...but create/make the delay TIME range respond to some modifier that is moving randomly (not constant ratio), such as pitch or envelope.... Unfortunately "random" replacement for sine wave LFO form does not work on this as far I can have found, but I may have missed something. But you can get quite a stereo spread this way that responds to how you play. The other tips above like Ian's can work in concert with this.
Click to expand...

And Strangely enough, today is the last day of a Sweetwater promo with WAVES and you can the Abbey Road reel ADT for only 99$..70% off!!
http://www.sweetwater.com/store/det...e=&network=g&gclid=CICMos3Ytc8CFQ4yaQodAU0B_w
 
austinbuddy said:
You need modulation -- but you need it to be truly random within parameters that don't make things go stupid out of tune. That seems to be the secret.
Click to expand...
I have reasonably good pseudo-random modulation in the parameters on the preset I created up above. I'm happy to show people how to do that.

I have an outstanding wish for a truly pseudo-random controller on the unit that's not a randomized step pattern, repeated.

But even with what we've got you can make the step controller mostly unguessable to get a random control series.
 
Let me add something important for true ADT sound. REAL ADT means the track was already laid down on tape BEFORE any ADT process.

The Engineers were playing it back through multi-head tape machines, and the second copy (modulated) was laid out from the original, to emulate a second performance.

As I understand it, that also meant a second track (because of multiple tape heads and sources you could tap) could in theory also be varied to play slightly AHEAD of the original track laid down, not just behind it. If you ever get the WAVES plug-in you can see what I mean - visually, the tape heads move both left and right over the original source track, by milliseconds.

For us live players, that is impossible to do! Because we are playing our track live, the ADT second track effect can ONLY be a delayed sound, behind our original playing.

Once we lay down a recorded track into a DAW though, in processing it is theoretically possible that the second track can be sometimes be shifted ahead and sometimes behind our original track with modulation, which is why plug-ins like REEL ADT can do this.

Now -- you truly CAN get some amazing wide doubling effects with experimentation and great guitar thickening in the AXE FX, it's killer. Ian has it right!

But to get the real "Abbey Road" Beatles type of ADT sound, that is very hard to do because of this issue (the AXE FX is amazing BUT it can't anticipate what you will play before you play it to generate that second copy modulating both before and after the notes you play!).

As I understand it, engineers look at 24ms and 28ms as magic numbers in terms of setting delay for double track. (Try it, pan dry one side, then 100% wet delay with 24 or 28ms to the other, see what you think -- hear that Beatles-esque guitar bathroom reverb sound?)

But when I look at WAVES Reel ADT plug-in, it is varying by up to 15 ms BEFORE original source track (left of original tape source head, so tapped from a tea placed before the main playback head) all the way to 15 MS AFTER track (right of tape head) -- which is basically a 28-30 ms total span...

Overall, that's why I love the WAVES REEL ADT plugin - it really emulates the Abbey Road process. (BTW I have no affiliation with WAVES).

Perhaps I''m not technically right, that's not my forte, but I wanted to lay out why this is such a challenge to nail it dead-on in AXE FX.
 
You can do the same thing by delaying both copies of the signal. If you offset the dry with a bit of delay you an place the copy before the dry in time. This is the same way through zero flanging works. The cost is latency since you are delaying both signals, but for a thick doubled type sound it's not as noticeable since the timing of everything is blurred a bit anyway.
 
(mr fender touched on some of what I was writing up for a post but I'm going to paste what I already had written)

Seems to me original DT effects were created in 2 manners. Either one original source being duplicated and then one of those sources being manually delayed or the creation of two original sources which offer delays naturally due to human characteristics.

The Waves plug is using one source and so is the Axe. Sounds like the Waves plug is mimicking the first method mentioned pushing identical signals in front and back of each other. In my brain since they are copies of each other neither could really be called 'the' original and I could see the Axe being able to replicate this method via oscillating delays on 2 channels utilizing a single source. The guitar signal.

IMO the second method has to boil down to variations in either timing or pitch and at times a combination of both. Which I could also see, in theory, the Axe being able to replicate with very slight range random delay and pitch oscillitions much like we would treat a MIDI drum track we are trying to humanize.

