Re: [Linuxsampler-devel] Envelopes , filtering and zipper noise

[Frank N. <bea...@we...>]

Hi,
Steve wrote:

> BTW do we know for sure than samplers use exponential envelopes? I guess we
> need linear ones too, but they are easy to implement. We should probably
> get some recordings from samples of high freq sinewaves with envelopes.
> I think Frank N. has done this allready for the S2000, Frank are you on
> the list?

Yes, I'm here :-). I just made a short example of this for you; I hope this
is what you were asking for, I have a hard time trying to catch up with all
the mailing lists :-).

I loaded a sine wave into the sampler, set the program's attack and release
times to 0 and played a couple of high-pitched notes, sampled this into the
PC (44.1 kHz) and had a closer look at the attack and release phase of the
sample (for analysis, Conrad Parker's "sweep" is highly recommended; I just 
learnt recently that you can even vertically zoom into a sample by using
Shift+Cursor up/down, which helps a lot in finding the exact starting point
of a sample :-)

Please find the following files for analysis:
http://lakai.sf.net/sinetest.wav.gz (the sample, 440 KByte packed)
http://lakai.sf.net/sine1_attack.png (a screenshot of a close look at the
attack phase)
http://lakai.sf.net/sine2_release.png (the same for the release phase)

It was interesting for me to discover that the release phase is always
faded out somewhat exponentially (and "rather slowly", taking about 2.2ms),
while the attack phase is very short; looking _very_ close at it, I'd say
it's about 28 samples long, ~0.6 ms. But both are _not zero_ (so really 
"hard clicks" can never occur). That's the same behaviour I'd expect from
any software instrument, too.

If you need more analysis, let me know; however, I'll be on a company trip
(to Bristol - yeah, UK, unfortunately not So'ton :-}) from Sunday - 
Thursday, so don't expect answer before Friday.

Greetings,
Frank