Are synth square waves more prone to phase cancellation?

[Sundown]

Hey guys,

 

I"m trying to finish up a long overdue track, but the opening 16th-note arpeggio sounds harsh and unpleasantly resonant to me. It"s a single oscillator square wave sample from my Roland XV-3080 and while I"ve EQ"d it to death, I may need to replace the sound. Fabfilter"s Q3 will allow you to perform laparoscopic surgery, but as much as I like the timbre, it"s still harsh no matter how many cuts I apply.

 

Every time I model a similar square wave patch on a VSTi or my Kurzweil (with VA-1), I seem to get noticeable phase cancellation. Any overlap of notes results in some phasey behavior.

 

Is there something inherent in square waves that might make them more susceptible or prone to phase cancellation?

 

Thanks in advance.

 

Todd

[KuruPrionz]

I am just guessing but true square waves go straight from positive peak to negative peak, they are never in transition.

Exactly half the duration of the tone is at full positive peak and the other half is at full negative peak.

 

With a sine wave, the positive and negative peaks are very brief.

 

A phase discrepency in a square wave should be more noticeable, particularly if the out of phase is also a square wave. The duration of full intensity is longer.

[Mike Rivers]

Two square waves that are exactly out-of-phase would cancel completely, just like two identical copies of any waveform.

 

What may be happening in your case is that the the two presumably identical waveforms are slightly displaced in time, not 180 degrees of time offset. So what you're getting is a spike when both are on, and a null when they're both off, an entirely different waveform that what you started out with.

 

Use your DAW as an oscilloscope. Feed the analog output of the mix to an analog input to your computer interface, record it, and look at it.

[Markyboard]

Try adjusting the pulse width for one or both square waves. They"ll still 'phase' but perhaps in a more desirable way.

 

Edit: never mind as I re-read the op and see this is a single oscillator patch.

[The Real MC]

Try setting one of the square waves at different amplitude, then they won't cancel.

[Nowarezman]

Ok, if you've got an arpeggiator running 16th notes and there's some overlap note to note, first question is - do you want the overlap? Or do you just want legato? If you can shorten the notes to legato with no overlap, then no phasing can happen.

 

You can also dial down the release on the synth ADSR for the amplitude and/or filter.

 

Also might try, if your synth will do it, and your velocities are all pretty close together - route the velocity to move the filter cutoff a little bit up and down. Just that in itself might clear up phasing.

 

If you want some overlap, you might still go legato and put a reverb or delay on the track to smear things up a bit. You probably won't hear any phasing in the complex sounds that will result.

 

I bet you can clean it up. It ain't the waveform's problem, it's the interaction between two waveforms. Get them far enough out of synch with each other and phasing will go away. Or get them 100% in synch.

 

nat

[Markyboard]

....It"s a single oscillator square wave sample.....

Todd

 

 

I edited my previous response but thought this part from your initial post worth repeating. I don"t play with samples much but assuming the notes being played are different makes this all the more perplexing.

 

I can see certain notes phasing just based on the randomness of how each square wave waveform from each note lines up. But it could be the arpegiator tempo makes this not so random after all. If you slightly increase or decrease the tempo does that change things? Maybe try modulating the sample start time to skew the waves a bit. Not sure this will help but it"s something to consider.

 

Interesting problem.