all analog pitch shifting idea (without BBD's)

[Brian Marshall]

Hey guys, I haven't been around much lately, but thought I'd share this with you all.

There is an idea I have had rattling around in my head that stems from being around a lot of amateur radio people in my youth among other things. 

Anyway, when listening to a broadcast transmitted using sideband modulation there is often an audible effect on sound.  If you heard it you would probably recognize the effect you'd hear in a science fiction movie for a robot voice or radio transmission.  (this leads me to believe someone has already done this, but not sure it has ever been applied to actual musical instruments)

As far as I know sideband is not used for commercial audio signal transmissions any more, partly because this effect was so noticable, and required more complex demodulation circuits to negate the effect.  Although it was very efficient, the frequency range of the audio signal was limited, and from what i understand required many all pass filters in series to expand the frequency range.

Now the actually pitch shifting is not really pitch shifting.  After a bit of searching around the internet the term "spectrum shifting" appeared a few times. 

THE DIFFERENCE- pitch shifting is by percentage of frequency where as spectrum shifting is by hz
example-
5% pitch shift- a signal of 100hz with a 2nd harmonic of 200hz shifted to 105hz, and the harmonic is shifted to 210hz
5hz spectrum shift- a signal of 100hz with a 2nd harmonic of 200hz shifted to 105hz, and the harmonic is shifted to 205hz

I believe the shift in hz is equal to the difference between the rejected transmission carrier and the demodulation carrier, but I will fully admit I don't know much about radio broadcast electronics beyond AM/FM

now if you really wanted "normal" pitch shifting this sound would probably be dramatically disappointing.  Some might find it utterly useless... it may also require a ridiculously large circuit, but it could certainly be an effect unlike anything else.

[Mark Hammer]

Mike Irwin came up with this frequency shifter module design for modular synth company Modcan.  I think this is what you are describing.
http://www.modcan.com/modhtml/freqshift.html

Oh, and welcome back, buddy.  Long time no see.

[Taylor]

Yep, there's also the old Bode frequency shifter in Moog modulars, which I'm pretty sure is the genesis of the effect, at least as a purposeful audio effect. It's something I've thought about stompboxifying, but seems much more complex to achieve than ring modulation, which is a pretty similar sound (except that a frequency shifter at low setting sounds like a phaser).

[aziltz]

THE DIFFERENCE- pitch shifting is by percentage of frequency where as spectrum shifting is by hz
example-
5% pitch shift- a signal of 100hz with a 2nd harmonic of 200hz shifted to 105hz, and the harmonic is shifted to 210hz
5hz spectrum shift- a signal of 100hz with a 2nd harmonic of 200hz shifted to 105hz, and the harmonic is shifted to 205hz

the way this is explained makes me think the pitch shift would sound deeper on the lower notes, which I think would be pretty awesome.  i wonder if we can find a block diagram of this process.

ps brian I super dig your pedals.

[Brian Marshall]

Mike Irwin came up with this frequency shifter module design for modular synth company Modcan.  I think this is what you are describing.
http://www.modcan.com/modhtml/freqshift.html

Oh, and welcome back, buddy.  Long time no see.

Yep, that is EXACTLY what i was talking about.  and it's only $937    Granted, if it works the way i think it does there's probably quite a bit of work that goes in to each one.
I guess I'll have to put it on my wish list.

[Taylor]

THE DIFFERENCE- pitch shifting is by percentage of frequency where as spectrum shifting is by hz
example-
5% pitch shift- a signal of 100hz with a 2nd harmonic of 200hz shifted to 105hz, and the harmonic is shifted to 210hz
5hz spectrum shift- a signal of 100hz with a 2nd harmonic of 200hz shifted to 105hz, and the harmonic is shifted to 205hz

the way this is explained makes me think the pitch shift would sound deeper on the lower notes, which I think would be pretty awesome.  i wonder if we can find a block diagram of this process.

ps brian I super dig your pedals.

