yes i notice you say a lot of things.fluffy_little_something wrote: Anyway, I am not saying that ________, all I am saying is that _______
What I am also saying is that _______
native SE sawtooth with gibbs smoothing. those regularly spaced spiky things are harmonics i'm not sure where the 'metal' is on that. my observation is that jeff's using BLIT with linear interpolation (a few years ago i probably could have correctly gauged the size of the table). i like my blit tables closer together and a sharper gibbs dropoff, which is besides the point.
what you are failing to appreciate in your eagerness to demonstrate your powers of discernment is that SE is a modular and multipurpose environment, which makes the BLIT (bandlimited interpolated tables, not the bandlimited impulse train BLIT) a very good choice, exhibiting fair audio fidelity and low cpu. jeff was probably thinking of the little people you aren't, who want to experiment or use SE for the same things people use PD for (eg. god knows what, but probably not dick measuring the purity of a basic oscillator). you can run dozens of these oscillators on a pc that's old enough to get a driver's license.
an impulse train BLIT or MinBLEP would have a lower noise floor, but the cpu of these methods is several times higher. i'll venture jeff left it at one oscillator type for a few reasons... (1) SE is overtly user expandable (2) dude has stated he had an ulterior motive of luring people into c++ programming via SE and the SDK (3) n00bs would be all "WTF?!?!?!" when confronted with multiple osc choices (4) and maybe he figured the noise floor was good enough for practical purposes (which used to not include dsp snobbery at kvr, but maybe that's all that's left in the world, i don't keep up).
do i work for synthedit? i work for people who can use synthedit. it's a tough job because people keep dumping this ignorant dismissive bullshit in threads, dude. hating on shit they don't even know about like a punk.