Hi,
does anyone has a good suggestion for a very steep filter? Preferably
something that cuts through the frequencies without any rollof at all? (if
possible)
I've tried the iemlib hp10_** filters, also 2 in a row, but the result is
still too gradual. Putting 16 lop~ in a row make makes roughly the same
result. I wanted to avoid fft filters, because they might alter the
original signal. Or is there any other option? (even if it has to be fft)
cheers,
João
chebychev filters look good, elliptic ones look better:
http://en.wikipedia.org/wiki/File:Electronic_linear_filters.svg
yes, I had seen that graphic as well. but trying out the highpass
iemfilters with noise, turns out that 10 hip~s work better.
I looked in my list of pd-ext objects, but found no elliptic filters. are
there any, or any way of getting their coefficients to use with biquad~?
chebychev filters look good, elliptic ones look better:
http://en.wikipedia.org/wiki/File:Electronic_linear_filters.svg
Hallo, Jo?o Pais hat gesagt: // Jo?o Pais wrote:
yes, I had seen that graphic as well. but trying out the highpass
iemfilters with noise, turns out that 10 hip~s work better.
Hehe, I guess that's because you filter noise: It's harder to recognize the noise that the 10 hip stages added. :)
I looked in my list of pd-ext objects, but found no elliptic filters. are there any, or any way of getting their coefficients to use with biquad~?
Elliptic filters are rather hard to design, most of the time you will find Butterworth or Chebyshev designs in our little audio world.
A biquad~ has only two poles and zeroes, so you would need to cascade several stages to get a really sharp filter.
Ciao
Frank
yes, I had seen that graphic as well. but trying out the highpass iemfilters with noise, turns out that 10 hip~s work better.
Hehe, I guess that's because you filter noise: It's harder to recognize
the noise that the 10 hip stages added. :)
that's true. I started then using other signals with varying complexity,
and the output was better. btw, I was testing the signal only by looking
at it, through a spectral analysis.
I looked in my list of pd-ext objects, but found no elliptic filters.
are there any, or any way of getting their coefficients to use with biquad~?Elliptic filters are rather hard to design, most of the time you will
find Butterworth or Chebyshev designs in our little audio world.A biquad~ has only two poles and zeroes, so you would need to cascade
several stages to get a really sharp filter.
yep, I only found those. I'll try cascading other types of filters as
well. I thought there would be a magic filter somewhere that would do any
operation imaginable by men.
On 16 Mar 2009, at 09:25, Frank Barknecht wrote:
Jo?o Pais hat gesagt: // Jo?o Pais wrote:
I looked in my list of pd-ext objects, but found no elliptic
filters. are there any, or any way of getting their coefficients to use with
biquad~?Elliptic filters are rather hard to design, most of the time you
will find Butterworth or Chebyshev designs in our little audio world.
hey, joão
you still haven't made friends with fft? ;)
i've just finished a max external that does chebyshev filtering with
a variable number of poles and adjustable % of ripple in the
passband. repackaging this into a pd external wouldn't be too hard.
if you are interested in trying this out, drop me a mail and i'll see
when i can get it done.
volker.
Does anybody know how to wire a midi in on to the arduino.
any help appreciate markus
Does anybody know how to wire a midi in on to the arduino.
See the attached schematic. The Rx pin is whichever digital input on the Arduino you are using. The two MIDI wires are pins 4 and 5 of a DIN-5 connector. You must not connect the ground (shield or pin 2) from the MIDI cable. R2 and D1 are there to let you see the data as it arrives by flashing the LED, so they're not essential.
Martin
I looked in my list of pd-ext objects, but found no elliptic filters.
are there any, or any way of getting their coefficients to use with
biquad~?Elliptic filters are rather hard to design, most of the time you will
find Butterworth or Chebyshev designs in our little audio world.hey, joão you still haven't made friends with fft? ;) i've just finished a max external that does chebyshev filtering with a
variable number of poles and adjustable % of ripple in the passband.
repackaging this into a pd external wouldn't be too hard. if you are interested in trying this out, drop me a mail and i'll see
when i can get it done. volker.
ah ah, that's true. actually I didn't need anything so far that *had* to
be done with fft, so I never learned it properly (just unproperly a couple
of times, like yours, in which I didn't exercise afterwards).
of course I would be interested in the filter, as well as the rest of the
community should be. don't make any rush because of me, because it can
happen that after a while I don't need it anymore (in case you're too busy
with something else). but if you want to join your code to pd-ext,
everyone will be happy. (and windows compatible, please)
On 16 Mar 2009, at 17:58, João Pais wrote:
of course I would be interested in the filter, as well as the rest
of the community should be. don't make any rush because of me,
because it can happen that after a while I don't need it anymore
(in case you're too busy with something else). but if you want to
join your code to pd-ext, everyone will be happy. (and windows
compatible, please)
for anyone still insterested in this, i made a pd version of the
chebyshev filter, called vb_cheby~.
