KVR Audio

Concretefx · Post by **Concretefx** » Wed Mar 10, 2004 11:54 am

Hi there

AM is interesting , I'll have to have a look into it

Cheers

Jon

no_signal · Post by **no_signal** » Wed Mar 17, 2004 5:28 pm

Just listened, I think they sound and look OK.
I think i'll experiment with it.
Thanks,
A.F.

stefancrs wrote:
Can you show us some examples of your waveform ?
observe: my waveforms are not alias free, nor are oversampled waveforms, they are just anti-aliased (reduced aliasing).

ftp://217.215.143.219/

both wav's and mp3's
duty cycle modulation on all the examples, except on the latter part of the triangle example where it is a pure sawtooth all the way. nothing but one oscillator is active. It was a few years since I wrote this softsynth, so it's not really up to date with anything

stefancrs · Post by **stefancrs** » Wed Mar 17, 2004 5:49 pm

Well. I have replaced them by now with self-phase-feedback oscillators or whatever one should call them

I found that method way cooler and more flexible. (way less amounts of code, less cpu usage, higher sound quality)

no_signal · Post by **no_signal** » Thu Mar 18, 2004 3:05 pm

stefancrs wrote:Well. I have replaced them by now with self-phase-feedback oscillators or whatever one should call them I found that method way cooler and more flexible. (way less amounts of code, less cpu usage, higher sound quality)

You mean, the samples I listened yesterday : trianglething... are not using that method you were talking about. They sound really analog to me... What is it all about ? Can you send me a short code to get started ? Please

Cheers,
A.F.

stefancrs · Post by **stefancrs** » Thu Mar 18, 2004 3:11 pm

The stuff you listened to was my old oscillators, the ones using integration and stuff. Now I use the y = sin(phase + y * feedback) and the different variatons of them. You don't need no code-snippet to try those out.

no_signal · Post by **no_signal** » Thu Mar 18, 2004 3:35 pm

stefancrs wrote:The stuff you listened to was my old oscillators, the ones using integration and stuff. Now I use the y = sin(phase + y * feedback) and the different variatons of them. You don't need no code-snippet to try those out.

Nice, i will try for sure

Hope I'll have the time.

Taika-Kim · Post by **Taika-Kim** » Thu Mar 18, 2004 6:07 pm

Muon Software Ltd wrote:
This is why I have to assume your implementation might have gone wrong somewhere - even putting aside my own work, there is Iblit and PolyIblit out there that sound OK to me, not mention Mercury 1.

Funny, soundwise those synths are my absolute favourites :)

no_signal · Post by **no_signal** » Thu Mar 18, 2004 6:52 pm

stefancrs wrote:The stuff you listened to was my old oscillators, the ones using integration and stuff. Now I use the y = sin(phase + y * feedback) and the different variatons of them. You don't need no code-snippet to try those out.

Just tryed. Nice, waveform ...
BUT: It's way too mouch bandlimited. For 440 Hz it's limited at ~5KHz. So how did you made it ? If the feedback is too big, it distorts the whole thing.
I'm thinking of creating 1-4 harmonics of this wave and mix them using amplitude 1/n (n=1...harmonics).
Any other methods ?
Thanks,
A.F.

stefancrs · Post by **stefancrs** » Fri Mar 19, 2004 8:05 am

Well, actually it isn't bandlimited, it's just a matter of not having that many overtones in it. You will get aliasing at really high tones, you could avoid this by decreasing the feedbacks at higher frequencies (I will try that out later). And I have found that you get more overtones with negative feedback than with positive feedback. Or if it was the other way around :)
If you like the sound of it, use it. I have these different waveforms (shape goes from 0.0 to 1.0)
sine->saw
out = sinf(phase + lastOut * (shape - 0.5f) * 2.0f);
lastOut = out;

saw->square
out = sinf(phase + lastOut * (1.0f - shape) * + lastOut * lastOut * shape);

pwm square1
square1 = sinf(phase - lastOut1 * lastOut1);
lastOut1 = square1;
square2 = sinf(phase + shape * pi * 2 - lastOut2 * lastOut2);
lastOut2 = square2;
out = (square1 + square2) * 0.5f;

pwm square2
float square1 = sinf(phase - lastOut1);
lastOut1 = square1;
float square2 = sinf(phase + shape * pi2 - lastOut2]);
lastOut2 = square2;
out = (square1 - square2) * 0.5f;

though they aren't exactly like this, I have tweaked the feedbacks everywhere, having them at like 1.1 instead of 1.0, giving for instance float square1 = sinf(phase - lastOut1 * 1.1f); in the pwm square2. Some trial and error :)

no_signal · Post by **no_signal** » Fri Mar 19, 2004 9:15 am

stefancrs wrote:Well, actually it isn't bandlimited, it's just a matter of not having that many overtones in it. You will get aliasing at really high tones, you could avoid this by decreasing the feedbacks at higher frequencies (I will try that out later). And I have found that you get more overtones with negative feedback than with positive feedback. Or if it was the other way around
If you like the sound of it, use it. I have these different waveforms (shape goes from 0.0 to 1.0)
sine->saw
out = sinf(phase + lastOut * (shape - 0.5f) * 2.0f);
lastOut = out;

saw->square
out = sinf(phase + lastOut * (1.0f - shape) * + lastOut * lastOut * shape);

pwm square1
square1 = sinf(phase - lastOut1 * lastOut1);
lastOut1 = square1;
square2 = sinf(phase + shape * pi * 2 - lastOut2 * lastOut2);
lastOut2 = square2;
out = (square1 + square2) * 0.5f;

pwm square2
float square1 = sinf(phase - lastOut1);
lastOut1 = square1;
float square2 = sinf(phase + shape * pi2 - lastOut2]);
lastOut2 = square2;
out = (square1 - square2) * 0.5f;

though they aren't exactly like this, I have tweaked the feedbacks everywhere, having them at like 1.1 instead of 1.0, giving for instance float square1 = sinf(phase - lastOut1 * 1.1f); in the pwm square2. Some trial and error

Yes, I noticed that a feedback of 1.1 makes better results. And at higher freq i will need to reduce the feedback. The results were nice, but still not bright enough. And the waveform was not really saw, it has is V- peak rounded a little. I also compared with a pure saw wave made with FFT, and this sounded mouch cleaner, but darker also.

helium · Post by **helium** » Fri Mar 19, 2004 9:21 am

stefancrs is a MSVC user

no_signal · Post by **no_signal** » Fri Mar 19, 2004 10:27 am

helium wrote:stefancrs is a MSVC user

Ok...But why are you telling me this ?

stefancrs · Post by **stefancrs** » Fri Mar 19, 2004 10:32 am

And the waveform was not really saw

Most aren't if you by saw mean a linear sloping |\|\|\| or /|/|/|. Take a look at a few from different synths.

helium · Post by **helium** » Fri Mar 19, 2004 10:39 am

'cos it's so damn boring here. And I just found it out by looking at

Code: Select all

sine->saw 
out = sinf(phase + lastOut * (shape - 0.5f) * 2.0f); 
lastOut = out;

no_signal · Post by **no_signal** » Fri Mar 19, 2004 6:23 pm

I found a way to make it more like a saw wave...
Instead of phase feedback, i feedback the waveform itself. The waveform it's closer to a saw wave, the spectrum is bright. Little more alias.

out =(sin(phase + out* feedback*3)+(out* feedback*3)+0.78)*0.3;//0.78 (tweaks for DC offset)

Oscillator, about wavetable size and interpolation