KVR Audio

karmafx · Post by **karmafx** » Tue Sep 21, 2004 7:39 am

I tried the minblep code from the archive for generating bandlimited saw and square waveforms. Works excellent.
Now, can anyone explain how to generate other waveforms than saw and square, e.g., invsaw, super-saw and triangle (with pw) with this code ?

Thanks a bunch !

- KarmaFX

stefancrs · Post by **stefancrs** » Tue Sep 21, 2004 7:45 am

Supersaw is just a bunch of detuned saws added up together.

Apart from that, can't you basicly make any "function based" (out = waveform(time) kinda) bandlimited using minbleps at the discontinuites?

karmafx · Post by **karmafx** » Tue Sep 21, 2004 7:52 am

stefancrs wrote:...can't you basicly make any "function based" (out = waveform(time) kinda) bandlimited using minbleps at the discontinuites?

Maybe...my guess is that it is really simple... but since I don't know how blep's work, I don't know how ? Any code or explanation appreciated

Thanks.

stefancrs · Post by **stefancrs** » Tue Sep 21, 2004 8:06 am

Basicly, what you do when you use minbleps is that you apply those in the discontinuities of the waveform to "simulate" a discontinuity in the analouge realm (they don't exist, I'm just trying to explain here). Hence you remove the aliasing a normal, "ugly", signal step would introduce.

Do I make any sense? Otherwise I guess you'd have to read the papers better, or check out the previous thread on the subject:

http://www.kvr-vst.com/forum/viewtopic.php?t=45778

karmafx · Post by **karmafx** » Tue Sep 21, 2004 8:34 am

Cool, so this means that is is possible to produce "bandlimited" versions of any kind of signal...even presampled stuff,e.g. in a wavetable ?

Is this easy to derive from the code in the archive...I'll have a look at the paper now...but I'm not really sure what discontinuities you are referring to ?

Thanks!

stefancrs · Post by **stefancrs** » Tue Sep 21, 2004 8:40 am

A discontinuity = a step, as when the square wave suddenly goes from -1 to 1. A perfect square wave does this in zero time, hence it has an infinite spectra, this will cause aliasing so you have to use lowpass filtered steps instead. That's what the minBleps are.

Read, and understand, http://www-2.cs.cmu.edu/~eli/papers/icmc01-hardsync.pdf

No, you can not use it on sampled waveforms since you have no idea about the derivatives etc.

karmafx · Post by **karmafx** » Tue Sep 21, 2004 9:09 am

Okay, I'll read the paper !
Though, I guess I was hoping for a "do-this-and-it-will-work" kinda answer

But hopefully I'll learn something new in the process.
Thanks again, stefan.

texture · Post by **texture** » Tue Sep 21, 2004 11:35 am

Would it work to integrate a bandlimited square wave in order to generate a bandlimited triangle?

mauseoleum · Post by **mauseoleum** » Tue Sep 21, 2004 11:47 am

hey, karmafx, your delay sounds lovely! simple and nice, clean sound.

karmafx · Post by **karmafx** » Tue Sep 21, 2004 12:59 pm

texture wrote:Would it work to integrate a bandlimited square wave in order to generate a bandlimited triangle?

Thanks, Yes I'm guessing it would... I seem to recall reading that somewhere too...but I dont know how to do that ?! Any thoughts ?

karmafx · Post by **karmafx** » Tue Sep 21, 2004 1:01 pm

mauseoleum wrote:hey, karmafx, your delay sounds lovely! simple and nice, clean sound.

Thanks man! Glad you like it

I'm working on a reverb effect and a synth now...

karmafx · Post by **karmafx** » Tue Sep 21, 2004 5:51 pm

texture:
Okay, I just tried what I think you suggested. I simply sum the square wave contribution for each getsample iteration and output that (scaled) as triangle output, and resetting the sum when resetting the phase.
I get something that looks somewhat like a triangle, but not quite... the result sounds like it is bandlimited though - which I guess makes sense.
Has anyone tried this ? and got good results ???

stefan:
I read the (slightly confusing

) paper...but think I grasped the general idea, however, I'm still not sure where to add bleps for a triangle wave... I mean saw and square waves have big discontinuites, but the triangle does not. Any hints anyone ?

stefancrs · Post by **stefancrs** » Tue Sep 21, 2004 6:45 pm

But the triangle too has discontinuites, every time it "changes direction". As far as I understand the paper, you have to pretend to switch between ramps instead.

The interpolation method works so-so. When resetting the sum value you produce a discontinuity, hence it will alias.

I'd go for solving a duty cycle ("symmetry") modifiable triangle wave instead, then you'd "automagically" have the sawtooth covered as well. It is doable with ramps and so on.

Note: I have never coded anything with minBLEP's. So I'm no expert in any way at all.

dwerner · Post by **dwerner** » Wed Sep 22, 2004 4:02 am

karmafx wrote:Maybe...my guess is that it is really simple... but since I don't know how blep's work, I don't know how ? Any code or explanation appreciated
Thanks.

I have code for generating MinBLEPs posted on my site at http://experimentalscene.com/dsp/minbleps.php

Enjoy

karmafx · Post by **karmafx** » Wed Sep 22, 2004 10:09 am

dwerner wrote: I have code for generating MinBLEPs posted on my site at http://experimentalscene.com/dsp/minbleps.php
Enjoy

Yes, I visited your site yesterday actually

... Looks cool, but it doesn't clarify how to build a triangle wave with minbleps ?!
Do you have code snippet that generates that ?
Eli talks about it in the end of his paper, but does not go into detail.

Also, I'm curious.... in the text you write that you add minbleps on both side of the discountiuity... but (as far as I can see) the MinBLEP code in the archive only adds a single blep. Is the code in the archive wrong, or am I missing something ?

Thanks.

Other waveforms with MinBLEPs ?