Octave Splitter

[Eb7+9]

I recently came up with a new form of piece-wise non linear octaving ...
this time by doing the flip right down the middle //

it's not something I've seen elsewhere



an interesting sound, not too much unlike an ideal FW rectifier but different in its own way

very ringy ...

the design presents a few challenges, one of which is to produce clean switching at very high speeds
I'm not gonna say switching of what, but it's not hard to guess how I'm doing it ... or is it?

I'll leave it at that ... it works, it's reality at this point

those who explore the idea will likely run into one issue that is inherent in the concept
I'm curious to see how it's dealt with ...

like my Nyquist Aliaser it would be cool to see others explore it - their own way ...
and so for now, best not to cloud the imagination with my schematic

[ElectricDruid]

Ok, I'm curious. A few things about this pique my interest.

1) Why "high speed switching" is a problem. This is audio rate switching. That's not high speed for modern CMOS switches.

2) How you've got the switching signal aligned with the peak of the input waveform and not the zero crossing. I've seen a trick for this somewhere I remember, but I can't remember where or how it was done.

3) What the "issue" is that you mention. Seems to me if you've solved (2) and you clearly have, then you've got it.

[Eb7+9]

(1) is not a problem ... but rather a challenge - that is, using simple discrete circuitry

(hint: otherwise you end up loosing a bigger piece of the pie in the transition ... and distorting what's left over)

(2) I'll let you think about that a bit more

(hint: it's Calculus related ... and something I haven't seen used in any other audio signal processor)

(3) yes you're right, it basically works ... but I'd like to see what other peeps do if/when they get "there"

(hint: something that creeps in as signal dies off)


I should box it up first and do another test vid to show what I mean ...
but for now just a novel idea // in the same way the Super Full-Wave octaver was

I'll just say that the Octave Splitter is interesting in itself

but doesn't sound as pretty as my other new Octaver fix ... an offshoot of my original Green Ringer mods
demo'd here in complete absence of internal or external clipping:

[PRR]

> How you've got the switching signal aligned with the peak of the input waveform and not the zero crossing.

Differentiate. Zero-cross.

It's standard in electronic ignition (because the usual sensor is a differentiator).

[highwater]

I can't even begin to guess at the circuitry, but the waveform looks like what I'd expect from a Boss-style octave-down without the frequency divider.

[ElectricDruid]

> How you've got the switching signal aligned with the peak of the input waveform and not the zero crossing.

Differentiate. Zero-cross.

It's standard in electronic ignition (because the usual sensor is a differentiator).

Ah, yes, that'd do it. Thanks PRR. That was going to nag me!

That's nice and simple, Eb7+9. It'll go wappy when the signal gets very small and runs out into noise, since the differentiator will tend to boost any noise within its bandwidth causing a lot of false triggers on note tails. That's your problem in (3) right?

[Eb7+9]

That's nice and simple, Eb7+9. It'll go wappy when the signal gets very small and runs out into noise, since the differentiator will tend to boost any noise within its bandwidth causing a lot of false triggers on note tails. That's your problem in (3) right?

you got it Druid ...