Octave down

[Kipper4]

Hey guys.
I'm thinking of building an octave down jobby.
What's reliable no matter where I play on the neck.
Reccomendations please.
Thanks
Rich

[EBK]

To do reliable all over the neck, you may have to do either a really complicated crossfaded multiple delay circuit (to sample at one clock rate and seamlessly play back at half the clock rate) or go the DSP route.

Otherwise, I like the OC-2 circuit for a quick and dirty octave down.  For some good in depth info on this sort of thing, check out Freppo's CMOS workshops:
https://www.parasitstudio.se/articles.html

Merlin has something interesting on this subject too:
http://www.valvewizard.co.uk/uboat.html

Ok, one more bit of reading material for you:  There is a Stompboxology article, which you can find through a quick Google search for "stompboxology octave".  Someone else can provide a clean link, perhaps

[maartendh]

Built several oct downs, Chopped OC-2 is very good. Also Anderton's Rocktave.

[KarenColumbo]

http://moosapotamus.net/stompboxology.html

[ElectricDruid]

I've never built one, but what I always liked about the Rocktave design was the use of a 570/571 compander to re-introduce some of the dynamics that the division process removes. That's a neat trick to not only make the sound a bit more interesting, but also to hide that spattery decay you can get with things like this. The rest of it is more standard stuff, much like the OC-2.

[Mark Hammer]

If you want "reliable all over the neck", Rich, then you need to use medium-to-heavy gauge flatwound strings that keep their pitch, with no wobble or bending, and have minimal harmonic content to trick the flip-flop.

The weakness is not in the circuit; it's in what we force them to contend with, and the unreasonable demands we make.

Having said that, as threshold-dependent devices, flip-flops require a dependable signal such that the threshold can be set appropriately.  One can use actual compression ahead of the flip-flop, or simply diode-based hard limiting.

The old EHX Graphic Fuzz used a sort of compression/expansion strategy to feed the distortion section and re-expand the signal to restore original dynamics.  I wonder if a similar strategy could be deployed to yield a more reliable octave, but still have some feel to it.

[Kipper4]

I saw a guy on a tube with what sounds like an auto wah and a pog or some kinda octave down and it sounded good to me. So I thought I'd look into the possibilitys.

Thanks guys
Rich

Too busy to build right now, maybe in a couple of weeks or so.

[Kipper4]

https://www.facebook.com/groups/909398102509055/permalink/1728493407266183/

[vigilante397]

Honestly I'd vote for a POG You can usually get a used one for a reasonable price, and it's really hard to get close to the POG sound with analog. Merlin's U-Boat is a really neat clean sub-octave and sounds incredible coupled with a fuzz factory (or similar), but the tracking is hit-or-miss.

[ElectricDruid]

Yeah, have to say, don't try and compete with digital pedals with analog pedals when it comes to octave-up/octave-down. You'll lose, for sure. With a lot of effort, you might get close(ish), but the kind of resources that something like POG can throw at the problem are in a different league to what you can manage in an analog circuit. Try doing an FFT with analog!

That doesn't mean it isn't fun or worthwhile. Just that you're not playing the same game as the digital pedals.

T.

[Kipper4]

Hey Guys
Thanks against I have to say when I've attempted octavers in the past I've been very disappointed with the results and the need for a special technique.
I do own a zoom multi effect that has a much more reliable (digital) pitch shifter.

FFT ?  Whassat please?

Is there anything that could be done with one of your trusty pic chips Tom ?

Bedtime. It's been a long one.

Rich.

[EBK]

FFT ?  Whassat please?

Is there anything that could be done with one of your trusty pic chips Tom ?

Fast Fourier Transform.  Let's you think in terms of frequency buckets instead of time. 
Could Tom do a reliable pitch shifter with his PICs?  Tom is a wizard, but alas, I'd almost bet a fortune against it.

Some have done a decent attempt with the FV-1, but there are annoying artifacts, like pulsing from phase mismatch as the program crossfades from one memory bank to the other.

To do it well, you need a phase vocoder.....

