R3 osc test


(BpZ) #1

Hi
I test one cycle R3 osc with sxx command and i find some strange behavior.
So i make a small pattern (16 step or loop point) and i use two track with the same wave.
At the first i write a note C-4 and at the second C-4 with S80.
Now i try the sine and triangle hear fine (cancellation phase).
But in track scope still i have the wave view.
Lets try the saw and square…Wow i have an octave up!!!
Why this?


(oneunkind) #2

It depends on the waveform’s symmetry. Curious, what did you expect them to do?


(BpZ) #3

It depends on the waveform’s symmetry. Curious, what did you expect them to do?

Phase reverse and this means no sound (like sine and triangle).


(oneunkind) #4

But, a phase shift is not the same as flipping the polarity. It’s true that if the waveforms are symmetrical you will get complete cancellation at 180 degrees, just as you would from an in phase waveform of opposite polarity.

Assymentrical waveforms will not add and subtract at each sample to always equal zero so you will get different results.


(BpZ) #5

Why do you think that the square and saw wave are asymmetrical.

Every wave symmetrical or not have the same behavior.

Make a wave saw square or a 25%square (asymmetrical) and duplicate this,use the invert phase button at the sample editor and write the first at the 1st track and the ather at the 2nt track put the play and you will hear (NO SOUND).

I use 128 nambers of samples.


(oneunkind) #6

Let’s build a peculiar (and underpowered) sampler that samples every waveform exactly 8 times. It’s called the Bitzoblaster 301. The BB 301 uses Bitzoblorps to measure amliptude. Bitzoblorps go from +12 to -12.

Here’s what seems to happen when we use it in the same way as you used Renoise:

0 deg. +12 +07 +04 +01 -02 -05 -08 -11
180 deg. -02 -05 -08 -11 +12 +07 +04 +01
----------------------------------------------------------
 +10 +02 -04 -10 +10 +02 -04 -10

See. Simple addition has doubled the frequency, although it has made it a little quieter.

(Ahh, but I do not know how you got the square wave to do what it did, if it is a true square wave.)

Why do you think that the square and saw wave are asymmetrical.

Every wave symmetrical or not have the same behavior.

Make a wave saw square or a 25%square (asymmetrical) and duplicate this,use the invert phase button at the sample editor and write the first at the 1st track and the ather at the 2nt track put the play and you will hear (NO SOUND).

I use 128 nambers of samples.

OK, but you said you used the Sxx command which does not ‘invert’ phase, but shifts it.


(BpZ) #7

Shaper kaihttps://forum.renoise.com/t/new-tool-3-0-selection-shaper/41240


(BpZ) #8

Yes i use the sxx for sift to phase… where you glue?

I think i was clear at the start.


(oneunkind) #9

Make a wave saw square or a 25%square (asymmetrical) and duplicate this,use the invert phase button at the sample editor and write the first at the 1st track and the ather at the 2nt track put the play and you will hear (NO SOUND).

And this…

So i make a small pattern (16 step or loop point) and i use two track with the same wave.
At the first i write a note C-4 and at the second C-4 with S80.

Are not the same thing, although they may sometimes produce the same results. It seems to me you are saying they should do the same thing. But they really aren’t supposed to do the same thing.

Anyway, the unexpected square wave results are due to a small error (sampling error, generation method, sample length size, something, perhaps Sxx not being able to cut it at exactly the correct point) that has caused the two ‘halves’ to be very slightly different lengths. (Or due to sample length, S80 does not find exactly the correct point?) If playing these two together with one phase inverted and the other not, you would not notice. But combining them as two instances of the same wave, but one moved over 180 degrees you will notice and get different harmonics depending on the resulting pulse width.

Similar tiny errors might cause sines and triangles to almost cancel but not completely.

Everything seems to be working to me as expected. Any ‘problems’, in my opinion, are inherent to sampling in general and not Renoise in specific or the maybe the nature of the Sxx command (256 divisions not working out just right sometimes).

(thanks for the nifty tool link… in this case the zero crossing of the square created seems to be one sample off? what do you think?)