Combining many oscillators?

[Nigel Goodwin]

Hey, I wrote that flute is actually not that I wanted to have. A theremin, just like the violin dont have separated note steps, they slide between them. Isnt that going to be easier?

If you build a theremin it's going to be considerably easier, but that's got nothing to do with what you keep asking? - if you want a theremin then build one of those, there are plenty of projects for doing so on the net (basically it's just a beat note between two RF oscillators).

 

[audioguru]

Isnt that going to be easier?

A theramin is fairly simple because its sound is always the same but a circuit with a switch can change the sound. It has one antenna for pitch (frequency) and another antenna for loudness which can turn the tone on and off.

 

[MrAl]

Is this it?
https://www.google.co.uk/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&cad=rja&uact=8&ved=0CDQQtwIwAQ&url=https://www.youtube.com/watch?v=B5EicXYEfs8&ei=7rQ-U_biMIWqhAfLtoCABg&usg=AFQjCNEep6EuOIz14Oaxye4njKI9gGK-pA&bvm=bv.64125504,d.ZG4

Hi,

Well no it's not that one, but thanks for pointing to that link as i was able to find it!

If you never listen to any other classical piece in your whole life you GOT to listen to this flute section. It starts around 1:24:40 in this link:

might start a little before that or a little after that.

Unfortunately this isnt the best rendition but it isnt too bad.

 

[audioguru]

If you never listen to any other classical piece in your whole life you GOT to listen to this flute section.

That is THE one. It is too bad it was recorded in a reverberant washroom and the flautist was in a hurry to get out-a-dere.

 

[MrAl]

Hi audioguru,


Ha ha, yeah that describes it perfectly.

That is THE one. It is too bad it was recorded in a reverberant washroom and the flautist was in a hurry to get out-a-dere.

I also did not appreciate the poor chordal accompaniment...was that a dang guitar? Supposed to be a cello for God's sake (with other intruments too). Also sounded like the flute player was a little bit lazy and skipped a few note changes! That's nuts.

There's another flute piece too that is sort of in the middle of a full piece, but i have yet to find that one. Really really nice too...if done by the right orchestra of course. I heard a good French orchestra do these pieces and they were better than perfect! I wish i could find those recordings now.

 

[MrAl]

Hi again Audioguru:

This one is really the ultimate:

The orchestra is much better although i feel the flute player put a little too much of his own interpretation into it. Not bad though. The piece itself is really beautiful though. Flute solo with cello accompaniment starts at 48:30, entire piece starts around 47:10.
There are several other flute passages in that link too though.

 

[synox89]

If you build a theremin it's going to be considerably easier, but that's got nothing to do with what you keep asking? - if you want a theremin then build one of those, there are plenty of projects for doing so on the net (basically it's just a beat note between two RF oscillators).

Well, yes - my goal is to make a primitive model of a theremin: this means i only want to build 3 oscillators: 2 mixed up for the frequency and one for the loudness.

Well, maybe I should reveal my whole situation. I am making a presentation (a pretty important one), which tries to answer the question: "Could a theremin be able to replace a violin?" So my idea was just to make a simple model of the theremin and then theoretically explain to what extend it would be possible to modulate it in a way to sound violin-like. So my first idea was that it should be possible to connect several oscillators, whose frequencies added together, should produce a certain sound wave. So if lets say a violin has this sound wave (hypothetically):

i could just add these two oscillations together:

and then get the sound.

The question is, how these two oscillations can be plugged together in a way, so they would produce a correct wave in every frequency pitch.
I hope this is understandable - the relation between the oscillators should be always the same in order to produce the correct result.

And AFTER THIS is solved, there would be the question: since a sound wave of a violin changes over time (attack, decay, sustain, release), is there any way to program them to behave this way? This hasnt to be realized but I should try to answer it theoretically..

I hope this isnt too much...

Alex

 

[audioguru]

A musical instrument like a violin has MANY harmonics but you show only one harmonic (its second harmonic that is double the frequency of the fundamental frequency).

You begin with a frequency that is high then divide it down with a digital divider IC to get the harmonics and fundamental frequencies you want, then each must be filtered into a sinewave because the outputs of a digital divider are squarewaves like a buzzer. Then you must adjust the level and phase of each harmonic then mix them together.

A voltage-controlled-attenuator or voltage-controlled-amplifier can produce attack, decay, sustain and release.

I found the waveform of a violin in Google Images. It has only a few harmonics:

 

[MrAl]

Hi again,

synox:
I gave a pretty good overall description of what you would need in post #34 but it seems you did not read that one.
It will take quite a bit of electronics but it can be done.

 

[synox899]

Hello, MrAll.
Your discription was very nice, but I am not sure, if it quite helps me, because as I said, I do not need 12 seperated notes. (Thats why you suggest to make 12 of them i suppose, I hope I got it right, if not correct me). I only need need maybe 4 or max 10 (depends on whether it gets exponentionally harder to it with more) oscillators, which are connected with each other and are factorized by one main oscillator. It is supposed to be just a presentation and a simple simulation of fundamental frequency and the overtones and not really a sophisticated model of a digital violin. That wont be possible since I only have 6 days till the presentation and am not that bright in electronics.