[ElectricDruid]

Fast Fourier Transform.  Let's you think in terms of frequency buckets instead of time.

That's the one. Can be handy if you're trying to work out what an incoming pitch is.

Could Tom do a reliable pitch shifter with his PICs?  Tom is a wizard, but alas, I'd almost bet a fortune against it.

I think you've got safe money there...a pitch shifter on a PIC is a big ask at anything beyond "toy" quality. It might be possible to manage something like that god-awful HT-whatsit "robot voice" chip that sounds like it's broken even when it's working, but I wouldn't personally hope for much better than that. It does sound like a challenge though, doesn't it?
dsPIC is a different story. I've heard some very accomplished pitch shifting algorithms on dsPIC, even with Suzanne Vega's "Tom's diner" vocals, but a 16-bit processor with DSP functions offers a lot more power.

Tom

[Mark Hammer]

I bought a dbx 100 "Boombox" subharmonic synthesizer in 2001, but never really used it much after taking it apart to see what was in there.  As expected, it uses a 4013 flip-flop, but a LOT more lowpass filtering than we would normally use for an octave down, since its function was to isolate the stuff that was low already and make it  "boom" by adding an octave down.  That would include the "thunka-thunka-thunka" of disco kick-drum beats.

[bluebunny]

Tom is a wizard

Aside: Tom is a druid.  Merlin is a wizard.

As you were...

[samhay]

Fast Fourier Transform.  Let's you think in terms of frequency buckets instead of time.

That's the one. Can be handy if you're trying to work out what an incoming pitch is.

Could Tom do a reliable pitch shifter with his PICs?  Tom is a wizard, but alas, I'd almost bet a fortune against it.

I think you've got safe money there...a pitch shifter on a PIC is a big ask at anything beyond "toy" quality...
dsPIC is a different story. I've heard some very accomplished pitch shifting algorithms on dsPIC...

I have done pitch shifting using a ring buffer with a range of PICs including the dsPIC 33FJ64GP802 - the chip Tom's digital delay uses. You can actually use both outputs from this chip to give you 2 different outputs, like the micro POG.
However, I never released these as they don't quite sound good enough. There are glitches from the ring buffer, which you can't completely remove.
I did intend to try using a FFT/IFFT approach, as this should be much cleaner. However, I could not get this running fast enough to work usefully. Perhaps with the better coding, dsPIC libraries and some more time, it would be quite doable. Might require a faster chip, which is not likely to be available in a DIP package.

[EBK]

Honestly I'd vote for a POG You can usually get a used one for a reasonable price, and it's really hard to get close to the POG sound with analog.

Probably the best advice.  The more I think about it, the more I feel that the price of a POG is well justified.  But, be wary of POG imitaters.  Some might be good, but there are some terrible ones out there, so I'm told.

There is probably some good "real-time" DSP software that can run on a laptop with a guitar plugged into a USB-based adapter.  You have to watch the latency though. 

[EBK]

However, I never released these as they don't quite sound good enough. There are glitches from the ring buffer, which you can't completely remove.

A couple tricks some have tried to combat the phase mismatch artifacts include having a lot of overlapping ring buffers, which complicates the cross-fading quite a bit, or matching the cross-fading to a fundamental frequency detector, which quickly gets you back to "why am I trying to do this rediculously overcomplicated thing again?" 

But again, I'd recommend a phase vocoder approach if you are a glutton for punishment interested in making something that sounds good.

[samhay]

>A couple tricks some have tried to combat the phase mismatch artifacts include having a lot of overlapping ring buffers...

Yes, but then you need to add extra memory as the dsPICs don't have enough on board.

>or matching the cross-fading to a fundamental frequency detector

Which typically leads to the glitchy response one is used to with analog methods.

>But again, I'd recommend a phase vocoder approach if you are a glutton for punishment interested in making something that sounds good.

I agree. But not sure the PIC is the right platform.

[rutabaga bob]

I have wanted an octave-down for occasional use with a bass.  Ran across one the old-time funk players used to use, but can't remember the name of it now.  Anybody help me out?