[sdiy] Digital audio demo on Youtube

[ASSI]

On Tuesday 18 February 2014, 09:18:02, MTG wrote:
> I don't think this guy does any favours at 8:00 when he says that
> Wikipedia is confusing, "the stair-steps aren't really there".

Because they aren't.  His point was that the sample values of a digital 
simply don't stretch out to the next sample.  If your reconstruction 
involves a first-order S&H it introduces distortion, which is the difference 
between the bandlimited signal and the "staircase".  To the to first 
approximation this is a sawtooth at the sampling frequency and your 
reconstruction filter ought to take care of it.

> All the
> (non-1-bit) dacs I see in these old synths generate pretty visible stair
> steps... plus I'm sure there are lots of "control" dacs on the planet so
> we shouldn't just assume every DAC is on a PC motherboard (for instance).

No such assumption needs to be made (the superficial difference between a 
control DAC and one on a PC motherboard is really only a different tradeoff 
between precision and accuracy).  If the digital signal is properly 
bandlimited and the reconstruction filter after the DAC is any good, then 
the reconstructed signal matches the input and any difference is distortion 
and noise.  "These old synths" quite often start off with an already 
distorted signal in the digital domain and whatever their reconstruction 
filter looks like, it's not going to remove all that distortion, because it 
mainly resides inside their pass-band.  Specifically, the staircasing you 
see typically comes from naive oversampling of a low-resolution wavetable or 
sample.  There's no incompetence involved either, these synths simply did 
not have enough horsepower to do this correctly (at least originally, but 
those distortion artifacts are now part of their signature sound).


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

SD adaptation for Waldorf microQ V2.22R2:
http://Synth.Stromeko.net/Downloads.html#WaldorfSDada