Time-staggered overdrive

[Mark Hammer]

Dunlap/MXR introduced their "Dookie" pedal at NAMM this week.  It does something that Boss has been doing for a while; namely providing two parallel blendable overdrive channels governed by a master gain control.  The stated intent was to mimic the dual-amp setup that Green Day guitarist Billie Joe Armstrong uses on stage.

It occurred to me that use of two amps on stage might result in some very slight differences in ear arrival time, and perhaps some slight cancellations as a result.  Nothing major, but perhaps the basis for why using two amps at once might sound good to someone.

Forum semi-regular Thomeeque posted some layouts a while back for producing a short fixed delay.  The goal was to time-stagger a clean signal in order to produce through-zero flanging (the swept time delay has to be able to briefly move "ahead in time" of the clean/dry signal.  The circuit can use a MN3007 or more available MN3207/V3207, in addition to a 4047 to produce the clock and a 4049 to buffer it and push the 3207 a little higher than usual, for a delay time as short as 0.5msec.

I'm curious about inserting this between the output of one of a pair of different-character overdrives and a mixing stage.

[Rob Strand]

It occurred to me that use of two amps on stage might result in some very slight differences in ear arrival time, and perhaps some slight cancellations as a result.  Nothing major, but perhaps the basis for why using two amps at once might sound good to someone.

Most things like that do make it sound better.  Like stereo chorus.  Or even listening to a wall of speakers instead of a single speaker.   Part of it is delay part of it is when you have a sound source off-axis (in left/right sense) the frequency response of the left and right ear are different.   If you rub your fingers together at arm's length in front of you, then move you hand to the left and right  you can see that one ear gets more highs than the other.

Look up HRTF.   You get a similar thing when trying to make headphones sound more natural using "crossfeed simulators".   When you listen to speakers your left ear head both left and right channels but it hears each side differently.  There is also a difference in the delay.   (Visa-versa for the right ear.)

I've got this cool experiment:

Experiment 1:
- Generate some pink noise in the left channel.
- Create copy the left channel to the right channel so the signal is monaural.
- Listen to that with headphones

Experiment 2:
- Generate some pink noise in the left channel.
- Generate some pink noise in the right channel. 
   Don't copy the left channel, you must create a new sequence of noise
  so you have uncorrelated "stereo" noise.
- Listen to that with headphones

(Experiment 3: You can do the same experiment where the right channel is left channel delayed, experiment with the delay time.)

The result of experiment 2 sounds quite different even though you are listing to random noise.
It's more spacious and bigger (kind of brighter and bassier at the same time).
In experiment 1 your brain links the left and right ears and detects the correlation in the noise.

[idy]

A short fixed delay with attendant notches is a bit like a phaser in the manual setting. The kind of sounds your idea makes me think of are similar to a fixed phaser. Page with the Evantide Instant Phaser. Isn't a phase stage delay-ish in the sense that a univibe is a vibrato?
Not to knock the BBD solution,  but is the phase solution easier and better fidelity?

[Mark Hammer]

It is a simpler solution, if I knew precisely what sort of phase delay was needed, and it was short.  Thomeeque's TZF daughter board is small, reasonably uncomplicated, and provides for being able to vary the delay to taste.

[pinkjimiphoton]

i used to do this in audacity kinda all the time, was one of my recording "secrets" and it can make for one hell of a sound.
i used to record a stereo track, split it to mono, then use the alignment function to drag the left track "back" in time by a couple ms, and move the other track "forward" in time by a different amount of ms... maybe 5ms one, 13 ms or so on the other.

the result was a real curious doubling where ya heard a predelay and a delay, but no fundamental. very ear catching and standout-y

listen to the wah wah solo on this song, about 1:53 in. its very pronounced on the wah lead

https://drive.google.com/open?id=0BzMJxNNgY3_yeFRYb0wybU50cjA

or on the vocal about 1:56 in on here

https://drive.google.com/open?id=0BzMJxNNgY3_yUzNKYldkQXIySkk

its all over this one... the slide guitar and vocals especially. adding stereo flanging makes the vocals ridiculously sick and evil sounding cuzza the tricks this kind of sound plays on your ears.

https://drive.google.com/open?id=0BzMJxNNgY3_yUUI0WVR6OFk0Q0k

i like the idea of doing it in a stompbox. so i'm'a makin me some popcorn!!

[Mark Hammer]

Compelling evidence, Sir James.

