MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

Frequency Shifting possible?
MUFF WIGGLER Forum Index -> Wiard  
Author Frequency Shifting possible?
chamomileshark
Just wondering, with all the many wonderous things the Wiard 300 series can do - can it do frequency shifting in some way?
numan7
hmm, let's see. we have:

- balanced modulation (model 361)

- bandpass filters (2x model 351 or model 352)

i suspect that a basic frequency-shifter can be patched from just these functions.

but i haven't thought about the details yet, and still need do some more reading on the subject...

cheers
drumsofd00m
In theory you can do it with two ring modulators and two quadrature phase shifters for 90° shifted versions of the carrier and modulator. You can approximate this shift in 8bit with a custom EPROM for the Waveform City (for one or both of the signals, depending on how many WFCs you have). The 311C does have two quadrature sine wave LFOs, and the Mixolator does ringmodulation (albeit with very heavy bleedthrough, on my module), so you might get there in a noisy, crude DIY way. I'll try as soon as I have an EPROM burner and/ or the 311C. Actually I posted this very patch idea to the Wiard group 12 years ago... funny how some things just take their time.

Oh yeah, about bandpass filters, most of these have a too soft rolloff to really cut off the unwanted sideband, and this is especially true for the Borg filters. Grant used to describe their BP mode as a Helmholtz resonator rather than a true bandpass. You *can* use filtering for frequency shifting, but you'd want a very steep lopass to isolate the LSB (lower side band) and a very steep hipass for the USB. Not very practical IMO. However, if you want to perform frequency shifting to get a thru zero phaser, there is a heterodyning approach where (IIRC) you create a very high pitched/ shifted USB, cut off the LSB as best as you can and then shift the USB back by only *almost* the same amount and mix it with the original dry carrier signal... AFAIK it didn't work all that well in vintage FreqShift units tho and it might end up a noisy mess with the above mentioned "8bit quadrature" patch. But of course it's always worth trying, who knows where it leads.

While I'm at it, I keep asking this without success... frequency shifting is single side band modulation, so isn't ring modulation really *two* quadrant modulation (because it produces a 50/50 mix of sums and differences of the partials), not "four quadrant" as most people write? whereas AM (amplitude modulation) would be the real four quadrant modulation (sum, difference, AND the two original signals)? Any experts on this?
chamomileshark
hi thanks for the replies. I also got the impression that ring mod and frequency shifting were related but I've never seen it explicitly expressed.

I have two ring mods in Frac which might be better than using a mixolator.

That leaves me with trying to find 2 90 degree phase shifters. The boogie filter outputs have different phases, can you use the 6db/octave output from the Boogie filter as 90 shifted from the input?

Alternatively I have two polarizing mixers - can they be used in some way?
drumsofd00m
They are very closely related - frequency shifting in effect is nothing but getting the sum and difference (upper and lower sidebands) separately instead of summed. I'm just not clear about the two/ four "quadrant" terminology regarding RM and AM, just like the terms carrier and modulator are often mixed up.

Interesting idea to inspect the Boogie for a quadrature phase shift! However, just like with the (theoretical/ custom programmed) 8bit quadrature output of the WFC, the signal may be too strongly altered to ultimately produce a frequency shift without lots of bleedthrough or additional noise. From memory, the Boogie seems to add a lot of bass to whatever I put through it, even with the 6dB/o output, which is amazing, but probably not a pure enough "through" which you want here.

A polarizing mixer does not provide 90 or 270° shifted outputs, but 180° only. (Edit: if you mean the D*epfer module, it's just another term for "attenuverting" mixer)

I think getting two Waveform Cities (so you don't have a clean 0° signal and an 8bit 90° signal, but two identically quantized ones) plus a 311C and two external ringmods that are cleaner than the Mixolator looks like the best option.
chamomileshark
ok thanks. I'll probably give it a go.

Can you explain a little more why there are two ring mods?

The polarizing mixers I have are the Banalogue 3P ones which were based on serge designs.
drumsofd00m
Off the top of my head (I haven't done this in more than a decade): You want to (in terms of voltages, not frequencies) sum the ringmodulations of the two 0° and the 90° signals, and IIRC the actual sum will be one sideband, the difference (where one ringmod is inverted before the mixer) the other sideband... this could be nonsense though, better do some research of your own and correct me if you find something different.

