[sdiy] Analog bandwidth

[Tom Wiltshire]

On 22 Feb 2014, at 13:45, cheater00 . <cheater00 at gmail.com> wrote:

> On Sat, Feb 22, 2014 at 1:47 PM, Michael Zacherl
> <sdiy-mz01 at blauwurf.info> wrote:
>> 
>> On 22 Feb, 2014, at 1:33 PM, cheater00 . <cheater00 at gmail.com> wrote:
>> 
>>> I believe adding shaped, ultrasonic noise before a VCA is a way to
>>> shape transients during fast A/D/R stages. It should also work in a
>>> compressor (you must not add the noise to the sidechain, though).
>> 
>> I wonder if this also applies when using programming environments, like
>> SuperCollider, Max/PSP, Puredata, Csound etc. In theory it should, given
>> that the basic elements used are pure and 'un-shaped'.
>> 
>> Michael.
> 
> Yes, but when doing multiplication or non-linearities or IMD in
> digital the typical rules for oversampling apply. So you'd have to
> take, say, a signal that goes up to say 100 kHz, and then oversample
> it quite a bit.. that's computationally expensive. Why 100 kHz? For
> example in multiplication you obviously get sum and difference
> components. So depending on the frequency content of the envelope of
> your VCA, you'll at most push the noise (which is at 20 kHz - 100 kHz)
> down by that much. If you want a nice thump you'll make the envelope
> push down the noise by those 20 kHz or more. As the envelope goes
> faster, I imagine you would get typical "overdrive" type effects
> happening on the noise.
> 
> 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.

Sorry, I'm missing something. Why is ultrasonic noise supposed to help with a fast attack?

Thanks,
Tom