[sdiy] Novation peak NCOs

[Gordonjcp]

On Fri, Apr 21, 2017 at 12:34:34AM +0100, Tom Wiltshire wrote:
> 
> On 21 Apr 2017, at 00:22, Michael Zacherl <sdiy-mz01 at blauwurf.info> wrote:
> 
> > 
> > On 21.Apr 2017, at 1:13 , Tom Wiltshire <tom at electricdruid.net> wrote:
> > 
> >> That's equivalent to the 1199th harmonic of  20KHz ramp wave.
> > 
> > which also then would be just more than 60dB below the level of the 1st harmonic of a ramp wave, IIANM. 
> 
> Yep, that's what I got too. 60dB below the level of a harmonic which is almost an octave above what I can realistically hear seems good enough to me!
> 
> If you were actually bothered about the aliasing at that level, use a "band-limited" waveform with only 500 harmonics. No-one will ever know (20x500 =10KHz). My own listening tests suggested that it's already hard to tell the difference between band limited waveforms with 100 harmonics and 150, and by the time you're at 200 harmonics, it's all just high end fizz.
> 

Or just use a polyblep, which suppresses aliasing by an acceptable amount even at low audio rates (24kHz is quite usable) and bit depths.  You could afford to reduce the sample rate to something far more reasonable - say, 8MHz - and run dozens and dozens of utterly alias-free sawtooths.

Right, I'm off to buy an FPGA dev board.

-- 
Gordonjcp MM0YEQ