I think the major problem comes from trying to duplicate a dual source analog effect with single source digital equipment. I really can't see the plug capabilities being any different than the Axe in theory.
 
MIDI humanize is a different creature because you are manipulating the triggering of the note. This results in a much more natural sounding track that is more similar to a separate recorded take. It's a performance manipulation, not an audio manipulation. All artificial DT effects suffer from the limitation of using copies of only one source signal. You can manipulate the signal copy to make it more unique sounding, but it will never be exactly the same as combining two separate takes. Randomization would need to occur on a note by note basis as it would with two separate performances. Any procedural manipulation of the copy, even random ones, will not quite sound the same. Using an LFO to modulate pitch or delay times becomes more obvious during sustained notes. You can randomize the LFO to help hide any obvious cyclical pattern, but it still does not follow on a note by note basis. Timing and pitch irregularities in natural performances tend to track the note envelope and usually don't just occur at random. Ideally you would need to divide the source signal up into separate note envelopes and manipulate each of those separately for a more natural sounding double track effect. You can do this in most DAW's these days, using time stretching and pitch shifting to correct flubs in recorded takes. Many even offer time align and randomize functions similar to MIDI humanize. Doing this in real time on a live signal would be tougher though. You could perhaps combine the use of random values and envelope based triggering to make the manipulations more closely follow the natural note envelope.
 
mwd said:
IMO the second method has to boil down to variations in either timing or pitch and at times a combination of both. Which I could also see, in theory, the Axe being able to replicate with very slight range random delay and pitch oscillitions much like we would treat a MIDI drum track we are trying to humanize.
Click to expand...
The preset I used to do the clip above varies pitch, delay and frequency content of the second track. Sometimes randomly, sometimes based on an envelope. Usually using both approaches in a single block actually.

But, despite this, it doesn't change fundamentally what I'm playing. That's the ultimate failing for all automated double tracking algorithms. They can't be another human. The copies are still too perfect.

"Humanizing" something is completely different. That's where you actually introduce anomalies and alterations to the actual notes being played. You make the copy less than perfect and it actually sounds a whole lot better when paired with the original.

I think the major problem comes from trying to duplicate a dual source analog effect with single source digital equipment. I really can't see the plug capabilities being any different than the Axe in theory.
Click to expand...
I used to agree. Or at least: I used to agree that you could only do the human-izing on MIDI tracks where you had the precise note and timing data. Then Celemony came along. Their algorithms do fantastic multi-timbral and multi-pitch detection on signals to synthesize less-than-perfect copies of the original signal. It can be done for analog signals though it takes a lot of processing power and may not be useable in realtime.
 
It seems like the best way to accomplish double tracking with the Mimiq pedal without the weirdness of applying a pitch shifting after one amp's distortion stages would be to put it in the Axe's effects loop, then place two amp blocks in parallel after the pedal, one amp receiving the Mimiq's left out and the other receiving the Mimiq's right output.

In the real world with actual tube amps you'd need two preamps which is much less common, but with the Axe it's not really an issue.

Of course that would really only work in the truest sense with the pedal set to only 1 additional track. Setting the pedal to "2" (three guitars total) or "3" (4 guitars total) would change things quite a bit I'd imagine.

The idea behind the Mimiq is awesome though, and the demos seem really promising. I'd love to try one out with my setup.
 
Last edited:
APOGEE123 said:
Be nice if FAS would make an algorithm that would get you there fast
Click to expand...

Frankly, I think Cliff won't bother making this. He feels the enhancer block does the double-tracking effect well enough, which it doesn't for many people. The enhancer block does a beautiful stereo spread (similar to the TC Mimiq's "1 dub" setting of making a wide stereo spread) and I love the enhancer block for this. The enhancer block, however, can't do the "thickening" aspect of 2 or 3 "dubs" like the TC Mimiq pedal appears to be able to do. He's obsessed with amps.

When was the last time a gigantic improvement in an effect block or a new variant of an effect was implemented in the Axe? It's been a long time. Not to be a Debbie-downer, but that's not his thing these days.
 
You must log in or register to reply here.
Back
Top Bottom