smooth sounding octave?

[Bill Mountain]

Any smooth sounding octave pedals?  Everything I've played had a splatty fuzz sound.  I'm looking for something with an overdrive feel.

Any ideas?

[pappasmurfsharem]

Never tried this, but maybe attaching some sort of octave stage to the end of an overdrive?

[Bill Mountain]

Just single note lines/solo's on bass.

[mth5044]

Did you want octave down or up? Both? How many in which direction?

[R.G.]

It's quite difficult. This is one issue that both the JFET and MOS Doublers and the PLL/synthesizer was intended to address. They had varying degrees of success.

A multiplier chip offers some hope, as does analog pitch shifting, but they're both difficult to implement well.

[Bill Mountain]

Looking for octave up but I don't want it to be clean.  Just not abrasive.

Is the spattyness the nature of the beast?

[Mark Hammer]

We've seen lots of octave up circuits pass through this forum, but you know, I can't recall too many where there may have been separate paths for an octave up and straight sound, with perhaps a tone and level control specific to the octave path.  If the octave-doubled signal could be "tamed" by suitable low-pass filtering and then mixed back in, maybe it would sound less harsh.

At least part of the splattyness is, I think, a byproduct of needing to combine two halfwave rectified copies of the signal.  Because the diodes used to do that impose a threshold for conducting, they will introduce crossover distortion.  The higher the forward voltage of the diode, the more crossover distortion is produced.  My gut sense is that using Schottky diodes with the lowest forward voltage you can find, would help.  Maybe.

[Quackzed]

seems counter intuitive, but maybee adding some decent clipping after the rectifier and then low pass filtering might seem smoother over all. sure more harmonic content/fuzz but more even in volume, and with heavy handed low pass filtering i'd think it might be smoothish...
less blatty anyway... could just stick a distortion with the tone rolled down after it and see..

[amptramp]

Check Figure 30 here:

http://www.onsemi.com/pub_link/Collateral/MC1496-D.PDF

for an audio frequency doubler.  If you stay within the linear range, there will be only the octave up without the extra harmonics that fullwave rectifier circuits generate.  It uses the trig identity:

cos 2F = 2cos²F - 1

so a pure cosine input at frequency F when squared will give a frequency 2F and a DC component that gets blocked by a coupling capacitor.  If you use a multiplier in a feedback loop, you can get a square root function and a half frequency or octave down following this formula:

cos²(F/2) = (1 + cos F)/2

taking the square root of both sides:

cos (F/2) = SQRT((1 + cos F)/2)

Frequency x time = angle so this is how frequencies map into angles suitable for trig transformations.

[R.G.]

Yeah, that's the multiplier way.

The problem with analog ways in general is that they work GREAT on sine waves, but intermodulate badly on waves that are not simple sines.  Still, a simple multiplier isn't as bad as full wave rectification.

Here's the PLL way I came up with: http://www.diystompboxes.com/smfforum/index.php?topic=105622.msg952116#msg952116

It uses a phase locked loop to lock onto the note, then synthesizes two octaves down and five octaves up digitally (i.e. square waves out). The saving grace for this is that you can then use resistor networks to add the higher octaves into a much smoother-sounding sawtooth approximation, and filter down from there.

[mac]

Never tried this, but maybe attaching some sort of octave stage to the end of an overdrive?

See my gallery, ring fuzz & snd demo.

mac

[aron]

This might be something better done with digital. So far I haven't heard a really good analog octave up that's clean and smooth.

[Bill Mountain]

Digital is a little beyond my abilities but I think my plan will be to build a basic overdrive and split the output into two channels with one of them being an octave section with filtering and then a blend control to mix it with the overdrive tone.

Would this work for the phase splitter?  Section 2.4 from this application note uses a resistor network to create a balance out from on unbalanced signal:

http://www.jensen-transformers.com/an/an003.pdf

[R.G.]

It's not clear to me why you'd want a phase splitter, or a splitter of any kind. I think I'm missing something.

[Bill Mountain]

It's not clear to me why you'd want a phase splitter, or a splitter of any kind. I think I'm missing something.

Isn't that how you achieve the octave effect?  Once you have the two phases you put the diodes in series to get the octave up sound.  Or am I missing something?

[anchovie]

Digital is a little beyond my abilities

There is no shame in being a member of this forum and buying a commercial pedal - especially when it comes to digital pitch shifters! 

[deadastronaut]

+1

just rehouse it...

[amptramp]

It's not clear to me why you'd want a phase splitter, or a splitter of any kind. I think I'm missing something.

Isn't that how you achieve the octave effect?  Once you have the two phases you put the diodes in series to get the octave up sound.  Or am I missing something?

This is the fullwave rectification method of frequency doubling.  The octave up is in there, but so are a lot of other harmonics.  You can get sound samples here:

http://www.falstad.com/fourier/

by running the Java applet, choosing a source such as sine and a function such as full rectify and you can hear what the waveform sounds like.  Then drag your cursor over the response peaks to get the amplitude and phase of each harmonic.

[R.G.]

Isn't that how you achieve the octave effect?  Once you have the two phases you put the diodes in series to get the octave up sound.  Or am I missing something?

Ah. I was not understanding.

One way of doing an octave effect is by doing full wave rectification, and one way of doing that is with a phase splitter.

Yes, you can do it that way. The Super Fuzz, Fox Tone Machine, and Fender Blender do it that way. The Octavia with the transformer in it does it kind of that way, but the transformer is center tapped. You can also do it with a two-opamp precision rectifier circuit and some others. I sometimes get lost in the "all the others" and trying to be both complete and concise.

All the rectification-based ways will not give a particularly smooth sounding octave. Maybe acceptable, but still with distortion.

[Mark Hammer]

I think Bill is following up on something I suggested earlier in the thread: processing a derived octave-up separately from a clean and/or overdriven undoubled sound.

If you do that, Bill, it might be appropriate to trim a fair amount of treble off the basic overdriven sound in order to make the octave-up a little more evident.  You'll likely want the basic overdrive to give you the fundamental, and the rectified octave-up to give you the octave and any other harmonics.

Thinking about this, and not wishing to extract any industry secrets, what I described seems to be what the VS Angry Fuzz does.  It has a control for blending in octave-up, so I assume there are two paths internally.