madtheory wrote:Seems to me that someone is not using the 'scope properly, for example it's easy to ignore slight level variations over time on older/ analogue 'scopes (which a lot of people use for audio work because they're fine for that and inexpensive).
I don't think so mate.
It also sounds the exact same and it's also the same in theory. In addition to looking the same on a scope.
madtheory wrote:So I guess anyone who thinks PWM via sync sounds different to PWM via LFO is suffering from a perception bias.
See? That's much better
Also, you still need an LFO with the sync method.
DesolationBlvd wrote:There are two aspects of sync-PWM making it sound subtly different from regular PWM.
First, in regular PWM, the amplitude decreases towards half as you approach 0/100% pulse width. In sync-PWM, there is no amplitude loss as you approach 0/100% pulse width.
Second, you can't get the sound of modulating from, say, 20% to 80% pulse width when using sync to generate PWM. The other half of the range is the sync sound, not the PWM sound, as the pitch is raised from the original.
I see what you mean now - I did notice that the top half of PWM on my 3P starts adding in a second pulse (and therefore an octave up). In practice though, you could just modulate from 20% to 50% at half the speed. I can't say I've used a synth with regular PWM, so I don't know about amplitude loss.
I still have to say - if regular PWM and sync-PWM look pretty much the same on a scope, how can they sound different? Seems to me that people are hearing inherent differences between two different synths than between the two different types of technology.
Regarding the two points from DesolationBlvd that I've just written off as "wrong" before, some clarifications.
1) It's true that you only get "silence" in one direction, when the slave is more than two octaves down compared to the master. But you do get it there and that means you can emulate the classic PWM "fade in/out" effect (if you want).
2) This only applies (more or less-ish) when both oscs are the same frequency. As written before, two octaves down on the slave gives silence, one octave up "folds" in the first second-order pulse. Thus, as also mentioned earlier, if you tune the slave down half an octave (i.e. 6 to 7 steps), you get the widest range. Which is easily "20-80%".
Note that I'm not at all talking about the specifics of the JX3P. Anywhere in the thread.
I also hope I didn't confuse anything now, writing it all up from my office chair right now, with no synth in sight.
I've tested this extensively on my A6 that can do both "methods", using the same LFO and so on, and it sounds and looks exactly, precisely, 100%, on the spot, the exact same. And theory tells us that that's right as well.