Of course where what you described involves duplicating the exact same track at two points in time, what I'm proposing is staggering two tonally different versions of the guitar in time.  Again, the idea is to mimic two different-sounding amps that are not flawlessly on-axis at all time, whether in the studio or onstage.

The result could either be used in stereo, or could be mixed down to a mono blend.

[Rob Strand]

It is a simpler solution,

I suspect it's not so simple when you are pushing the sound down one channel (ie. a guitar lead) you will have to fight/trade against cancellations.  One idea is not to combine with 50% of each signal.

[Mark Hammer]

I don't know about that.  Certainly one doesn't want the equivalent of having speakers out of phase.  But I would imagine that a small portion of whatever magic comes from running two different amps IS some cancellations.  Not a lot, but some.

[pinkjimiphoton]

Compelling evidence, Sir James.

Of course where what you described involves duplicating the exact same track at two points in time, what I'm proposing is staggering two tonally different versions of the guitar in time.  Again, the idea is to mimic two different-sounding amps that are not flawlessly on-axis at all time, whether in the studio or onstage.

The result could either be used in stereo, or could be mixed down to a mono blend.

diggit, imagine that with the two different tones, if ya could make the time delay envelope dependent, say, play harder=longer time and you had each side go thru separate circuitry i think it would probably sound amazing.

think of it as using the envelope to adjust the difference between tone 1 and 2.

maybe make it as an outboard box, where ya could chain whatver effects you choose into a and b and then be able to offset them just slightly. i bet it would pop.

i didn't mean the phasing aspect as much as the way it makes the sound seem to pop out of the speakers in a really peculiar way.

but with envelope control of the delay, say from 5ms to maybe 25ms or so, you could kinda control the dominance of either by the phasing that occurs.. all from your pick attack.

oh man....


if you guys could only hear the sounds in my head

[pinkjimiphoton]

duh...
stoned AND stupid, here...

i mean what you describe would kinda sound a little like the examples i posted. a real short delay between two signals, but with NO DRY SIGNAL... so short that the slightly delayed signal is used AS the dry signal, and then the "double" maybe be generated separately. i just got this feeling that being able to control it by envelope would be really cool sounding.. an "evolving" distortion sound, if you will...

like i said... if ya could hear the sounds in my brain

[Rob Strand]

I don't know about that.  Certainly one doesn't want the equivalent of having speakers out of phase.  But I would imagine that a small portion of whatever magic comes from running two different amps IS some cancellations.  Not a lot, but some.

It's a complex process which gets more complex when you add you brain into it (like my noise experiment shows) and the room.   It's while since I've looked at the details but this should give you an idea:

The Set-up:  Suppose you have two stereo speakers L and R, and you are standing in front of them in the usual manner.   For the sake of the argument the speakers are spaced at 30deg each side of centre.

The Aim:      Analyse what one ear hears, say the right ear.

What happens to the acoustic waves:
- The sound from right speaker has an emphasized high frequency response compared to listening in front of you.  That's caused by the ear.
- The sound from left speaker has a  de-emphasized high frequency response compared to listening in front of you.  That's caused by the ear and the fact the sound diffracts around your head (think of your head as a ball).
- The sound from the left speaker is delayed (by about 220us) compared to the sound from the right speaker due to the long path length of the sound (which also includes the sound diffracting around the curve surface of your head).

What happens to the sound:
- Because the high frequencies from each channel have different high frequency emphasis there is a natural imbalance in the levels.   Off hand that covers 1.5kHz and above; not to sure of the specifics here
- The 220us delay can only cause cancellation in the high frequencies (notch at 2.2kHz and peak a 4.5kHz).
- The imbalance of hf levels prevents the delay causing the notches getting too deep.

(The high frequency emphasis can be heard with the finger rubbing experiment I mentioned before.)

The bottom line is there more going on than just a delay.   Those headphone "crossfeed simulators" try to emulate this exact effect.  It's a quite subtle effect overall.

When you consider the left ear the roles of the speakers reverse and the opposite channels get emphasized and de-emphasized.   The brain puts all that together and no channel sounds more emphasized than the other, it's kind of a flat overall impression.

The trick pinkjimiphoton uses (which is a common trick for voice) goes way beyond a 220us delay.   It's a different and more exaggerated and noticeable effect.  The longer delays push the combs closer together which prevents a large suck-out in the spectrum which would occur with the 220us delay.   

In a room you have short delays from direct signals and long delays from the room and also more complex filtering.