Just BTW, I've been so busy just exploring the basic timbres of the Wiard (tho to be fair also some of its bugs & quirks), that I simply haven't yet tried many of the more esoteric/ effect-like patches I've been wanting to. I've always been interested/ would have vowed for a dedicated frequency shifter & thru zero phaser for the Wiard lineup, tho; kind of a utility/ processing module that accomodates this (with internal and external carrier) and since that doesn't require a lot of controls & jacks, a few other bits that haven't been available in the 300 and 1200 (at least not as seperate patch points, like a S&H/ T&H/ shift register).
ndkent
I'm surprised no one has come the the fairly obvious conclusion I have, in that if it were something you could patch up on your modular system you would have seen plenty of "how to" youtubes and web pages by now. Frequency Shifting has been around for decades and I think must have been an accidental byproduct of broadcasting. The first famous unit was the Bode sold through Moog in the early 70s and not simple or cheap. I am impressed though that the much newer Cwejman module doesn't seem to be digital and isn't comparatively expensive in his lineup, he clearly knows something no so common. Many out there are part or entirely digital to cut size and costs. I believe accuracy is critical so from a design standpoint it probably is how much do you spend to do it all analog for the sake of analog. Most, digital or not use a sine as the only carrier option since it's easier to design an internal wave and is the classic effect. I'm sure if one does digital then other built in waves (like the Synthesis Tech) are easy if the designer wanted them in there but an external carrier is a good bit more hardware.

As mentioned, it's ring modulation with a separated up and down shift. If both are combined then you get a ring mod again... but the point is they are separated and it takes lots of circuitry to do it acceptably.

http://www.jhaible.com/fs1a/fs1a.html might be interesting to some as to the circuitry, as hinted at, not a lot of DIY or simple circuits.
drumsofd00m
Erm... What *is* your conclusion? ;)

Those years ago I used to read discussions of expert owners of vintage frequency shifters on AH, but never used these myself (I preferred to do it in digital in the first place for cleaner signals and for full pitch control over the carrier). From what I recall, the carrier of the Bode module could not go below a certain frequency, so for thru zero phase shifting, one needed two for the heterodying patch I described (which is used in radio also as you say). The old EMS unit IIRC could go lower but introduced stairstepping because the phase shift was done via digital table lookup (so essentially like the Waveform City).

It's great there are better units out there today (personally I've only used the very well working, albeit not too fat sounding - digital? - Buchla 297 Infinite Phaser shifter so far, not the usual frequency shifters per se), but I thought Mark wanted to implement it within his Wiard system as far as possible just for the fun of it - right? People buy whole Serge systems just to patch something up that they could get much easier on a Buchla or Wiard...


EDIT - Ok, I got it on third or fourth re-reading, sorry. I disagree tho - I think the knowledge of certain old tricks and scientific methods has actually declined as the availability of both very complex modulars and online videos has increased.
Graham Hinton
drumsofd00m wrote:

While I'm at it, I keep asking this without success... frequency shifting is single side band modulation, so isn't ring modulation really *two* quadrant modulation (because it produces a 50/50 mix of sums and differences of the partials), not "four quadrant" as most people write? whereas AM (amplitude modulation) would be the real four quadrant modulation (sum, difference, AND the two original signals)? Any experts on this?


You are confusing the instantaneous voltage product of multiplication and frequency products. Quadrants refer to the four parts of a cartesian graph of the input voltages plotting two bipolar voltages. A linear multiplier operates in all four quadrants, i.e. two negative voltages will produce a positive voltage output according to normal multiplication rules. A VCA only operates in two quadrants, even if the CV has an offset, it does not change the sign of the input.

Four quadrant multiplication is balanced modulation and two quadrant multiplication is amplitude modulation. The latter does not contain both original signals, it contains the carrier plus two sidebands at the modulator frequency apart.


ndkent wrote:
I'm surprised no one has come the the fairly obvious conclusion I have, in that if it were something you could patch up on your modular system you would have seen plenty of "how to" youtubes and web pages by now.


First you would need a 90 degree phase shift module, or two, that works well across the whole audio range. These aren't exactly thick on the ground, but if anybody wants one please let me know as I'm considering making a batch.

Frequency shifting is often a disappointment and is expensive to implement well. The concept sounds exciting, but it is not that much more useful than ring modulation. Pitch shifting is more useful.

drumsofd00m wrote:
I think the knowledge of certain old tricks and scientific methods has actually declined as the availability of both very complex modulars and online videos has increased.


To understand what is possible you would have had to have remained awake when you did trigonometric identities at school and most people just glazed over.

See this thread: how do you patch a frequency shifter?
drumsofd00m
Graham Hinton wrote:

You are confusing the instantaneous voltage product of multiplication and frequency products. Quadrants refer to the four parts of a cartesian graph of the input voltages plotting two bipolar voltages. A linear multiplier operates in all four quadrants, i.e. two negative voltages will produce a positive voltage output according to normal multiplication rules. A VCA only operates in two quadrants, even if the CV has an offset, it does not change the sign of the input.