(prefix vb_ to avoid name clashes - popular theme...)
you can get it from here: http://www.esbasel.ch/Downloads/MaxMSP-Objects.htm
included is a compiled version for osx, standard makefile and source
code.
it also compiles fine on my linux machine - no idea about windows
though.
please feel free to send me your comments/complaints/suggestions. volker.
I have a question: are your externals included in pd-ext? that is, can
they be downloaded with the latest nightly builds? I would advise you to
get them into "the system", so that it's garanteed that (almost) everyone
has them automatically.
On 16 Mar 2009, at 17:58, João Pais wrote:
of course I would be interested in the filter, as well as the rest of
the community should be. don't make any rush because of me, because it
can happen that after a while I don't need it anymore (in case you're
too busy with something else). but if you want to join your code to
pd-ext, everyone will be happy. (and windows compatible, please)for anyone still insterested in this, i made a pd version of the
chebyshev filter, called vb_cheby~. (prefix vb_ to avoid name clashes - popular theme...)you can get it from here: http://www.esbasel.ch/Downloads/MaxMSP-Objects.htm
included is a compiled version for osx, standard makefile and source
code. it also compiles fine on my linux machine - no idea about windows though.please feel free to send me your comments/complaints/suggestions. volker.
On 24 Mar 2009, at 14:26, João Pais wrote:
I have a question: are your externals included in pd-ext? that is,
can they be downloaded with the latest nightly builds? I would
advise you to get them into "the system", so that it's garanteed
that (almost) everyone has them automatically.
hm, i'm certainly not against this idea, as i'm not opposed to
sharing what i've done, if it is interesting for others.
but i don't use pd-extended myself a a lot and actually i don't know
the process of "getting externals into the system".
also, i don't know about the policy of pd-extended or where it's
aiming at,
but i guess that the externals that get in there, must have somehow
proven to be useful to many users.
volker.
hm, i'm certainly not against this idea, as i'm not opposed to sharing
what i've done, if it is interesting for others. but i don't use pd-extended myself a a lot and actually i don't know the
process of "getting externals into the system".
that is quite easy: contact directly H-C Steiner to get a svn account,
then upload your code - or ask him more about how to do it.
also, i don't know about the policy of pd-extended or where it's aiming
at, but i guess that the externals that get in there, must have somehow
proven to be useful to many users.
ahhh, there's no policy. anyone who manifests interest in joining in joins
in, I guess. there's also no admittance test, so everyone just puts in the
code he wants (sometimes even buggy code). which makes the whole a thing
not really totally ÃŒbersichtlich (one of the GSC projects was to organize
and catalog the externals on pd-ext).
but pd-ext is quite used (only hardcore people/code developers prefer
pd-van), and it's the easiest way to make the code available to everyone.
of course it can happen that no one notices it's there, because there are
around 2000+ files in the /extra folder of pd-ext.
On 27 Mar 2009, at 01:41, João Pais wrote:
hm, i'm certainly not against this idea, as i'm not opposed to
sharing what i've done, if it is interesting for others. but i don't use pd-extended myself a a lot and actually i don't
know the process of "getting externals into the system".that is quite easy: contact directly H-C Steiner to get a svn
account, then upload your code - or ask him more about how to do it.also, i don't know about the policy of pd-extended or where it's
aiming at, but i guess that the externals that get in there, must have
somehow proven to be useful to many users.ahhh, there's no policy. anyone who manifests interest in joining
in joins in, I guess. there's also no admittance test, so everyone
just puts in the code he wants (sometimes even buggy code). which
makes the whole a thing not really totally übersichtlich (one of
the GSC projects was to organize and catalog the externals on pd-ext). but pd-ext is quite used (only hardcore people/code developers
prefer pd-van), and it's the easiest way to make the code available
to everyone. of course it can happen that no one notices it's
there, because there are around 2000+ files in the /extra folder of
pd-ext.
"not really totally übersichtlich", yeah i had that impression, too :) anyway, thanks for your comments, joão. i'll think about it. volker.
Are there any FIR filters in Pd? Those can have really steep slopes like 96dB per octave.
On Sun, Mar 15, 2009 at 1:37 PM, hard off hard.off@gmail.com wrote:
it's not possible to filter without ANY roll-off as far as i know.
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
-
brandt＠subnet.at -
chris clepper -
Frank Barknecht -
hard off -
João Pais -
martin.peach＠sympatico.ca -
volker böhm