[Ripthorn]

One more wrinkle is that with two separate amps, you have the separate volume velocities of the cones from the speakers which you are not able to get from a single amplifier with the two pre-combined signals.  You can get close, but you do lose some effects. Some of this stuff is something that I was discussing at work with some coworkers.  I'm an acoustic physicist and I work on audio and acoustic signal processing semiconductors. There are a ton of factors, however, that doesn't mean that you can't try to get the most significant ones.

While a large amount of info is related to the HRTF, there are things you can do to at least give the impression of more space and girth to a sound. Delay with some hard panning (or at least wide panning) gets you a long way there.  I think a box with a very short delay and a dual gang pot for the panning width sending the signal out via TRS could be really interesting. Of course, I've been on a microcontroller kick lately and have toyed with the idea of using something like a Teensy 3.6 for something like this. It would be super simple and size and cost would probably be on par with discrete components.

[Mark Hammer]

Thanks, Rob.  I appreciate an analysis that takes how we hear into account.  I have several psychoacoustic "imagers", that provide a bigger stereo stage by means of a bit of delay and crossfed signals.  I recommend them highly.  Hell, I even built one from a Radio Electronics article, 30 years ago.  So, I have no illusions about what I'm planning and what it can or can't do.  Really I'm just curious about what staggering two overdrives a teensy bit in time will sound like.  It might be pointless.  We'll see.  It may depend on what sorts of voices I'm blending.  I was thinking of a mono blend, rather than anything more sophisticated or stereo crossfeeds.

I etched the board this evening, and plan to drill and stuff it tomorrow, contingent on whether the garage is too cold to work in or not.  Here's the view from the kitchen this morning.

[Rob Strand]

So, I have no illusions about what I'm planning and what it can or can't do.  Really I'm just curious about what staggering two overdrives a teensy bit in time will sound like.  It might be pointless.  We'll see.  It may depend on what sorts of voices I'm blending.  I was thinking of a mono blend, rather than anything more sophisticated or stereo crossfeeds.

It's probably easier to play around with wave files in some audio editing software (like audacity etc).  Adding delays is pretty easy.   You might need some good VST's for filtering.  I find the packaged filtering in those programs very crude and imprecise and the numbers don't translate to back to circuits so well.

I guess what I was thinking is you can use the idea of the cross-feed but apply it to one channel.  That means one overdrive would get the "bright" path and the other overdrive would get the "dull" path.  The aim being to prevent the notches from getting too deep.   Overdrives with greatly different voicings probably won't interfere with each other too badly.  The ultimate goal would be to feed the same overdrive into both arms and have it still sound thicker with no cancellation effects.

If you went with pinkjimiphoton's method that would go down a whole different route.

When I think of sound "biggeners" the Boss Dimension C always comes to mind.   A more subtle version of that wouldn't be bad.

Hell, I even built one from a Radio Electronics article, 30 years ago. 

Can you remember the article?

Here's the view from the kitchen this morning.

You guys have tougher winters than us.   The guy's in the US I worked with were always sending their chilly backyard pics.   I can just survive winter here in T-shirts and shorts.

[pinkjimiphoton]

fuzzmonger here cuttin in for a second..
biasing, or rather MISbiasing transistors i think you may find some of the parallel distortion paths can sound kinda hip.
wanna build a better fuzzbox? its tough.
but to get that double amp sound, i'd kinda think that two different kinds of dynamic fuzz
in parallel, or split by frequency (or like i said, control the frequency with the amplitude)
you could probably mimic enough outta phase overtones and get the illusion of  more than one amp tonally... the equiv of summing it to mono, obviously, unless ya kept a true stereo path, but then ya gotta use two amps anyway!
certain classes of misbiased transistors can get to be a very phasey kinda distortion and do some surprising things to your mind..
like the ludwig phase II thing where it sounds like its behind people sometimes and freaks 'em out. i've been yelled at by people for using that thing live cuz it freaks 'em out.. in that case, phasing is making it sound like the other stereo field, remember, we have not only
left and right, but front and back... you can mess with "where" the sound appears in the room. its all cuzza the way it ends up mixing back into the air.
sorry.. just random thoughts on this. the idea of a time staggered overdrive i think is possible, but i really think it would sound better dynamic than static.... let the envelope either move the delay time, or sweep the distortion level <that would make it act more like the boss i bet> or a flilter. hell, maybe all three. i think it would sound like little purple fishes swimming laughing thru my fingers as i listen in awe.. lol truly a psychedelic sound.

ya can do it with vsts after the fact, for sure... but to do it live is the thing thats gotta be done.
maybe a combo of good old analog distortion and dsp?

[deadastronaut]

look up 'haas' effect....