Thanks, that reminded me and I'm clear now, at least as far as practical musical use. To explain it in practical terms and stick with Wiard systems, it's what we hear when switching the Mixolator between AM and RM while e.g. an LFO is going into the Z input - AM will produce unipolar amplitude swings and RM bipolar swings (at, IIRC, twice the frequency? and one of the swings inverting the phase of the signal, again IIRC? Hm, I've actually lost some of my facts memory over the years once I was doing things more intuitively and more musically focussed).

Quote:
Four quadrant multiplication is balanced modulation and two quadrant multiplication is amplitude modulation. The latter does not contain both original signals, it contains the carrier plus two sidebands at the modulator frequency apart.


Interesting (about actually not retaining the originals) - in that case there's some widespread misinformation. I'll look into that with a software modular tonight (they may not sound as good but are often really clean and mathematically more ideal).

Quote:

ndkent wrote:
I'm surprised no one has come the the fairly obvious conclusion I have, in that if it were something you could patch up on your modular system you would have seen plenty of "how to" youtubes and web pages by now.


First you would need a 90 degree phase shift module, or two, that works well across the whole audio range. These aren't exactly thick on the ground, but if anybody wants one please let me know as I'm considering making a batch.


Might be interested later this year; I have two WFCs and a PSIM-1 (an early collaborative Wiard product) so can do it in digital as I said above, but with 8bit quantization and/ or higher resolution but some latency. How would yours be implemented? Would it be based on the old EMS unit?

Quote:

Frequency shifting is often a disappointment and is expensive to implement well. The concept sounds exciting, but it is not that much more useful than ring modulation. Pitch shifting is more useful.


The first sentence is true but I totally disagree about the others. Maybe it's my musician's experience versus engineer's thinking, or maybe it's because many musicians' imagination is limited by the fact that *analog* frequency shifting is in effect so problematic, especially when there's no external carrier input and/ or no 1V/oct CV input for the internal carrier. Personally I've never understood what should be desirable about ringmodulating or frequency shifting a signal with a sine wave of *static* pitch other than for Dalek voices, and I find it boring and too commonplace. But I've done some patches in software were frequency shifting with precise control over the shift amount was 1) crucial to the sound, 2) worked quite well thanks to reasonably good usability of an external carrier and 3) the single side bands produced a much more interesting, cleaner, less grating/ clangorous soud than the sum of the sidebands = a ringmodulator, but rather timbres like a PPG or Prophet VS (i.e. waveforms quite removed from the standard analog ones). In terms of sound aesthetics and usability it's a bit like comparing exponential FM (which often sounds "dirty", especially if the index is varied so that pitch drifts) to linear FM or even thru zero FM or PM (much more potential for cleaner and better controllable timbres, less audible beats etc).

Quote:

drumsofd00m wrote:
I think the knowledge of certain old tricks and scientific methods has actually declined as the availability of both very complex modulars and online videos has increased.


To understand what is possible you would have had to have remained awake when you did trigonometric identities at school and most people just glazed over.

See this thread: how do you patch a frequency shifter?


I'll read that tonight, but actually I meant something different, maybe didn't express it well. It's going off topic here in the Wiard forum, but I meant that as more modules with built in patches, more complex waveforms by default, and more free-style controllers became easily available, people didn't have to search years and "pay king's ransoms" (quote Don in 2001) for small Buchla systems etc, they didn't have to ask a handful experts anymore or study EM at universities, but could just tweak away at it and share videos in a matter of minutes, and it's not really necessary anymore to study synthesis in depth. It has become so much easier to get fascinating sounds. Plus some of the older experts from the AH days have passed or do not participate anymore, while the number of amateurs with possession of modulars has grown by maybe a hundred or thousand times, so while there may be profound info, it is not easier to find because of all the stuff to wade through. Actually, I did find some tutorials on what ndkent said there weren't any, but maybe even with online searches, doing it somewhat systematically and a bit pf patience is not in fashion anymore ;)
So, you may be right about drifting off in maths classes, and I was one of those candidates at one point, but with "scientific" here (maybe not the right term) I meant approaching one's personal synthesis journey in an at least somewhat systematic way of looking at a module's principles instead of just going through module functions (which is what most youtube videos seem to do).
/OT
MUFF WIGGLER Forum Index -> Wiard  
Page 1 of 1
Powered by phpBB © phpBB Group