[PD] click remover
noise.now at gmail.com
Thu Jan 8 06:35:55 CET 2009
click remover in the "practical sense" of the meaning work quite
well. as practical meaning I refer to what is written here
10 years ago we spent fortunes for cedar hardware processors (which
did and do an incredible good job). what is difficult is the
difference between "wanted" transients" like very short sounding
percussion and "discontinuities" like unexpected jumps in level,
waveform etc. In order to make the difference it takes complex
analysis processes that determine tempo so that "musically logical
discontinuities" are untouched - even if they're not metronomically
Low-pass cannot do the job because the "click" is typically broadband
- but you're right, we may perceive a click as high frequency. In
this regard, the process would identify a 20 kHz saw tooth as
musically correct because it is periodic. clicks are not.
On Jan 8, 2009, at 10:31 AM, Mathieu Bouchard wrote:
> On Wed, 7 Jan 2009, Charles Henry wrote:
>> Wait... you said discontinuities. I've got something that takes
>> out impulses...
>> A low-pass filter should do a reasonable smoothing, but it would take
>> a lot away from the signal, in the other spots.
> In the discrete domain, a loud very-high-frequency tone is
> indistinguishable from a sequence of discontinuities.
> So, what is it that people are calling a "discontinuity" in this
> If discontinuities are so hard to remove, it's because they don't
> correspond to anything in objective reality. It's all in the mind.
> _ _ __ ___ _____ ________ _____________ _____________________ ...
> | Mathieu Bouchard - tél:+1.514.383.3801, Montréal,
> Pd-list at iem.at mailing list
> UNSUBSCRIBE and account-management -> http://lists.puredata.info/
More information about the Pd-list