Waveforms in motion

Ok- Been scouring my samples to see if I could find anything remotely similar to your waveforms, and never did..

But I did find this clip with the EW playing bass and modulation being applied - Look at the variations in the waveform!

Fred.

Above sample was grabbed from This video:

Using http://www.youtube-mp3.org/

Snapshots of your waveforms:

My first guess is the audio waveform is mostly a square wave but the signal is being recorded with a speaker and microphone setup, acoustics are causing the effect. This is why I always go with theremin direct to sound card so there is no acoustic influence when I want people to evaluate a wave form.

My next guess if acoustics do not come into play... it is a divided down fundamental frequency sine wave and remixed with the square wave which gives the amplitude modulation on the square wave peaks.

Before I captured authentic sound I did something like that wave shape to enhance sound.

Listen.mp3   more a faux cello sound.

Christopher

Well I'm wondering if its just sampling issues with the plug-in. There's just too much jaggedness to what I recorded.

Rich

Yea, I have to believe my video is just displaying sampling induced abnormalities. Maybe that plug-in needs more processing power. Have to see if I can change the sampling rate or put a real scope on it. It's just to jagged.

Rich

Fred wrote: Thierry - Help!  ;-) ... Because I dont recognize this as an EW! - So either my EW's are odd, or this one is!

In fact, that does not look like a healthy Etherwave wave form, taken at the audio output. There, you don't normally have neither these fluctuations, nor this ringing. The etherwave's output is (due to its tanh function wave shaping by overdriving and then diode biasing the LM13700 inputs = variable soft clipping) always a kind of square wave where the duty cycle varies with the position of the waveform knob and the raise/fall time (slope) varies with the brightness knob. In ever case, the floors and ceilings of the LM13700 soft-clipped output are always flat, the edges somewhat rounded by the following LPF.

What we see here has many additions which must come from "outside", i.e. grounding issues or a dirty SMPS in the neighborhood.

Thierry,

I have dome detective work to do here. This was going through a tube preamp channel strip with some EQ to a mixer powered through wall power conditioners and then through cubase monitoring through that plug-in, so there's a lot of stuff between the actual Etherwave output and the scope.

I'll start by just connecting the Etherware output to a Tex scope and see what that shows. I expect that will be more in line with what you are saying here.

Thanks

Rich

Fred, what you say about waveforms in motion makes a lot of sense to me - I guessed a while ago that there was some sort of frequency and amplitude modulation going on during vibrato and staccato - so it's nice to see this validated by experimental results.

To me it suggests a way of improving digital emulation of theremin sounds as per, for example, the theremini, by considering not just the momentary frequency and amplitude of the theremin signal but also the rate of change of frequency and amplitude when synthesising the waveform. Maybe even the second derivatives over time too - i.e. the position of the players hands in the capacitive fields, and also their velocity and acceleration - I guess that these are also a factor - I also guess that subsequent derivatives - jerk and jounce - are of diminishing significance. 

Hi Gordon,

I agree - I think that if the rate of change (delta) was separately derived, it should be possible to use this to apply harmonic changes to the sound.

If the above was done, it might even be possible to correct somewhat for sluggish front-end response - as in, a detected delta could be 'amplified' and applied to the audio engine / processor - but to be honest, I think that would probably be more destructive to playability than of benefit.

The thing about direct heterodyning is that any change in any cycle of the HF (Variable) oscillator is instantly reflected on the audio waveform - at say 250kHz, this is an AUDIO update every 4us - but in fact, its even faster than this, it is, quite literally, instant - because even the HF waveform will be distorting as the capacitance changes.

I have no idea at all how much anyone can hear of the harmonic changes that occur to the audio waveform - Any change to harmonics above the instruments audio roll-off frequency will obviously be lost, so I doubt that anything above 10kHz will be of any significance.. and from this we can say that updating the audio waveform every ~100us should allow everything that can be heard, to be heard.

