Question: Does more random modulation improve modulated delays?

[Mark Hammer]

A post on another forum got me thinking this morning.

More and more commercial delays and delay projects are coming with modulation capabilities; typically in the form of a pair of Rate/Depth controls, and typically using a dual op-amp LFO to produce a triangular waveform.

But I got to thinking that maybe the tape transport systems such designs try to emulate the wow and flutter of aren't THAT consistent.  Or maybe they are.  I don't know.  Are there commercial devices, or homebrew experiments, that provide aperiodic or "random" modulation of delay?  And if there are, does it sound discernibly better or more realistic than a standard dual-op-amp LFO?

I'm wondering if there would be any value added by running dual LFOs to get something a little more aperiodic.  Not committed to it, just wondering.

[Groovenut]

I have often wondered the same. It would seem that the tape drag on the reels would tend to make the speed fluctuation somewhat random. Maybe only 10-20% though. So is there an easy way to incorporate a 10-20% sweep/drift in LFO frequency to account for this in modulated delay designs?

[robthequiet]

Yes. I vote for aperiodic patterns. Back when Trower and Hendrix made the univibe central to their sound, I found it a bit annoying after a bit. Too repetitive, although the licks were epic.

On the other hand, the Leslie sweeps were far more interesting, unless a tech decided to put a fixed mic in front of the horn  -- in churches where you had dual Leslie cabs, the angels sang along, or so it sounded.

One weird variation was the ANRS noise reduction in JVC cassette players -- if you slightly overcranked the record level and did not decode, esp on another deck, you'd get a weird sweep that you either loved or hated. Not sure if you got the same effect from dolby or dbx.

[Mark Hammer]

I'm not sure what you mean by 10-20%.  Do you mean the depth of any modulation, or are you implying some semi-quantifiable degree of semi-randomness?

Mixing two LFOs, running unsynced (and they can even be at identical frequencies) creates something that, while not purely random, is aperiodic enough to not sound like a predictable up/down wobble.

But the question is whether such less repetitive modulation sounds more like a tape machine.  I.E., is it worth the effort to tinker with this, or would it the sort of thing one doesn't even notice?

[Ice-9]

Yes, I think if you can implement a wow and flutter then the randomness can really improve the sound, a static modulation will tire the ears very quickly. Using two LFO's that counteract with each other will produce a random modulation and the modulation should sound more natural.

A wow should be slower than the flutter and the intermodulation of the two make the random pattern, easy to do in digital

[Groovenut]

I'm not sure what you mean by 10-20%.  Do you mean the depth of any modulation, or are you implying some semi-quantifiable degree of semi-randomness?

Yes, implying some semi-quantifiable degree of semi-randomness

[amz-fx]

A semi-tone is about 6% so you probably want to keep modulation way below that to avoid sounding like vibrato. Cheap consumer decks are probably going to wobble less than 0.2% unless they are defective.

It all depends on what effect you are trying to achieve.

regards, Jack

[Groovenut]

So we might be looking for 1% variation in LFO speed and 10-20% variation in waveform (randomness), yes?

[Scruffie]

I've tried a dual LFO type set up with similar thoughts, I found it to get very irritating in the end, a single triangular LFO just isn't as in your face.

Envelope modulation is a little more natural.

[Transmogrifox]

I have done some DSP experimentation with envelope modulation and various modulation waveforms.

Something that happened by accident is my power source feeding the delay time pot was really noisy and was so bad it sounded like trash.  I added a 2-pole digital filter to it to get all the noise down into the <10 Hz range and the result was something rather pleasant sounding -- basically low-pass noise modulating delay time.  I have since fixed the "problem" but may re-implement it on purpose (the on purpose part means I can switch it off if I don't wish to use it).

A semi-tone is about 6% so you probably want to keep modulation way below that to avoid sounding like vibrato.

Sinusoidal modulation of delay times at barely-discernable levels also adds a nice flavor to the delay sound.  I find this combined with envelope modulation could be worth your tinkering time.

One idea would be to copy the Maestro Sample/Hold effect noise generator but instead of S/H (although that could be something interesting in its own right), follow the noise generator output with some low-pass filtering to put most of the noise down into the <20 Hz band so you don't get audio-frequency noise modulation out of it.