It really doesn't sound at all like "pitch shifting" we're used to - like digital multi effects, Whammy, POG, etc. It sounds a lot more like ring modulation. So it does sound cool (I'm a ring mod fiend) but it's not like a "deeper" Whammy. There's a schematic available. It is very, very complex.

[R.G.]

I-Q Modulator

[soggybag]

What about that guy that makes the PLL pitch shifter? I forget the name it was posted here ... searches forum and comes up with ... http://www.schumannelectronics.com/pll.html

[Mick Bailey]

Interesting. I have a 'Rambler' regenerative Radio that I built a couple of years ago and this has a similar robotic ring-mod type effect when the regen control is turned up too much (but not so much as it oscillates). The rambler uses a single space-charge triode so it could probably be reproduced with a single transistor and adapted to give a similar effect with a guitar signal.

Igor - to the lab.......

[cpm]

What about that guy that makes the PLL pitch shifter? I forget the name it was posted here ... searches forum and comes up with ... http://www.schumannelectronics.com/pll.html

PLL pitch shifting is useful if you are working o a single and uniform wave. That's the concept used in octavers like the OC2, where a square wave is generated from fundamental extracted, then dividedd by2, and by2 again to get 2 octaves down, however its a raw synthesized sound.
The fun comes with PLLs where you can precisely shift the frequency to any factor.

[markusw]

From my experience PLLs will always have a glitch at the beginning of each note. It can be reduced with appropriate measures (nice edges of the square signal feeding the PLL, sample-hold-circuit that "remembers" the previous note) but at least to my experience you will never be able to completely get rid of it (at least for the lower notes on fretboard).
Cool thing it's pretty easy to generate 3rds, 5ths, octaves up and down......

[Mark Hammer]

Pitch shifting is one of those things, very much like ring modulation, that appeals principally to synthesists, and is NOT a set-and-forget type of effect.  One uses it VERY sparingly.  As a result, nobody gets volume discounts on manufacturing these.

That's not a reason to be disinterested in it, merely a reason to not expect Danelectro to sell one for $20 anytime soon.

The feel is very different than either a harmonizer, or analog interval generator (octave-down or something like the E&MM harmony Generator), or ring modulator, or pitch wobble of a chorus-pedal set to vibrato mode.  One of the ways it's different is that it becomes more "pitched" as you go higher.  In other words, if my pitch offset is 5hz, then the harmonics of a 1khz fundamental won't be nearly as far off from where they ought to be as they might with a 200hz note.  With ring modulators, you can get that general result too, except that frequency shifters CAN do things like a 5hz offset, whereas the modulation frequency has to be in the audio range for a ring/balanced modulator effect (i.e., >20hz).  Most commercial ring modulators will likely modulate at frequencies of 50hz or more.

Another thing that's different is that the shift is, from what I understand, unidirectional at a time.  At the very least, the two "directions" are separable.   So, a 5hz shift added to a 1khz square wave would provide a 1st harmonic at 2005hz, a second harmonic 3005hz, a 4th at 4005hz, etc.  A ring modulator would produce sum and difference products all the way along.  So a 1khz square wave gets you 900 and 1100hz, but it also gets you 1900 and 2100hz, 2900 and 3100hz, etc.; a much "busier" output.

[Strategy]

Frequency shifters are a pretty big topic in the synth-DIY community.

One of the gurus, Juergen Haible, a German guy who is an engineer and has developed a number of clone and original projects on a hobby/open source basis, sells PCBs, etc. for a Frequency Shifter of his own design.

As pointed out in this thread, it's not much like pitch shifting exaclty, not like a Digitech Whammy would do. Its more like ring modulation, but just with one sideband instead of sum and difference sidebands. But it has a similar bell tone sort of thing. When you get into lower frequencies, you can get into really strange "barberpole" (illusion of constant rising/falling) modulation, tremelo type sounds, etc. Frequency Shifter circuits are notoriously difficult project.