But at the front end, these changes need to be 'picked up' in order for them to be used... I do not actually see any way to differentiate between different gestural actions fast enough to allow their use in a theremin.. Any such processing will take time, which will bring us back to theremini cookoo land (although in the thereminis 100ms lag, you should be able to do a hell of a lot of processing! ;-)

For playability one needs faster than 10ms, and I think if one has this, it may be possible to separately derive delta and use this to accentuate harmonic changes - I think this could be really interesting in fact.. having vibrato for example that had a greater influence on harmonic content as the vibrato frequency increased (it will have this anyway on a heterodyning instrument, but would be interesting to mess with adjusting this relationship) .. But really, the most interesting would probably be to have 'amplified' delta acting with a fast (1ms or less) pitch tracking - I think that doing anything to the waveform which is not synchronous with some other 'cue' such as pitch or volume, will probably sound 'unnatural'.

(not that 'unnatural' is always bad ;-)

Ok - Lets get back to YOUR suggestion ;-)  ... Sorry - I went off on one again..

"To me it suggests a way of improving digital emulation of theremin sounds as per, for example, the theremini, by considering not just the momentary frequency and amplitude of the theremin signal but also the rate of change of frequency and amplitude when synthesising the waveform."

If the front-end is fast enough, and one has a digital engine capable of updating the waveform during an audio cycle, then this should be possible.. A digital voice which only updated the audio every audio cycle or 1/2 cycle would not be able to synthesise the waveform.. This does not mean however that there is no solution.. If one had some digital 'processor' (distorter) that was fed the (digital or analogue) composite audio, and this block was fed a delta signal and could process (distort) the audio in real time as a function of the delta signal, then this could be a route..

I do not think the "theremini" should even be mentioned - there is absolutely no hope for 'improving' it other than by completely redesigning it - hardware and software.. The theremini isn't  a toy  one could turn into a theremin - it would be difficult (probably impossible) to even turn it into a playable toy theremin!

Fred

Fred,

I see what is going on and it is really interesting. I took the computer out of the equation and went right to a TekScope. When I just create a fixed tone playing on the Etherwave and recorded direct to scope I get a nice steady wave.

But then I scoped the output of the ART Voice Channel Tube Pre-Amp that parallel feeds a Lexicon MX200 Reverb processor I am running the Etherwave through. And I get something very similar to what the plug-in scope gave me:

Etherwave through ART Voice Channel Tube Pre-Amp

It turns out when I remove the Lexicon, the waveform of the Etherwave alone is similar through the tube-preamp.

That Lexicon reverb unit is adding the weirdness!!

So I connected a pedal reverb unit instead (Hall of Fame stomp box) and still get a bit of fluctuation, but now on a waveform that is more reasonable. I took out the reverb in the middle of the little video and you can see the waveform totally stabilize then. 

Etherwave through Hall of Fame Reverb & ART Tube Pre-Amp

I also notice that Lexicon unit really cuts the gain and I have to turn gain up on the pre-amp which likely added some distortion as well.

So I need to determine if that Lexicon unit is doing what it is supposed to.

Rich

Rich,

That is much more like the EW waveforms I am used to seeing!

I use my Lexicon MPX1 a lot (its my favorite processor by miles) but have never bothered to look at the waveforms from it - Its just such a complex processor that making any guess about what any output waveform "means" is pointless IMO...

The MPX1 is superb IMO - Great for theremins even if only used for 'overdrive' distortion, but the reverbs and flange and chorus and chord generators etc actually allow some astounding "synth" pads and sequences to be generated from simple theremin input...

But unless only used lightly, the MPX1 consumes the theremin.. And even if used lightly, I wouldnt trust what I see on its output...

What I am not sure about though is this.. Are you saying that when the MPX1 is loading the theremin or amp, you are seeing this strange squared ringing signal on its INPUT?

- Because this wouldnt make any sense to me.. I could hazard guesses if it occurred when either the theremin or the amp was connected - but not for both.. By the time the theremin has gone through the amp,Iwouldnt expect RF levels to be high..

This leaves only one common that I can see - the grounding to your Lexicon.. That would be the first thing I would check.

Fred.