[sdiy] Analog bandwidth

[cheater00 .]

On Sun, Feb 23, 2014 at 3:22 AM, Andrew Simper <andy at cytomic.com> wrote:
>> Andy said he was going to check whether adding some ultrasonic noise
>> makes sense for fast attack settings in his Glue VST but I don't
>> recollect if we spoke about that afterwards.
>>
>> Cheers,
>> D.
>
>
> I have no idea what you are talking about and have never considered
> adding "supersonic" anything to signals for any purpose.

We were having lunch near your old flat in Angel when I presented the
idea to you, and I remember you got some nice chunky fries which ended
up being much better than whatever side I was having. It was likely a
passing thought for you, but I was hoping you may have taken this
somewhere. Anyways, luckily enough we have spoken about ultrasonics,
whereas we haven't spoken about actually accelerating anything of
yours above Mach 1.2 :-) But that might have been a fun experiment :-)

> It is possible to anti-alias non-bandlimited discontinuities if you
> can estimate the discontinuity and add a corrective grain to
> band-limit it. This is insanely hard and it ends up being  cheaper
> just to oversample in the first place unless you have a priori
> knowledge of the signal being processed, which is the case with a
> sawtooth / pulse etc oscillator.

I remember seeing someone talk about this method a few years back and
have been wondering why it never went anywhere. Thanks for summing it
up.

Cheers,
D.