Quite interesting result of “disassembling” square wave with Fourier transform and “assembling in different order”
Insane! That’s getting added to my mental list of “things I didn’t know I needed to know, that I now know”, lol
Also, someone in the comments pointed out, the wave will interact differently with distortion and also sound differently in lower octaves. I just toyed around with it in Vital, and you can get some pretty old-school sounding imperfect square waves by randomly phaseshifting different frequencies and passing them through a clipping distortion to get back the volume. Sounds like something straight out of Atari and NES sound chips, cause I suppose the sound chips themselves on console or/and on the tv produced a ton of distortion. Playing several notes close together or in a chord and pitch shifting is completely wild, though perhaps everything I just outlined isn’t a surprise to any chiptune veterans out there.
clipping distortion to get back the volume. Sounds like something straight out of Atari and NES sound chips
A-ha! that’s why all my chiptune-like attempts sounding way too clean. I’m feeling dumb now.
Hey man, I only know about distortion in old hardware from documentaries, it surprised me too when it produced the sound straight out of a console! I just tried one patch, but there is probably more you can do with triangles and perhaps some FM-synthesis, that’s something to feel excited about!
I’m glad I’m not the only one who was thinking the square waves sounded cleaner than the other ones. I feel like the alternative ones would create a lot more noise in chords than it does as a single line, the way a whole bunch of people using vibrato sound more out of tune than they would be without vibrato. But I have no easy way to check right now.
The alternative squares do exactly that, and the effects become way more obvious with distortion. Ultimately, changing phases on the same frequencies you need to build a square changes the volume due to phase cancellation and introduces some micro-changes to the sound which is what impacts the way the sound interacts with processing (distortion, compression, all that).
Also, using less frequencies to build a square overall will produce a softer and nicer square sound. Modern software really pushes it when building a proper square by adding a ton of high frequency sines to make it “textbook square” (something that wasn’t done in older hardware due to limitations). That sharpness can be easily removed to get essentially a square that doesn’t cut into your ears. Something to keep in mind if you feel the need to lowpass your squares later in the chain.