I would like to suggest adventurous Pedal DIY'ers to go check out Juergen's projects in particular. While most synth DIY is based around large modular systems, Juergen develops most of his projects as standalones. His Tau and Krautrock phaser projects, and even the Frequency Shifter, would be friendly to pedal formats; I have seen pics of a pedal format krautrock. In complexity, probably equal to or exceeding the lovetone clones, BUT, his designs are so solid and the documentation so good that there are often no or few questions that arise in assembly, making troubleshooting very straightforward.

I would say that the Frequency shifter is probably the most complex project he's developed though. And agian, it won't have the benefits of PLL harmonizing or digitech/eventide style digital pitch shifting. But you WILL have a kick ass synthetic effect to work with 

- Strategy

[earthtonesaudio]

I found this on accident while searching for something much less interesting:
http://www.jhaible.de/tonline_stuff/hj_fs.html (This is the page Strategy was referring to, I believe.)

[Strategy]

That's the link, yeah. his domain recently changed and I couldn't remember which was the right one.

I am building the Tau Phaser right now as a standalone. It is pretty crucial- 20 stage filter!!!

Paul

[Taylor]

That's the link, yeah. his domain recently changed and I couldn't remember which was the right one.

I am building the Tau Phaser right now as a standalone. It is pretty crucial- 20 stage filter!!!

Paul

Let us know how that goes. I'm always interested in far-out stuff like that.

[Strategy]

@ Taylor:
Yeah, I'll do a build report maybe with the intention of introducing the Haible stuff to this forum.

A friend of mine, actually my friend who got me into DIY electronics in the first place, is building the Freq Shifter. I would have to say it's over my head.

I'll soon post some pics of my other builds to date, also some of them are at the point where synth meets pedals. Just touching up the enclosures.

- Strategy

[puretube]

Mr. Haible himself is a member here... (<5 posts, IIRC),
and has been cited since > 7 years...   
(search: "haible" )

[Taylor]

Can someone who has used a frequency shifter answer this:

Ring modulators sound different when playing different notes because the addition/subtraction is a constant number, while the input pitch is a varying number, so the amount of modulation depends on how far your input note is from the oscillator note. What sounds like a bell playing a G just sounds like fast tremolo on a C.

I have read that frequency shifters retain the same tonality across all notes; is this true? Or is it applying a static addition just like ring modulators, just doing it differently (and with separate up and down shifts)? I've been working on a tracking oscillator for a ring modulator, but if the frequency shifter works this way it would be much better.

[Mark Hammer]

The relationship between fundamental and harmonics in any signal is a proportional one.  So, the third harmonic of a 500hz tone is 1500hz, and the third harmonic of a 1000hz tone is 3000hz.  When a signal is fed to a ring modulator, although sideband products are added to "complexify" things, the proportional relationship between the fundamental and the harmonics remains constant.  Yes, the sum and difference will be "closer" to the original signal the higher up you go (assuming the modulating frequency is not deliberately tracking the fundamental), but if I play a 1000hz note (non pure tone), and modulate it with 100hz, I get 100hz added and subracted to/from my first harmonic at 2khz, second harmonic at 3khz, and so on.

With frequency shift, there are no sideband products as such, merely a copy of the input signal with all harmonic content added to by a constant.  So, a 50hz offset to a 1khz input signal (with harmonics) gets me a 1050hz output, with "harmonics" at 2050hz, 3050hz,etc.  Bump that up an octave and a 2khz input gets me 2050, 4050, 6050, etc.  In other words, the harmonic output is not multiples of the original.

I think if you listen to the sound samples for Mike Irwin's Modcan frequency shifter (and read the explanation) you'll get a sense of the differences ( http://www.modcan.com/modhtml/freqshift.html ).  Yes, it does sound sort of ring modulator-like, simply because there is now content which is not harmonically related to the original, but it has a different vibe.  $900 worth?  I'll let you be the judge.  But it's different.