[sdiy] Hiss, Crackle, and Pop

[Amos]

If I might offer a small aesthetic suggestion, I think the "crackle"
channel might sound a lot better (er, more crackly) with a simple highpass
filter on it.  Very cool project. :)

On Sun, Apr 2, 2017 at 12:25 PM, Tom Wiltshire <tom at electricdruid.net>
wrote:

> I'm not sure how you'd synthesise the vinyl bump sound, but getting a
> periodic noise wouldn't be hard.
>
> If you used the same firmware in two H.C.P. chips and started them up at
> exactly the same moment, I suppose you might hear something, since they
> follow the same LFSR pattern. However, it would be easy to change the
> random seed in one of the chips at programming and then they'd follow
> completely different paths. The LFSR is 47-bit  so doesn't repeat for 2^47
> bits. It's running at 90KHz or so, and I calculate it a byte at a time, so
> that brings it down to 2^44, but it's still a vast number and a long repeat
> time - over 6 years. Set one at a different point in the pattern and they
> could well be 18 months apart.
> The other significant fact is that the chip uses the internal RC
> oscillator not any external crystal so even if they started in sync they
> won't stay that way for long. The two sets of sample outputs will be close
> to the same rate, but they'll drift in and out. That means there'll be no
> relationship between when the random crackles and pops occur on the two
> chips, even at the sub-millsecond level we're talking about.
>
> Tom
>
>
> On 2 Apr 2017, at 16:19, Mike HEQX <mike at heqx.com> wrote:
>
> > Tom,
> >
> > That's a great idea, but my synths do this on their own LOL
> >
> > If you wanted a cyclic bump like on vinyl what would you need to do?
> Could this thing drive a timer externally? If you had two of H.C.P working
> together would it create any type of convergence that might sound like
> something periodic?
> >
> >
> > Mike
> >
> >
> > On 4/1/2017 3:02 PM, Tom Wiltshire wrote:
> >> On 1 Apr 2017, at 18:53, Neil Johnson <neil.johnson71 at gmail.com> wrote:
> >>
> >>> Hi Tom,
> >>>
> >>>> Yeah, I was probably aiming for something about 30 years earlier than
> vinyl! But I know what you mean, that "old 78s" noise would be lovely. You
> could probably get closer with a bit of fine tuning of the circuit.
> >>>>
> >>>> The mains hum…I dunno. I think it might be from recording the sample
> on the laptop with an unshielded cable. I should really have added a mains
> hum generator into the mix too, with a EU/US 50/60Hz selector!
> >>>>
> >>>> And no, I definitely didn't think of using a dsPIC! A 16-bit
> processor to make the noise of old wax cylinders going round?! Are you
> crazy, man?!
> >>>> No, I used a simple *8* bit processor for a job like that!! I even
> stayed away from op-amps because that just seemed too modern for the job.
> Maybe a 741 would be allowable.
> >>>>
> >>>> But you're dead right - it's one place the background hiss on the
> internal DAC might even be a benefit…
> >>> A few years ago a friend of mine (top-notch DSP engineer with an
> >>> interest in audio) developed a DSP system for recreating the sound of
> >>> 78s (I gather it was a homer project in between designing DSP code for
> >>> satellite comms).  It had everything you could imagine - wow&flutter,
> >>> scratches, hiss, and so on.  It was very convincing.  Not sure where
> >>> it went, but suffice it to say that doing it right is a lot lot lot
> >>> harder than it seems...
> >>>
> >>> Neil
> >> Hohoho! Yeah, I bet! You could definitely throw a pile of
> hardware/software at the problem.
> >>
> >> I had several ideas for extending my basic version - lowpass filtering
> a noise output to give a wandering CV, and then sing that to control the
> volume of the other channels. Obviously this could be repeated several
> times for several channels…
> >> I also think that two channels is still sufficiently simple that your
> ear can differentiate them. If there were three or four or five, I think
> you'd get a noticeably  richer effect.
> >> It might also be fun to build a version with knobs on. Each channel
> could have  a"frequency/probability" control for the basic rate the events
> occur, a "decay" control for the envelope, and then "lowpass" and "high
> pass" filter knobs. Resonance would be a nice touch too. Several channels
> of that would enable you to mix up many flavours of noise.
> >>
> >> There's a certain overlap with some other experiments I've done. One is
> a "granular oscillator" that I wrote as an alternative code load for Bruce
> Duncan's Modcan AHDBDSR envelope generator. That had
> raised-cosine-enveloped pitched grains, and you could alter the randomness
> of the pitch , start point, and the frequency of the grains in the grain
> cloud. It produced a range of textures from sputtery/spattery to rainy to
> sizzly.
> >> The other area that this borrows from is analog percussion synthesis. I
> borrowed Roland's VCA, but the basic idea of a blast of shaped noise with a
> tone filter is the guts of many a drum machine sound. This suggests other
> interesting directions, like trying alternative sources like the "Metal
> noise" combinations of CMOS oscillators used for cymbals, either XOR'd or
> not as the mood or your vintage inspiration take you.
> >>
> >> So…yes, I can definitely understand how they might have spent quite a
> lot of time and DSP power on the problem!
> >>
> >> Tom
> >>
> >>
> >>
> >>
> >>
> >> _______________________________________________
> >> Synth-diy mailing list
> >> Synth-diy at synth-diy.org
> >> http://synth-diy.org/mailman/listinfo/synth-diy
> >
>
>
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at synth-diy.org
> http://synth-diy.org/mailman/listinfo/synth-diy
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20170403/df65e06f/attachment.htm>