[bool]

L-C-R pan time-staggering were basis for some Biran May sounds, and perhaps most notably used by Boston's Scholz (and all subsequent Rockman-powered recordings) ... and it's of course the basis for all "stupid" guitar multi-tracking (L-R panned multi take recording).

[Mark Hammer]

look up 'haas' effect....

You're talking to a guy that had Albert "Auditory scene analysis" Bregman as one of his profs.  The Haas effect relies on identical sounds, staggered in time, and presented binaurally.  I'm aiming for two fairly different sounds (same note but fairly different timbre), eventually mixed down to mono.  If I was aiming for stereo, I imagine I might start to get involved with cross-feeding.

I don't remember specifically which issue, Rob, but it would have likely been in the early 1980's, when such things were popular.  And it was Radio Electronics.  I won't plow through the archive at American Radio History, but I think I have a photocopy of the article in one of my binders.  That will probably have the month and year at the bottom of the page.  Used an SAD-124.  Ah, those were the days.

For the uninitiated, the stereo image enhancers (Carver called theirs the "Sonic Hologram") worked as follows.  In the real world, off-axis sound sources (i.e., not directly in front of you) provide their own Haas effect, by arriving slightly later at one ear than the other, and with slightly less high-frequency content.  People refer to this as a "sound shadow".  The enhancer circuits extract left-minus-right and a right-minus-left difference signals. So, whatever is unique to each channel is identified.  It is then fed through a short delay and a bit of lowpass filtering to the opposite channel, where it is mixed with that channel.  The effect is to make the soundstage huge and each instrument in the mix very easily spatially locatable.  You can close your eyes and "see" this percussion instrument over there, and that guitar over here, etc.  You wouldn't necessarily notice it when it's turned on, but turn it off and your first thought is "Where did the music go?", even though it's still in stereo.

Historically, its Achilles heel was that it was introduced in an era of vinyl.  Sadly, since surface noise on vinyl is largely unique to each channel, unless the disc was pristine, the effect tended to increase the overall audibility of surface noise, by hearing it with a sound shadow on the other channel as well.  But man, when it worked, it worked really well.  In more recent years, it was implemented in the digital domain.  If you remember the SRS (Hughes Sound Retrieval System) in Windows Media Player, that used the same principle, except that the MP3 or WAV file you used it with didn't have surface noise.

[Rob Strand]

I'm aiming for two fairly different sounds (same note but fairly different timbre), eventually mixed down to mono.  If I was aiming for stereo, I imagine I might start to get involved with cross-feeding.

In the end the longer delays might sound better.

I don't remember specifically which issue, Rob, but it would have likely been in the early 1980's, when such things were popular.  And it was Radio Electronics.  I won't plow through the archive at American Radio History, but I think I have a photocopy of the article in one of my binders.  That will probably have the month and year at the bottom of the page.  Used an SAD-124.  Ah, those were the days.

No problem, thanks.  I worked it out from what you said.
"Stereo Image Expander"
pt 1 Radio Electronics, June 1982, p45
pt 2 Radio Electronics, Sept 1982, p63

https://americanradiohistory.com/Archive-Radio-Electronics/80s/1982/Radio-Electronics-1982-06.pdf
https://americanradiohistory.com/Archive-Radio-Electronics/80s/1982/Radio-Electronics-1982-09.pdf

[Mark Hammer]

I'm aiming for two fairly different sounds (same note but fairly different timbre), eventually mixed down to mono.  If I was aiming for stereo, I imagine I might start to get involved with cross-feeding.

In the end the longer delays might sound better.

Simply changing the 68pf clock cap for maybe 82pf might address that, the clock frequency will still be high enough that the existing LPF should be fine.

I don't remember specifically which issue, Rob, but it would have likely been in the early 1980's, when such things were popular.  And it was Radio Electronics.  I won't plow through the archive at American Radio History, but I think I have a photocopy of the article in one of my binders.  That will probably have the month and year at the bottom of the page.  Used an SAD-124.  Ah, those were the days.

No problem, thanks.  I worked it out from what you said.
"Stereo Image Expander"
pt 1 Radio Electronics, June 1982, p45
pt 2 Radio Electronics, Sept 1982, p63

https://americanradiohistory.com/Archive-Radio-Electronics/80s/1982/Radio-Electronics-1982-06.pdf
https://americanradiohistory.com/Archive-Radio-Electronics/80s/1982/Radio-Electronics-1982-09.pdf

That's the one.  It was okay.  The two Omnisonic units I bought worked better.