An oscillator uses noise in the circuit to get it going. But a simulation program has no noise, you must "kick" the circuit with a pulse to get it going.

Your circuit has a few things wrong, I fixed it and explained why:

I will try this circuit in RL on Monday, I think it might work, but could it be advanced in a way to be able to controll the pitch with an antenna? Replace the one condensator with the antenna? (Sorry, im really not that good in electronics :S)

Alex

 

[Nigel Goodwin]

Hello, MrAll.
Your discription was very nice, but I am not sure, if it quite helps me, because as I said, I do not need 12 seperated notes. (Thats why you suggest to make 12 of them i suppose, I hope I got it right, if not correct me).

No, we've all suggested 12 - because that's how many notes are in an octave, and you generate the lower notes by division from those oscillators (at least in an electronic organ).

Your 'idea' isn't ever done because it's vastly over-complicated and very difficult to achieve (and you would need your 4 to 10 oscillators for every single note - so 48 to 120 oscillators for a single octave).

 

[synox899]

I am not sure, if talk at cross purposes... I really hope we get it right.
I know there are 12 notes in an octave. But if I am not targeting to have 12 separated notes but just a sine wave that can be changed to any frequency (for example between a c and a #c), why is there a need to create 12 oscillators?
A single oscillator could produce any frequency by changing the capacity or the value of the inductor. Thats what I mean - a single wave. And then connect a few more of them, in order to simulate the overtones of a violin.
For example if the main oscillator creates a 294 Hz (d'), the other oscillators should create 588 Hz, 882Hz and so on. I am missing something or do you?

Alex

 

[synox899]

I feel that you think I want to make something more complicated than it really is. ^^

 

[Nigel Goodwin]

I am not sure, if talk at cross purposes... I really hope we get it right.
I know there are 12 notes in an octave. But if I am not targeting to have 12 separated notes but just a sine wave that can be changed to any frequency (for example between a c and a #c), why is there a need to create 12 oscillators?
A single oscillator could produce any frequency by changing the capacity or the value of the inductor. Thats what I mean - a single wave. And then connect a few more of them, in order to simulate the overtones of a violin.
For example if the main oscillator creates a 294 Hz (d'), the other oscillators should create 588 Hz, 882Hz and so on. I am missing something or do you?

You're missing what everyone has been telling you all along (and why it isn't done), in order for such a scheme to work all the oscillators have to be precisely synchronised together at all times - which is why you would need multiple oscillators for every note.

But as you think it's so simple, feel free to go ahead and build one

 

[synox899]

You're missing what everyone has been telling you all along (and why it isn't done), in order for such a scheme to work all the oscillators have to be precisely synchronised together at all times - which is why you would need multiple oscillators for every note.

But as you think it's so simple, feel free to go ahead and build one

Well, I got that as well, I think there is a misunderstanding here: "...For example if the main oscillator creates a 294 Hz (d'), the other oscillators should create 588 Hz, 882Hz and so on."
Isnt it that we are all talking about? I got that there should be several synchronized oscillators. My problem was, that you were talking about 12 oscillators, what I didnt really got why. 4 would be okay as well.
Well, this is a secondary problem anyways.

I really rely on your help, so please be kind with me, I really try do understand the sense behind it, because I also have to know why its done this way (its still a presentation).

So... let's start over... What frustrates me in the first place is that none of the oscillators Ive ever tried in the simulation program really worked - audioguru's and cowboybob's as well. As soon i connect the ground in the bottom right with the transistor, Yenka starts to act awkwardly, the curser loads all the time... Probably I did something wrong but cant identify it.

So I am still stuck in the beginning, because none of all oscillators Ive found on the Net work for me. Maybe the program is "crap"? I dont know... do you know any other simulation program? I am going to try it on LTspice..

 

[alec_t]

I feel that you think I want to make something more complicated than it really is.

We think it really is more complicated than you think it is
Here's a quick-and-dirty relatively simple way of getting a tunable fundamental frequency square wave plus its second and third 'harmonics'. You would still have the problem of filtering each signal to get something resembling sine waves, and tunable filters would be reasonably complex in themselves.

 

[alec_t]

Probably I did something wrong but cant identify it

No inductance values of the primary and secondary windings of the transformer are shown. What have you defined them as?

 

[Nigel Goodwin]

No inductance values of the primary and secondary windings of the transformer are shown. What have you defined them as?

And an 11uF tuning capacitor

 

[alec_t]

And an 11uF tuning capacitor

That must be a close tolerance one: 10uF +10% tolerance

 

[synox899]

Oh.. i misread this

as 11 u. My bad. But the problem with the never-ending loading still lies somewhere else...

Transformer inductance values... I think this program is not that advanced.. ^^ I can only vary the windings.

The scheme you have done looks very nice, but I dont know how to read it (its because of my low knowledge). Would it be too much expected from you to explain it in a simple way? As I see, there are not obvious inductors, so what do these "boxes" do? I am bothered by the feeling that i go on your nerves, because I dont understand everything that fast. I hope you can tolerate it for some time.

Alex