[blackieNYC]

I built RG's pseudo random LFO.  2 or 3?  I found settings that were "off" from each other in such a way as to sound very random. Many settings don't. I imagine the speeds are as unharmonic as possible. A flatter fifth in single digits perhaps? I don't know.
This was used in the sweep of the Escobedo PWM. I would also like to hear it in a phaser or MS-20 filter. But in pitch modulation? I would expect to turn green and puke.

[Mark Hammer]

Well, to be fair, when modulating phasers and filters, you want substantive modulation.  But  if the goal is to mimic a tape transport system that is trying very hard to be stable and is coming up juuuuussssttt a little short, you're not going to use all that much modulation depth.  Which is more or less  why I asked the original question.  If one is only using just a smidgen of modulation, and trying to explicitly avoid boinginess, does randomness or aperiodicity matter?  Or is the application going to be so subtle that dickering with waveform is overkill?

[blackieNYC]

I see.  Like Ice-9 said, you need to explore the nature of wow and flutter. (Wow=low speed LFO, Flutter = higher speed? )  At least you'll know what to label the controls!

My two cents, even less sense: If there is a misaligned guide in the tape path, an unevenly worn head - these could cause a tape speed variation that would definitely be periodic - maybe siney, maybe more duty-cycley. I'm extrapolating from some experience here, but one scenario is that the tape gradually succumbs to a force pushing it "up" out of alignment until it maxes out in some other part of the tape path which forces it to pop back down, repeat. A badly bent reel could be another periodic modulation. But not sinusoidal, as most folks have visibly observed at some point.
Age or "budgetary-related" deviations in capstan speed or motor drivers or tape inconsistencies might be a little random - not sure about that one.
I guess I would expect tape speed misbehavior to be of the periodic variety.

[robthequiet]

Well, after pondering a bit, it does make sense to have some modulators at different speeds, -- i.e. periodic at feed reel rate such as bent flange, uneven wrap, rusty spindles, etc., or aperiodic with tape flap, power sag, tape stretch, etc. etc. etc. ...

The trick, to me, is to keep a single signal path, as your tape is likely to have one contact point with the PB head. So, mild vibrato, but maybe affected by two different warp engines? I guess a timbre modulation would count, as well.

[StephenGiles]

Mark, there was a magazine article I found once for a phaser where the outputs of 3 random noise generators were rectified, mixed and applied to the FETs to provide a very unusual phasing sweep. Could it have been Hobby Electronics I wonder?  Was it called Pseudo Random Phaser or something similar?

The Aphex "bounce" circuit from their Instant Phaser would help too.

http://img.photobucket.com/albums/v317/StephenGiles/Etide_contrl.jpg

[TejfolvonDanone]

Noise source to a LPF with corner frequency around 5-10 Hz as a modulation source is really great in a chorus. I did something like this as a VST long time ago (in a galaxy far far away) and it sounded rather pleasant.

[knutolai]

I prefer random modulation for most instances as you wont have the cyclical quality of the modulation to latch onto. Makes it more fluent and natural somehow.

[digi2t]

Son of Stormtide Flanger. I believe it uses a sample/hold circuit to "warble" the LFO. Maybe that can be applied here?

[Scruffie]

Son of Stormtide Flanger. I believe it uses a sample/hold circuit to "warble" the LFO. Maybe that can be applied here?

I believe in flangers and phasers that modulation type is called 'step' phasing or flanging.

Something derived of that might be as close as analog will allow but it might also be too random, I mean, a tape wow and flutter does vary but is also to a certain extent repetitive depending on what part of its rotation is causing the wow and flutter.

[pinkjimiphoton]

well, my echoplex says HELL yeah. modulation tho makes me think of two things... if say, ya wanna emu a tape echo.

first, is wow. long slow very shallow modulation
second, flutter... very random faster modulations.

two or more lfo's would be crazy good for that.

i still think all modulation should be envelope triggered... like, play stacatto and every new note triggers the lfo, and play legato and it just acts normally. but that's me, and a whole different thing.

i think having multi modulation capabilities could most definitely be exploited into some really good things.