Possible new way of clean pitch shifting??

[trad3mark]

That's cheating! You're not supposed to turn the audio into a square wave then just filter it, I mean, anyone can do that!  

yes yes, i know!! but, it is pretty effective. If a solution is good, i'm not going to argue with it. It's the old "If it aint broke, don't fix it" scenario.

[merlinb]

Well, the rocktave also makes a squarewave signal, the transistor arrangement at the end is basically a flipflop.
I"m sure you saw the synthbox thread.

Yeah, that's what I mean by octave fuzz. For a clean octave the idea is usually to invert every other cycle of the audio, so you get a sub octave but you still retain all the original harmonic/envelope information. Obviously this is harder to do.

[merlinb]

That's cheating! You're not supposed to turn the audio into a square wave then just filter it, I mean, anyone can do that!  

yes yes, i know!! but, it is pretty effective. If a solution is good, i'm not going to argue with it. It's the old "If it aint broke, don't fix it" scenario.

What would be really good is to do as you suggest but also add an envelope detect and voltage-controlled gain cell, so the filtered octave sound has the same attack and decay characteristics of the guitar.

[~arph]

Yes that would certainly improve things. One could also do several octaves up and then divide down with a counter and generate a staircase sine (see GEO) add some filtering and it would be a sine octave down with same attack/decay.

EDIT: spelling fixed

[merlinb]

Yes that would certainly improve things. One could also do several octaves up and then divide down with a counter and generate a staircase sine (see GEO) add some filtering and it would be a sine octave down with same attack/decay.

Interesting concept (I assume you were referring to this article http://www.geofex.com/Article_Folders/LFOs/psuedorandom.htm ).
Do you mean it would have the same attack/decay by using a separate envelope detector, or are you thinking of something more clever?

[Lurco]

The main reason i was doing this was a random pursuit of making a sort of home-made micropog. In terms of the octave down, there's a fair few decent ways of doing it, like a cross between the bluebox and the MXR Bass octave.
However, the octave up.... i never came across a good way to do it. I tried things like that Octup! but i could never get it to shift. and as for things that use transformers, they're a bit too big, bulky and awkward for my liking.

I too am fascinated by pitch shifting, although I would say octave-up is easier to do than octave down. The best method I know of is to phase shift the audio signal by 90 degrees, then mix it back with the original. This gives you a new signal of twice the frequency, half the amplitude, from the trigonometric identity:
2sin(f)cos(f) = sin(2f)

I have a schem from an 80s electronics magazine (not handy right now) for a very effective clean octave up pedal, although it uses quite a few opamps to do the 90 degree phase shifting and also some compression to keep the signals of similar amplitude. I expect there is a simpler way to do the same thing, with a little inventiveness.

Recently I have become interested in octave down, which is much harder to do without tracking problems. I am currently building Anderton's Rocktave as an experiment, since I could find no sound clips of it. I also have some ideas for my own octave down circuit, which will have to wait until I've finished the Rocktave. However, from experimentation I have found that it is virtually impossible to get perfect, clean tracking using simple analog circuitry, because the second harmonic from the string changes phase relative to the fundamental, after you pluck the string, which always causes a momentary false track. Avoiding this requires some overly-complex circuitry, like frequency dependant filters.

(I must say I think the Bluebox is an awful sounding pedal. I mean, if you're willing to go with octave down fuzz instead of clean, then you should be able to get very good tracking without trying very hard, but the bluebox just sounds like random spluttering!)

Has anyone built the original Shocktave? http://www.diystompboxes.com/analogalchemy/sch/shocktave.html
I couldn't find any soundclips.

Must have been this one: http://www.experimentalistsanonymous.com/diy/Schematics/Vibrato%20and%20Pitch%20Shift/Guitar%20Frequency%20doubleur.pdf

[merlinb]

Must have been this one: http://www.experimentalistsanonymous.com/diy/Schematics/Vibrato%20and%20Pitch%20Shift/Guitar%20Frequency%20doubleur.pdf

That's the one! (My copy is in English though of course)

[trad3mark]

that has an awful lot of opamps though. how disheartening!

back to looking for fun things to do with 555's....

[~arph]

Interesting concept (I assume you were referring to this article http://www.geofex.com/Article_Folders/LFOs/psuedorandom.htm ).
Do you mean it would have the same attack/decay by using a separate envelope detector, or are you thinking of something more clever?

Yes I was referring to that article, esp. the part about the walking ring counter. I would split the input, one end going into the envelope detection the other into the octaving section. Quick calculation tells me that if you make a staircase sine out of eight steps using a 4015. you need to quadruple your input frequency to get a staircase sine of one octave below the original. (Of course if you reduce the sine steps there are interesting things happening as you will get offset notes). Then apply the original envelope to the octave sound and presto.

But I'm sure someone allready thought of all this.

[earthtonesaudio]

I experimented a little bit with those digital sine approximation techniques like in the GEO article.  There are some really interesting sounds in there if you make them audio frequency.  You could make a "tracking ring modulator" kind of sound, multiple-crossover distortion sounds, etc.

[~arph]

I"m tempted to try this too. Problem is that I am building my binson echrorec like delay. Plus working on a pentode compressor. So no time in the near future.

EDIT:

R.G. Already posted some leads how to do this, five years ago 

http://www.diystompboxes.com/smfforum/index.php?topic=33027.0

[slacker]

But I'm sure someone allready thought of all this.

Do a search for electrax guitar synth, the schematic is on here somewhere. That uses a CD4046 and a divider chip as a frequency multiplier that could be used as the basis for a digital sine generator. It's also got a pretty cool looking octave down section.
I've got the main part of the input section on the breadboard at the moment and it works really well, the tracking is excellent, and if you want it to it will hold a note practically forever. It's a really clever design. I've simplified the CD4046 section, with a CD4024 divider in the loop, and then used a resistor DAC off that to give a digital sawtooth that can be anything from 2 octaves down to 2 up. I've then used some pretty standard analogue synth waveshapers to get triangle and sine outputs.
It's going to be used with my modular synth, but it would make a decent pedal in its own right.

[merlinb]

Do a search for electrax guitar synth, the schematic is on here somewhere.
the tracking is excellent, and if you want it to it will hold a note practically forever. It's a really clever design.

Here's the electrax and simplified version:
http://www.diystompboxes.com/smfforum/index.php?topic=55644.0

I would love to hear some sound clips of your design (and see the design itself of course too!). I am always on the lookout for examples of how well octave pedals track. Too many people cheat and don't show you what happens as the note decays.

[~arph]

I don't think that is the simplified versien slacker is talking about. It still needs the PLL to get a multiplied frequency. The slacktave only does squarewave octave down.

[merlinb]

I don't think that is the simplified versien slacker is talking about.

I know that; the linked thread includes both versions.

[slacker]

Yeah what I'm playing with at the moment is the actual electrax circuit. The input processor on page 14 and a modified version of the harmonic generator on page 15. The schematic below is basically what I'm using, except I've got the JFet part of the original in there as well, so it can hold notes.

If you stuck this in place of the CD4024 in the slacktave (input to pin 14) that would give you something to generate your digital sine wave with. You could use pin 4 of the CD4046 to clock RG's walking ring counter and then depending on which of the CD4024 Q outputs you connected back to the CD4046 you could move the sine wave up and down octaves. Or you can just tap straight off the Q outputs and get a range of square wave up and down octaves at the same time.

[faptdivers]

Did this project die out?